Enterprise Modelling and Information Systems Architectures

An International Electronic Journal

Volume 9 No. 1 June 2014 Special Issue on A Roadmap for Business Informatics

Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Table of Contents 1

Table of Contents

Editorial Preface 2

Henderik A. Proper and Marc M. 5 Enterprise Architecture – Towards essential Lankhorst sensemaking Ulrich Frank 22 Enterprise Modelling: The Next Steps

José Tribolet and Pedro Sousa and Artur 38 The Role of Enterprise Governance and Cartography in Caetano Enterprise Engineering Jorge L. Sanz 50 Enabling Front-Office Transformation and Customer Experience through Business Process Engineering Eng K. Chew 70 Service Innovation for the Digital World

Stéphane Marchand-Maillet and Birgit 90 Big Data Management and Analysis for Business Hofreiter Informatics – A Survey Thomas Setzer 106 Data-Driven Decisions in Service Engineering and Management Imprint 118

Editorial Board 119

Guidelines for Authors 120 Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 2 Editorial Preface

Editorial Preface

In 2011 the IEEE Technical Committee on Elec- There is no doubt that Business Informatics is tronic Commerce decided to broaden its scope an inter-disciplinary field of study. It endeav- and, accordingly, rename itself to the IEEE Tech- ours taking a systematic and analytic approach nical Committee on Business Informatics and in aligning core concepts from management sci- Systems. In line with this change in name and ence, organisational science, economics informa- scope it decided to rename its flag ship confer- tion science, and informatics into an integrated ence to IEEE Conference on Business Informat- engineering science. Consequently, the field of ics (CBI). Following these changes, it has been Business Informatics involves a broad spectrum a first priority of the technical committee to ex- of more specific research domains that focus on actly define the meaning of the term "Business important aspects of Business Informatics in the Informatics" in an IEEE context and to underpin above mentioned context. For the first edition un- the need for a Business Informatics Conference der the new title and scope, it has been important under the umbrella of the IEEE. to sharpen the future research directions in the domain of Business Informatics. Thus, we had Evidently, the IEEE as the Institute of Electrical carefully selected appropriate research domains and Electronics Engineers, the world’s largest that represent the IEEE understanding of Busi- professional association for the advancement of ness Informatics. In order to reach a common un- technology, takes a mainly engineering sciences derstanding of these domains in our community, direction when approaching Business Informat- we invited distinguished experts to introduce a ics. In order to find its own scope for the IEEE research domain by defining its scope, its existing body of knowledge, and most importantly Conference on Business Informatics, we have its future research challenges. These keynotes been inspired by Nygaard who defined inform- have been a means to guide the community in its atics as the science that has as its domain in- way forward and provide directions for Business formation processes and related phenomena in Informatics in the IEEE CBI context. artefacts, society and nature. In the spirit of this definition, we consider Business Informat- In this special issue of the EMISA journal we in- ics as a scientific discipline targeting informa- clude seven papers, each based on a IEEE CBI tion processes and related phenomena in a socio- 2013 keynote introducing a research domain in economical context, including companies, organ- Business Informatics. Evidently, these papers are izations, administrations and society in general. neither classical research papers nor pure surveys, since they focus to a large extent on the "fu- A key characteristic of Business Informatics re- ture", i.e. the open research challenges (without search is that it considers a real-world business providing a solution). In the following, we define context in developing new theories and concepts the scope of the seven research domains and in that enable new practical applications. Thereby, parentheses we name the author(s) who intro- Business Informatics research does not only ex- duce(s) the domain by a paper presented in this tend the body of knowledge of the information special issue. society, but at the same time provides a tangible 1. Enterprise Architecture (Henderik A. Proper impact to industry. Consequently, Business In- and Marc M. Lankhorst) formatics is a fertile ground for research with the Scope: In contrast to partial architectures such as potential for immense and tangible impact. Or IT architecture or software architecture, enter- put it in other words - Business Informatics is prise architecture focuses on the overall enter- research that matters! prise. Enterprise architecture explicitly incorpor- Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Editorial Preface 3

ates business-related concepts and artefacts in from a unique perspective, namely, the integra- addition to traditional IS/IT artefacts. By embra- tion of information and people’s behaviour. cing an enterprise-wide perspective enterprise 5. Business (Model) & Service Innovation (Eng architecture provides a means for organizations Chew) to coordinate their adaptations to increasingly Scope: Being successful in business no longer fast changing market conditions which impact depends on having the "best" product, but in- the entire enterprise, from business processes to creasingly depends on delivering high quality ser- IT support. vices, through attractive customer-centric busi- 2. Enterprise Modelling (Ulrich Frank) ness models, at affordable costs. This forces en- Scope: Enterprise modelling is concerned with terprises to continuously develop/ innovate their the modelling of different aspects of an enter- services and renew/innovate their business mod- prise (goals, capabilities, organizational struc- els. The world’s evolution toward services-based tures, business processes, resources, information, clusters also brings new trends that blur the tradi- people, constraints, etc.) and their interrelation- tional boundaries across conventional industries, ships. Accordingly, enterprise modelling offers thus generating new opportunities for economies different perspectives of an enterprise suitable of scale and scope. This has led to increasing in- for strategic planning, organizational design and terests by disparate industries around the globe software engineering. It covers the notation and in the "art and science" of the practices of ser- semantics of enterprise modelling languages, the vice innovation. A new concept, called service- processes involved in creating and managing dominant logic, has recently been introduced in models, tool support, as well as quality of model- the business discipline to study service phenom- ling. ena - one that has significant cross-disciplinary implications for the research and design of IT- 3. Enterprise Engineering (Jose Tribolet, Pedro enabled service innovations and the attendant Sousa, and Artur Caetano) service systems. Scope: The enterprise engineering domain aims to apply an engineering based approach to the 6. Empowering & Enabling Technologies (Ste- design of enterprises and their transformation. phane Marchand-Maillet and Birgit Hofreiter) As such, this domain is concerned with the de- Scope: Enabling technologies in Business Inform- velopment of new, appropriate theories, mod- atics integrate management practices with In- els, methods and other artefacts for the analysis, formatics and Information Technologies. design, implementation, and governance of en- Business Informatics tasks may be performed, terprises by combining (relevant parts of) man- supported or monitored by automated or semi- agement and organization science, information automated technologies. Running environments systems science, and computer science. range from thin mobile clients to large-scale distributed platforms, and newer areas such as ana- 4. Business Process Engineering (Jorge Sanz) lytics services, big data. Accordingly, we seek Scope: Business Informatics deals with informa- papers for original and innovative empowering tion processes in organizations, industries and and enabling technologies in domains related to society at large. This concept of "information Business Informatics. in motion" links to business processes deeply. Processes are the expression of the behaviour 7. Data-Driven Service and Market Engineering of organizations and this behaviour leaves foot- (Thomas Setzer) prints in the form of artefacts of all sorts, in- Scope: Economic problems faced by today’s or- cluding information. Thus, Business Informatics ganizations as well as society as a whole de- profoundly intersects with the social enterprise mand interdisciplinary knowledge from econom- Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 4 Editorial Preface

ics, management and informatics. Thus, eco- Guest Editors’ contact information: nomic modelling of IT-based solutions for ana- Dr. Birgit Hofreiter lytically and statistically formulated economic Electronic Commerce Group problems is subject to this track. In particular, Vienna University of Technology we are interested in the intelligent reduction of Favoritenstrasse 9-11 problem-relevant features from vast datasets In- 1040 Vienna, Austria cluding customer dynamics, market behaviour, resource usage, etc. Prof. Dr. Christian Huemer It should be noted that these research domains Business Informatics Group represent cornerstones of the CBI conference Vienna University of Technology series. However, it is our vision to complement Favoritenstrasse 9-11 the CBI picture on business informatics by other 1040 Vienna, Austria appropriate research domains. We plan to introduce these domains both at future CBI Keynotes and subsequent special journal issues. All articles in this EMISA special issue were handed in by domain experts that have given a keynote presentation at the IEEE Int’l Confer- ence on Business Informatics (CBI 2013), Vienna, 15th - 18th July 2013. These invited papers have then undergone a blind review for EMISA journal publication. Each paper had been assigned to two international reviewers. The reviewers for each papers have been chosen on the following criteria: The first reviewer has been a member of the IEEE CBI steering committee in order to ensure compliance of the paper with the scope of business informatics in an IEEE context. The second reviewer has been an accredited expert in the respective research domain not being involved in the IEEE CBI organization in order to ensure an open and unbiased representation of the domain. In the case where one guest editor is a co-author of the paper, the review process was managed by the other guest editor. Birgit Hofreiter Christian Huemer Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Enterprise Architecture 5

Henderik A. Proper and Marc M. Lankhorst

Enterprise Architecture Towards essential sensemaking

In this position paper, we discuss our view on the past and future of the domain of enterprise architecture. We will do so, by characterising the past, and anticipated future, in terms of a number of trends. Based on these trends, we then discuss our current understanding of the future concept and role of enterprise architecture. We conclude by suggesting vantage points for future research in the field of enterprise architecture.

1 Introduction 2.1 From Computer Architecture to IS Architecture Increasingly, organisations recognise enterprise architecture as an important instrument to steer The origins of enterprise architecture can be traced (or influence) the direction of transformations back to the concept of information systems ar- (Buckl et al. 2011; Greefhorst and Proper 2011; chitecture (IS Architecture), which in turn has Lahrmann et al. 2010; Lankhorst 2012b; Op ’t Land its roots in the concept of computer architecture. et al. 2008; The Open Group 2009). Over the past One of the first references to the term architec- decades, the domain of enterprise architecture ture, in the context of IT, can be found in a paper has seen a tremendous growth, both in terms from 1964 on the architecture of the IBM Sys- of its use and development in practice and as a tem/360 Amdahl et al. 1964. There it was used to subject of scientific research. The roots of the introduce the notion of computer architecture. domain can actually be traced back as far as the Later, in the 1980s, the term architecture star- mid 1980s. ted to become used in the domain of informa- In this position paper, which builds on Proper tion systems development as well. This occurred (2012), we will review the evolution of the field both in Europe and North America. The North of enterprise architecture. We do so by charac- American use of the concept of architecture in terising both its history (Sect.2), as well as its an information systems context can (at least) be anticipated future (Sect.3), in terms of a number traced back to a report on a large multi client 1 of trends. Based on these trends, we also dis- study, the PRISM project Hammer & Company cuss our current understanding of the concept (1986) conducted, as well as the later paper by and role of enterprise architecture (Sect.4). We Zachman (1987). The European origins can be conclude with a brief discussion of our view on traced back to the early work of August-Wilhelm research in the field of enterprise architecture in Scheer on the ARIS framework, also dating back terms of key vantage points for further research. to 1986 (Scheer 1986, 1988, 2000). In Europe, the ARIS framework as developed by 2 A History of Enterprise Architecture August-Wilhelm Scheer eventually formed the base of the well known IDS-Scheer toolset. In In this section we discuss the history of the field 1Not to be confused with the present day concept of of enterprise architecture in terms of a number PRISM http://en.wikipedia.org/wiki/PRISM_(surveillance_ of trends as observed by us. program) Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 6 Henderik A. Proper and Marc M. Lankhorst

North America, the PRISM project was a multi- is dependent on the key concerns/stakes of the year research project, led by Michael Hammer, stakeholders involved in an architecting effort. Thomas Davenport, and James Champy. PRISM, The basic idea to consider information systems in short for Partnership for Research in Informa- a holistic way, i.e., from multiple related perspect- tion Systems Management, was sponsored by ives, was actually already identified before being approximately sixty of the largest global compan- linked to the term information systems architec- ies (DEC, IBM, Xerox, Texaco, Swissair, Johnson ture. For example, Multiview Wood–Harper et al. and Johnson, Pacific Bell, AT&T, etc.). This re- (1985) already identified five essential viewpoints search effort produced an architecture framework for the development of information systems: Hu- known as the PRISM Architecture Model, which man Activity System, Information Modelling, was published in 1986. The PRISM framework has Socio-Technical System, Human-Computer In- strongly influenced other enterprise architecture terface and the Technical System. Even though standards, methods and frameworks (Beijer and the authors of Multiview did not use the term De Klerk 2010; Davenport et al. 1989; Richardson architecture, one can argue that Multiview is ef- et al. 1990; Rivera 2007). fectively one of the earliest explicit information Many years later, the PRISM report also influ- systems architecture frameworks. During the enced the IEEE definition of architecture, as many same period in which Multiview was developed, of the IEEE 1471 committee members were em- the so-called CRIS Task Group of the IFIP work- ployed by the original sponsors of their earlier ing group 8.1 developed similar notions (Olle work on PRISM. Key people involved in PRISM et al. 1982, 1983), where stakeholder views were later also spearheaded the wave on Business Pro- captured from different perspectives. Special at- cess Reengineering (Davenport et al. 1989; Ham- tention was paid to disagreement about which mer 1990), which is essentially an early business aspect (or perspective) was to dominate the sys- architecting effort. tem design (viz. “process”, “data” or “behaviour”). The Zachman (1987) paper is often referred to as In the early 1980s, the CRIS Task Group already human roles one of the founding papers of the field of enter- identified several (stakeholders!) in- prise architecture. It should be noted, however, volved in information system development, such responsible executive development coordinator that both the PRISM and ARIS frameworks pre- as , , business analyst business designer date the Zachman framework, although these , , quite similar frameworks have indeed been published in less to the stakeholder dimension of, e.g., the Zach- accessible sources. man framework. The important message of the ARIS, PRISM and In the 1990s, challenges such as interoperability Zachman frameworks is the need to consider and distributed computing resulted in the cre- information systems from multiple perspectives ation of reference architectures, including the based on stakes, concerns, as well as different as- CIMOSA (Open System Architecture for CIM) pects of the information systems and its business framework for computer integrated manufactur- or technology context, while at the same time ing systems ESPRIT Consortium AMICE 1993 focusing on the key properties of the informa- and the RM-ODP (Reference Model for Open Dis- tion system. The latter focus is also captured tributed Processing) framework for information by the phrase fundamental organization in the systems (ISO 1996a,b, 1998a,b) IEEE 1471 IEEE 2000 architecture definition: “the fundamental organization of a system embodied 2.2 From IS Architecture to Enterprise in its components, their relationships to each other Architecture and to the environment, and the principles guid- The awareness that the design of information ing its design and evolution.”, where fundamental systems needed to be seen in a broader business Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Enterprise Architecture 7

and enterprise context, triggered several authors Without an attempt to be complete, some enter- to shift towards the use of the term enterprise prise architecture approaches that indeed take a architecture rather than information systems ar- more co-design oriented perspective include: the chitecture. One of the first authors to use the Integrated Architecture Framework (IAF) (Goed- term enterprise architecture was Spewak (1993). volk et al. 1999; Wout et al. 2010), the ArchiMate The initial architecture approaches focused on (Jonkers et al. 2003; Lankhorst 2012b) language, the development of information systems, while as well as the DYA (Wagter et al. 2001, 2005) and taking the models/architectures of other relevant DEMO (Dietz 2006) methods. Also the most re- aspects of the enterprise as a given. However, cent version of TOGAF (The Open Group 2009) due to the strong connection between business does indeed suggest to co-design the business processes and the underlying information sys- architecture and the information systems archi- tems, it was only natural to not just treat such tecture. perspectives as a given, but rather to co-design these in tandem with the information systems 2.3 From Business-to-IT-stack to and their underlying IT support. Enterprise Coherence Earlier versions of TOGAF (The Open Group The realisation that information systems archi- 2005), including TAFIM (1996), treated business tecture and business architecture need to be co- architecture as a given thing. By defining Enter- designed in tandem, led most enterprise archi- prise Architecture Planning (EAP) as “the process tecture approaches to capture a business archi- of defining architectures for the use of information tecture in terms of building blocks such as busi- in support of the business and the plan for im- ness services, business processes, business actors, plementing those architectures”, Spewak Spewak etc. These business building blocks were then 1993 also seems to suggest to take business archi- linked to information systems, and ultimately tecture as a given. Boar (1999) in “Constructing IT infrastructures, resulting in a ‘Business-to-IT- Blueprints for Enterprise IT architectures” does the stack’. Among an increasing group of researchers same. and practitioners, the ‘reduction’ of ‘the architecture of the enterprise’ to a ‘Business-to-IT-stack’ The shift from taking a business architecture as a caused unease. In particular Graves (2008), Feh- given input, to the realisation that business and skens (2008) as well as Wagter (2009) have ar- IT should be co-designed as a whole, could be gued that such a Business-to-IT-stack centricity seen as the birth of modern day enterprise archi- is a major weakness of contemporary enterprise tecture. The strategic alignment model by Hende- architecture approaches, and that enterprise ar- rson and Venkatraman (1993) has played an im- chitecture should involve many more aspects of portant role in taking this step to the co-design an organisation, such as a clear connection to its of business architecture and information systems strategy, its financial structures, the abilities of its architecture. Henderson and Venkatraman (1993) work force, etc. More specifically, Wagter (2009) indeed suggests that aligning business and IT argue that enterprise architecture should not just should not necessarily require that the business be concerned with Business-IT alignment, but strategy should be treated as a given. There are rather with the alignment of all relevant aspects several ways to align business and IT. Also the of an enterprise. Therefore, rather than using work by, e.g., Tapscott and Caston (1993) contrib- the term alignment, Wagter (2009) suggest to use uted to this realisation, as well as the work by the term enterprise coherence to stress the multi- Ross et al. (2006). The earlier mentioned work on faceted nature. Business Process Reengineering (Davenport et al. 1989; Hammer 1990), essentially an early business A first enterprise architecture method to indeed architecting effort, also contributed to this shift. explicitly move beyond a Business-to-IT-stack Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 8 Henderik A. Proper and Marc M. Lankhorst

centricity is the GEA method (Wagter 2009). GEA As argued in Wagter et al. (2011), most tradi- argues that the coherence between several as- tional approaches and frameworks, including the pects of an enterprise needs to be governed ex- Sowa and Zachman (1992) and IAF (Wout et al. plicitly by means of an enterprise architecture. 2010) frameworks, the ArchiMate (Iacob et al. To indeed co-design the different aspects of an 2012; Lankhorst 2012b) language, as well as the enterprise architecture, and to use it (both the DYA (Wagter et al. 2005) and TOGAF (The Open co-design process, and the resulting architecture) Group 2009) architecture methods, essentially in governing enterprise coherence, it is necessary take a Blue-print perspective on change. The to take the concerns and associated strategic dia- need to really involve senior management, how- logues of senior management as a starting point. ever, suggests the use of another style of think- In other words, the way in which architecture is ing, involving internal or external stakeholder integrated into the strategic dialogue should take interests, strategy formulation processes, formal the concerns, language, and style of communic- and informal power structures, and the association of senior management as a starting point, ated processes of creating win-win situations and and not the typical domains, layers, or columns, forming coalitions. In terms of De Caluwé and as identified in the traditional architecture frame- Vermaak (2003) this would suggest to comple- works. ment the Blue-print perspective with the Yellow- The shift from Business-to-IT-stack centricity to print perspective, and arguably also a mix of the the broader notion of enterprise coherence also other perspectives. required a change in perspective on change pro- In the development of the GEA method (Wagter cesses in organisations (Wagter et al. 2011). De Caluwé and Vermaak (2003) have identified a 2009), this line of thinking was taken as a starting number of core perspectives on change processes point. As a result, the actual political power struc- in organisations: tures, and associated strategic dialogues, within a 1: Yellow-print thinking: Bring the interests of specific enterprise were taken as a starting point, the most important players together by means of rather than the frameworks suggested by exist- a process of negotiation enabling consensus or a ing architecture approaches. This leads to en- win-win solution. terprise specific frameworks of coherence gov- 2: Blue-print thinking: Formulate clear goals and ernance perspectives, to manage enterprise coher- results, then design rationally a systematic ap- ence. For example, in terms of ‘mergers & acquis- proach and then implement the approach accord- itions’, ‘human resourcing’, ‘clients’, ‘regulators’, ing to plan. ‘culture’, ‘intellectual property’, ‘suppliers’, etc. 3: Red-print thinking: Motivate and stimulate The existing Blue-print oriented frameworks can people to perform best they can, contracting and still be used to further structure the dialogue rewarding desired behaviour with the help of between the coherence governance perspectives, HRM-systems. especially where it concerns issues pertaining to 4: Green-print thinking: Create settings for learn- the Business-to-IT-stack. ing by using interventions, allowing people to become more aware and more competent on their It is to be expected that organisations aiming to job. use enterprise architecture to steer major trans- 5: White-print thinking: Understand what under- formations, will increasingly move from a Busi- lying patterns drive and block an organisation’s ness-to-IT-stack centricity perspective to an en- evolution, focusing interventions to create space terprise coherence perspective on their enter- for people’s energy. prise architectures. Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Enterprise Architecture 9

2.4 From Big-Design-Up-Front to framework used, etc. This situation triggered Fit-for-Purpose the agile software development community to talk about Ambler and Jeffries (2002); Beck et al. Early frameworks and languages for enterprise (2001); Cockburn (2002); Lankhorst (2012a)“Big- architecture (Lankhorst 2012b; The Open Group Design-Up-Front” (BDUF). Of course, experienced 2005; Wout et al. 2010; Zachman 1987) were primar- architects knew when to stop architecting. How- ily concerned with the identification of the as- ever, early architectural approaches did not provide pects, concepts and domains that should be in- clear guidelines to ensure that architectures stayed cluded in an architecture; hence the resulting con- Fit-for-Purpose, and rather invited architects to tent frameworks. This orientation brings along be over-complete. the risk that architects focus more on completeness of architecture descriptions, rather than on The need to tune an enterprise architecture to ensuring that the descriptions meet the purposes the purpose at hand and avoid overly detailed ar- for which they are actually needed. Accepted chitectures, triggered the authors of Wagter et al. standards for defining architecture, such as the (2001, 2005) to create the DYnamic Enterprise Ar- earlier quoted IEEE 1471 IEEE 2000:“the funda- chitecture approach, which incorporates notions mental organization of a system embodied in its such as “just enough architecture”, resembling the components, their relationships to each other and ideas that were also put forward (in parallel) by to the environment, and the principles guiding its the agile system development community. The design and evolution.” do not provide a clear most recent version of TOGAF (The Open Group ‘stop criterion’ for architects that allows them to 2009) also provides indications for different (pur- provide just enough architecture. This definition pose specific) ways to use its ADM to ensure points primarily at what the things are that an the resulting architecture descriptions are indeed architecture is concerned about:“its components, fit-for-purpose. their relationships to each other and to the environment, and the principles guiding its design and Greefhorst and Proper (2011) suggest to make a clear distinction between: evolution”. The risk is that inexperienced (and 1: The purpose that an enterprise architecture method obeying) architects loose themselves in serves. For example, to understand (make sense meticulous designs of the future enterprise. The of) the current/past situation of an enterprise in reference to “the fundamental organization” only terms of its fundamental properties and concepts, implicitly refers to the purpose for having an ar- to articulate and motivate (make sense of) a de- chitecture, i.e., understanding or expressing the sired future situation in terms of fundamental fundamental organisation of a system. But why? properties and concepts. And what part of organisation is to be regarded as 2: The meaning of an enterprise architecture as fundamental? This is of course dependent on the an artefact. For example, to express (for some purpose for which the architecture (description) purpose) the fundamental properties and/or con- is created. The more recent ISO (2011) version of cepts that underly the present structure of an this definition: “fundamental concepts or properties of a system in its environment embodied in its enterprise, or to express the fundamental prop- elements, relationships, and in the principles of its erties and/or concepts that should inspire, guide, or steer, the evolution towards the future. design and evolution”, does not remedy this. 3: The elements of an enterprise architecture in In our observation, the focus on completeness terms of the typical concepts used to capture this indeed quite often results in overly-detailed ar- meaning, such as its elements, relationships, and chitecture descriptions, involving long lists of the principles of its design and evolution as men- architecture principles, meticulously worked out tioned by the IEEE and ISO definitions, which models for each of the cells from the architecture may be captured by means of models and views. Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 10 Henderik A. Proper and Marc M. Lankhorst

This distinction enables a clear top-down reas- section, resulted in the awareness that another oning on the level of detail and completeness means was needed next to the traditional archi- needed from an architecture description. Given tecture descriptions involving the enterprise’s the purpose of a specific architecture (descrip- construction in terms of actual building blocks tion), one can identify the desired meaning of (value exchanges, transactions, business processes, the architecture, and following that, the kinds of actors, objects, roles, collaborations, etc). This elements needed to capture/express this mean- resulted in a strengthening of the role of archi- ing. For example, Greefhorst and Proper (2011) tecture principles as a way to translate an enter- focus on using enterprise architecture for the prise’s strategic intentions to more specific dir- purpose to align the enterprise to its essential ecting/guiding statements, without immediately requirements and ultimately its strategy: “... the ‘jumping’ to the use of actual building blocks main purpose of an enterprise architecture is to of an actual (high level) design. Several archi- align an enterprise to its essential requirements. tecture approaches indeed position architecture As such, it should provide an elaboration of an principles as an important element of enterprise enterprise’s strategy to those properties that are ne- architecture (Beijer and De Klerk 2010; Daven- cessary and sufficient to meet these requirements”. port et al. 1989; Op ’t Land et al. 2008; Richardson Even though it is only normative in nature, the et al. 1990; Tapscott and Caston 1993; The Open “necessary and sufficient” and the reference to the Group 2009; Wagter et al. 2005; Wout et al. 2010), enterprise’s strategy provide a (possible) stop- while some authors even go as far as to position ping criterion to keep an architecture Fit-for- principles as being the essence of architecture Purpose (i.e., steering transformations that aim (Dietz 2008; Fehskens 2010; Hammer & Company to establish an enterprise’s strategy changes). 1986; Hoogervorst 2009). In our view, the challenges of dealing with increased scope and com- 2.5 From a Constructing to a plexity really emancipated the role of principles Constraining Perspective as ways to constrain design space. The shift from computer architecture to information systems architecture, and then to enter- Fundamentally, we can see a shift from consider- prise architecture at large, also resulted in an ing an architecture as being primarily concerned constructing increase of scope of architecture efforts. Where with the (high level) design of an at the start the focus was typically on a limited enterprise in terms of building blocks to being constraining number of applications in support of an informa- concerned with the space of allow- tion system, the organisational scope gradually able/desirable constructions. A prime example constraining broadened to business-unit wide, then to enter- of an architecture from a point of prise wide, or sometimes even to a sector/branch view is the NORA (Nederlandse Overheid Ref- wide scope. At the same time, the potential time- erentie Architectuur (NORA) 2012) reference ar- horizon for architectures increased, from focus- chitecture for the Dutch government. It focuses ing on the situation after the next development primarily on architecture principles that should stage, to mid-term and longer-term planning be applied in the elaboration of more specific activities covering several intermediary stages. architectures and designs. The shift from Business-to-IT-stack centricity to It is important to note that the distinction between more overall enterprise coherence also resulted constructing an assembly of building blocks and in a wider range of aspects to be covered in an constraining the set of possible assemblies to an architecture. allowable/desirable subset, is orthogonal to the 2 The resulting increase in scope and complexity, deontic modality of an architectural description. combined with the Big-Design-Up-Front to Fit- 2See for example http://en.wikipedia.org/wiki/Deontic_ for-Purpose trend as discussed in the previous modality Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Enterprise Architecture 11

This refers to the question if the architectural Light-weight, iterative methods have gradually description is, for example, intended to be a sug- taken over much of the software development gestion (could), guidance (advisable), indicative community. Since the 1990s, evidence has been (should), or a pure directive (must). mounting that agile ways of working, using short iterations and close customer contact, have a 2.6 From Building to Integrating higher success rate than traditional, waterfall- Another trend also resulted in a similar shift to- like methods for software development, at least wards to the constraining of design space. In- for many types of software projects. Recent stud- stead of developing their own software, most or- ies provide theoretical and empirical evidence ganisations today use packaged solutions, cloud for the effectiveness of agile methods; see for services and other pre-defined solutions to sup- example the extensive overview by Lee and Xia port large parts of their business activities. These (2010). solutions may be configured with the organisa- The Agile Manifesto values “responding to change tion’s business rules, business processes, inform- over following a plan” (Beck et al. 2001). Many ation models, etc., but they inherently limit the proponents of agile methods are opposed to the design freedom of the architect. The upside, of use of architecture, categorically classifying it as course, is in the common gains of re-use: lower- Big-Design-Up-Front. They argue that stakehold- ing costs and risks, and speeding up developers cannot know what they really need and the ment. problem will change anyway before the project This trend, combined with the growing scope and is completed, so one cannot provide any useful complexity outlined in the previous section, also designs up-front. Moreover, the changing busi- leads to a growing emphasis on the integration ness environment makes stable requirements an of various business processes and IT compon- illusion to begin with. Hence, complex socio- ents, within and across organisations. Anyone technical systems cannot be designed solely be- who has spent some time in a large organisa- hind the drawing board. tion will recognise that the most common and at the same time most pernicious problems in On the other hand, many architects and man- architecture are at these integration points. The agers resist the agile movement, arguing that service-oriented architecture (SOA) paradigm Erl one should think before planning actions and 2005 was an important attempt to alleviate this building systems. They fear a loss of control and problem, but has not been the panacea that it claim that all these agile projects will build their was once thought to be. own silos, resulting in the same fragmentation of IT landscapes that the architecture discipline This shift towards integration also influences the promised to fix. design and development process. Whereas in the past, a large system was often designed in one go Both positions are misguided about the role of and as a single, coherent whole, an integrative architecture. A well-defined architecture helps approach will need to be more piecemeal and in positioning new developments within the con- iterative: adding and integrating various com- text of the existing processes, IT systems, and ponents one-by-one. other assets of an organisation, and in identifying necessary changes. A good architecture and in- 2.7 From One-shot to Iterative frastructure is an up-front investment that makes Approaches later changes easier, faster and cheaper, and a The agile movement in software development good architectural practice helps an organisation (Ambler and Jeffries 2002; Cockburn 2002) has re- innovate and change by providing both stability ceived much attention over the last two decades. and flexibility (Lankhorst 2012a). But this does Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 12 Henderik A. Proper and Marc M. Lankhorst

not mean that everything should be architected 3.1 From IT to IT up-front. As addressed in Sect. 2.4 and Sect. 2.5, In most enterprises the role of IT started with a good enterprise architecture is not overly de- the ‘automation of administrative work’. In mod- tailed and focuses on the essential inspiration ern day organisations, there continues to be a and guidance needed to foster enterprise-wide clear role for IT to automate administrative in- coherence. formation processing. However, the use of IT Architecture processes in many organisations has moved far beyond this. In some situations, still give the impression that architects should IT has given rise to new social structures, and do all the thinking beforehand and software de- business models. Consider, for example, the developers and others can only start their work velopment of social media, the (acclaimed) role of after the architects are done. Methods like TO- twitter in time of social unrest, the emergence of GAF’s ADM (The Open Group 2009) are also on-line music stores, app-stores, music streaming easily interpreted in this way. The measurable services, etc. The advent of ‘big data’ (Hurwitz success of agile methods and related develop- et al. 2013) is expected to drive such developments such as continuous delivery (Humble 2010) ments even further by allowing IT based systems creates an increasing need for the architecture to use statistical data to tune their behaviour to discipline to follow suit and embrace a more iter- observed and learned trends. ative way of working, closely tied to the entire At the same time, IT is becoming firmly embed- development process and not merely as a starting ded in existing technological artefacts. The cars phase. in which we drive now contain more lines of code than typical banking applications do. The The trend towards less detailed and more normat- next generation of cars will even be able to (par- ive enterprise architecture, as outlined in Sect. 2.4 tially) do the driving for us. The so-called smart and Sect. 2.5, matches well with this need for (power) grid, is likely to lead to the ‘smartening’ an iterative approach. Agile enterprise archi- of household appliances. Our houses are already tects provide assistance to projects to help them being vacuumed by dedicated robots, while in fit within the big picture, while refraining from some cases robots even play a role in the care of too much and too detailed guidance. Moreover, elderly people (Tamura et al. 2004). The military as Ciborra (1992) argued, bricolage, emergence use of all sorts of drones also spearheads more and local improvisation, instead of central con- peaceful applications of such self-reliant devices trol and top-down design, may lead to strategic that can, e.g., perform tasks on behalf of us in advantages: the bottom-up evolution of socio- hostile or unpleasant environments. technical systems will lead to something that is deeply rooted in an enterprise’s organisational In sum, we argue that we are moving towards culture, and hence much more difficult to imitate smart and more ‘sociable’ technology that is en- by others. Agile enterprise architects leave room abled by computer technology. One might indeed information intelligent for such local, bottom-up improvements and fit say, from technology to these within the larger scheme of things. technology, i.e., from IT to IT. When architecting modern day enterprises, one should treat these as (evolving) collectives of human actors 3 Future trends and computerised actors, where the latter might operate in a pure software world, or might be em- In this section we discuss the anticipated future bedded/embodied in other forms of (connected) of enterprise architecture in terms of a number technology. Needless to say, however, that hu- of anticipated trends. man actors will always need to remain (socially Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Enterprise Architecture 13

and legally) responsible for the actions of the also puts more emphasis on the need for se- computerised actors that operate on their behalf. mantic interoperability. Different semantic backgrounds in a multi-organisational setting make 3.2 From Syntax to Semantics this even more complicated. We need gradual, iterative approaches for coherent and collaborat- The trend towards an increased scope of integ- ive design, development and deployment of these ration, described in Sect. 2.6, brings its own set socio-technical systems. of design issues. Although paradigms such as service orientation promised to facilitate this in- 3.3 From State-thinking to tegration, they function mainly on a syntactic Intervention-thinking level, providing a stack of interconnection stand- We argue that contemporary approach to archi- ards for software systems. tecture ‘think’ in terms of as-is and to-be states of When the integration scope grows, the associated the enterprise. Some approaches may indeed go semantic problems grow as well. The informa- as far as identifying several intermediary stages tion shared across organisational borders may between as-is and to-be, e.g., leading to the con- be interpreted in ways that were not intended cept of transition architecture in TOGAF (The and do not match with the context in which this Open Group 2009) and plateaus in ArchiMate information originated. The same holds for the (Iacob et al. 2012; Lankhorst 2012b). What remains behavioural semantics of cross-border business common, however, is the focus on several states processes. The Semantic Web (W3C Semantic of the (construction of the) enterprise. This state- Web Activity 2013) provides some partial solu- oriented thinking might have worked well in the tions, but the premise of its methods is the uni- past when the focus was on architecting an en- fication of semantics in a single overarching on- terprise’s IT support. However, as soon as other tology, basically trying to standardise the mean- other aspects are taken into consideration, the ings of information. It is simply not feasible to story becomes more complicated. build such ontologies for the size and variety of As soon as non-technological aspects are taken real-world integration problems. Local variety into consideration, this brings about a shift of in semantics cannot be avoided or ‘standardised focus from technical systems to socio-technical away’, because of the inevitable loss of meaning systems involving a mix of human and technolo- this causes. gical actors. The enterprise and its environment, This problem is exacerbated by the rapidly grow- being socio-technical systems, will evolve out ing volume, variety and velocity of ‘big data’ of themselves. People working in an enterprise will make changes to the ‘design’ of the enter- (Hurwitz et al. 2013), as already mentioned in prise, if only to make the ‘design’ (continue to) Sect. 3.1. Applying statistical methods will not work in day-to-day practise. The people making suffice to create meaningful interpretations. This up the organisation, collectively ‘author’ their implies that novel methods are needed for archi- enterprise (Taylor and Van Every 2010). tecting the semantics of information and behaviour, taking into account the variety and context Even without the use of architecture as a plan- of meaning and the social processes needed to ning instrument, there are likely to be a plethora create understanding and agreement at different of projects and related efforts that will continu- scales. It is not feasible to provide complete top- ously change the enterprise in response to ex- down designs for large-scale socio-technical sys- ternal and/or internal stimuli. Some of these tems, as we have already argued in Sect. 2.7. The changes might not even be ‘visible’ as projects, shift from building towards integration (Sect. 2.6), as they are based on local initiatives taken within Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 14 Henderik A. Proper and Marc M. Lankhorst

the operational processes (i.e., actors switching It is important to realise that the humans in- between a role in the operational capability to volved in an enterprise can play a role towards the transformation capability). both the operational capability and the trans- We argue that a shift is needed from thinking formation capability simultaneously. For human of enterprise transformations as being a change beings this is actually quite natural. While ex- of an enterprise from one state (the as-is) to a ecuting our daily activities, we typically also future state (the to-be), but rather as primarily learn how to do these activities better and/or being an intervention in the natural evolution of adapt them to changing needs/circumstances. In the enterprise, resulting in a changed course of its these cases, we decide to ‘on the fly’ innovate our evolution towards a presumably more desirable operational capability. In doing so, we (briefly) direction. So, from an as-is trajectory to a to-be use our transformation capability. As a con- trajectory. sequence, it is advised to regard the operational For the focus of an enterprise architecture this capability and transformation capability of an en- aspect systems sub systems would lead to an even stronger emphasis of the terprise as and not as . constructing to constraining trend as discussed in When considering an enterprise from an architec- Sect. 2.5, as constraints are more suitable to artic- tural perspective, one can of course opt to focus ulate desired trajectories than specific building the architecture efforts on one of these capabil- blocks. Using, e.g., architecture principles enterprises can distinguish between desirable and less ities or both. In most cases that we know of, as desirable directions of its evolution, and from well as the illustrating case studies discussed in that infer interventions that can be undertaken the existing architecture approaches, the focus operational capability to drive, or lure, the natural evolution of the is on architecting the only. enterprise in the desired direction. These inter- An exception would be enterprises who have cre- ventions might indeed involve (re-)constructions ated a so-called development architecture focusing of building blocks of the enterprise. on the way the enterprise will go about developing new information systems. An example is 3.4 From Operational Capability to the development architecture from the Dutch Tax Transformation Capability Administration (Achterberg et al. 2000). In line with the previous trend, an enterprise is Whether an enterprise’s architecture effort should likely to evolve continuously. The capabilities focus on the operational capability and/or the needed to change an enterprise are quite differ- transformation capability depends on the enter- ent from the capability needed to run its day-today business. The latter capabilities of an enterprise’s strategy. For example, in terms of the prise can be referred to collectively as its opera- Discipline of Market Leaders from Treacy and tional capability, while the capabilities needed to Wiersema (1997), it would be logical for enter- transform itself are the transformation capability. prises focusing on: Teece et al. (1997) stress the need for modern day 1: operational excellence, that the operational cap- organisations to have a transformation capabil- ability requires architecting priority, ity that meet its rapidly changing environment, 2: product leadership, that the parts of the trans- leading to a highly dynamic transformation cap- formation capability dealing with product/service ability: “the firm’s ability to integrate, build, and innovation require architecting priority, reconfigure internal and external competences to 3: customer intimacy, that the parts of the op- address rapidly changing environments”. Teece erational capability and the transformation cap- et al. (1997) refer to this dynamic transformation ability that deal with client interaction require capability as “dynamic capability”. architecting priority. Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Enterprise Architecture 15

When indeed also architecting the transforma- In addition, due to the increasing amount of tion capability, it is again recommendable to real- shareholder value (and/or taxpayer’s money) that ise that the operational and transformation cap- is tied up in such transformations, one can ex- abilities are aspect systems, and that the different pect that the requirements on the transparency actors (be they human or be they technology) with which such decision are made, will increase. can play roles towards both capabilities simultan- Would it not be logical for companies that are eously. listed on the stock market, to also report annually In recent work on agile service development on their ability to transform in an effective way? (Lankhorst 2012a), it was also argued that an agile In other words, not just how well their opera- services context requires enterprises to move tional capability is able to earn a revenue for its from having only an efficient operational capabil- shareholders, but also how well their transforma- ity to an effective combination of operational and tion capability is able to ensure the continuation transformation capabilities. One should focus on of this revenue in a cost-effective way? designing the operational capability in such a In this sense, one can expect that senior manage- way that it lends itself to quick changes within ment will increasingly be held responsible (by given boundaries and ambitions, while the trans- shareholders, tax payers, and ultimately auditors) formation capability should be designed in such for their ability to steer and control transform- a way that it can use this built in agility of the op- ations. Even more, senior management should erational capability to meet anticipated changes not only worry about the cost effectiveness of in the environment, as well as the ability to take change, but also about governance, risk manage- appropriate actions to transform the operational ment, compliance, etc., associated to these trans- capability when having to meet unanticipated changes (in terms of Teece, it would have to be formations. Given the earlier discussion on the dynamic). purpose of enterprise architecture, and its role for informed governance, it shall not be surpris- In Lankhorst (2012a) some guidelines are pro- ing that we take the point of view that enterprise vided on how to balance an architecting effort architecture would indeed provide a means to between the transformation and operational cap- senior management to take more control over abilities. However, more research is needed. At the transformations and the associated decision the same time, the need for enterprises to be agile, making on the future of the enterprises for which does stress the need to be able to make explicit they are responsible. Using enterprise architec- tradeoffs on how to deal with this agility across ture, one can more crisply analyse problems in the two capabilities. an existing situation, articulate desired directions 3.5 From Intuition-based to (using architecture in a prescriptive way), ana- Evidence-based Management lyse the costs and benefits of different options (using architecture in a more descriptive way), Modern day enterprises need to change in order and guard that transformation projects are in- to survive. At the same time they need to do so in deed moving in the desired direction. the face of an increasing number of regulations on compliance and transparency. Furthermore, a In parallel to this, one can also observe an in- considerable part of an enterprise’s shareholder teresting trend in the field of management. As value is ‘tied’ up in the needed transformations. argued in (Pfeffer and Sutton 2006, 2011), there As a consequence, the processes needed to trans- is an increasing call for evidence-based manage- form the enterprise become a core business pro- ment instead of (yet not fully replacing) intuition- cess themselves, requiring ample management based management. The authors draw an inter- attention. esting analogy to the trend in medicine towards Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 16 Henderik A. Proper and Marc M. Lankhorst

evidence-based medicine (Evidence-Based Medi- IEEE definition (IEEE 2000) and the reference to cine 2012), which is defined in Sacket et al. (1996) fundamental concepts in the ISO definition (ISO as: “the conscientious, explicit and judicious use 2011). of current best evidence in making decisions about The reference to properties that are necessary the care of individual patients.”. If you think that and sufficient to meet its essential requirements doctors would always base their diagnose on does indeed introduce a strong form of relativity sound evidence and reasoning, then Pfeffer and to architecture: Who/what determines what the Sutton 2011 invites us to rethink this. essential requirements are? We argue that these When considering the promise of evidence-based essential requirements follow from the key stake- management, there is indeed a strong analogy holders and their core concerns. What concerns to the potential contribution of enterprise ar- them most about the artefact? In the case of an en- chitecture. Some early examples of how enter- terprise, the essential requirements can be linked prise architecture can be used for evidence-based directly to the enterprise’s (past/current) strategy, management of enterprise transformation can be next to other core concerns of the key stakehold- found in, e.g., Op ’t Land (2006, 2007); Op ’t Land ers (Greefhorst and Proper 2011). As such, we and Dietz (2008). We indeed argue that enterprise argue that enterprise architecture should first architecture can become a leading mechanism in and foremost be about essential sensemaking in enabling evidence-based management of trans- that it should primarily: formations. Or rather, the field of enterprise 1: make sense of the past and future of the enter- architecture should take upon it as its mission prise with regards to the way it has/will meet its to enable evidence-based management of trans- essential requirements as put forward by its core formations. We explicitly use the word enable stakeholders and captured in its strategy, to stress the fact that it is senior management 2: provide clear motivations/rationalisation, in who have to take the responsibility for making de- terms of the above essential requirements, as cisions based on evidence. It remains their choice well as, e.g., constraints, of the trade-offs that un- not to take that responsibility, and explain to the derly the presence of the elements (e.g., building shareholders, tax payers and auditors, why they blocks or architecture principles) included in the did not. architecture. purpose mean- 4 Redefining Enterprise Architecture In line with this, we argue that the , ing and elements of an enterprise architecture Based on the future trends as identified in the should evolve: previous section, we will now revisit our under- 1: Its purpose is (i) to understand the current evol- standing of enterprise architecture. In line with ution of the enterprise, including its past and its the definition provided in Greefhorst and Proper likely future evolution and (ii) formulate, as well (2011) we regard architecture as essentially be- as motivation/rationalise, the desired future evol- ing about: “Those properties of an artefact that ution and the interventions needed to achieve are necessary and sufficient to meet its essential this. requirements”. This view is shared by Fehskens 2: Its meaning is that it expresses, in relation (2008), who defines architecture as “those prop- to the (current) essential requirements: (i) the erties of a thing and its environment that are ne- understanding how the enterprise has evolved cessary and sufficient for it to be fit for purpose so-far, (ii) what the expected natural evolution for its mission”. The focus on the properties that of the enterprise is and (iii) the desired future matter, is also what distinguishes architecture evolution of the enterprise and actions needed to from design. It also resonates well with the refer- change/strengthen its current evolution. ence to fundamental organization in the original 3: Its elements will focus on the fundamental Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Enterprise Architecture 17

properties that have played a role in its past evol- the different models, frameworks, modelling lan- ution, as well as its expected/desired future evol- guages, model transformations, and associated ution. These properties can be expressed from a modelling processes for enterprise architecture. constraining perspective in terms of architecture 3: A sociological perspective concerned with the principles and/or from a constructing perspective role of culture, skills, attitudes, communication, in terms of the building blocks of the enterprise. etc, needed/involved during the formulation of an enterprise architecture, as well as in the inter- It is important to note that during the evolution vention needed to establish the changes proposed of an enterprise, it is likely that the understanding of what the essential requirements are will by a future architectural direction. change. This means that the boundary between what was included in the architecture and what 6 Acknowledgements is considered design may also shift over time. For the modelling languages used (be it from a This work has been partially supported by the constructing or a constraining perspective), this Fonds National de la Recherche Luxembourg, via means that they should better take a broad per- the ASINE (Architecture-based Service Innovation spective focus on enterprise modelling in gen- in Network Enterprises), ACET (Architecturebased eral, where what is considered to be “architectur- Coordination of Enterprise Transformations) and ally relevant” may shift over time; modelling ap- RationalArchitecture projects. proaches with a narrow view of what is “proper” architecture may find themselves obsolete before they know it. References Achterberg R. v., Frankema B., Jong-Ellenbroek 5 Conclusion M. d., Molen P. v. d., Proper H., Schut In this position paper we discussed our view on W. (2000) Handleiding SysteemConcept en ApplicatieArchitectuur – Startarchitectuur. the history, and the potential future evolution, Technical Report Version 2.0. Dutch Taxation of the field of enterprise architecture. It is our Office. Last Access: In Dutch firm belief that enterprise architecture can, and Ambler S., Jeffries R. (2002) Agile Modeling: Ef- should, play a crucial role in enabling senior man- fective Practices for Extreme Programming agement of enterprises to take their responsib- and the Unified Process. John Wiley & Sons, ility in steering, controlling and/or guiding en- New York, New York terprise transformations, based on sensemaking Amdahl G., Blaauw G., Brooks F. (1964) Ar- and evidence-based insights. It is certainly one chitecture of the IBM System/360. In: IBM of the driving hypotheses in our work. Journal of Research and Development We suggest that future research into the enter- Beck K., Beedle M., Bennekum A. v., Cockburn prise architecture domain should do so from at A., Cunningham W., Fowler M., Grenning least three important vantage points, that are J. Highsmith J., Hunt A., Jeffries R., Kern J., also likely to need different types of research Marick B., Martin R., Mellor S., Schwaber K., methodologies: Sutherland J., Thomas D. (2001) Manifesto for 1: An engineering perspective that focuses on Agile Software Development. http://www. strategies, methods and techniques to provide agilemanifesto.org. Last Access: Accessed 14 evidence-based underpinning of the design de- June 2013. cisions underlying enterprise architectures (both Beijer P., De Klerk T. (2010) IT Architecture: in the constructing and the constraining sense). Essential Practice for IT Business Solutions. 2: A modelling perspective focussing the role of Lulu Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 18 Henderik A. Proper and Marc M. Lankhorst

Boar B. (1999) Constructing Blueprints for En- “Enterprise Architecture” Ought to Mean. In: terprise IT architectures. Wiley, New York, Open Group Conference Boston. The Open New York Group Buckl S., Matthes F., Schweda C. (2011) A Goedvolk J., Bruin H. d., Rijsenbrij D. (1999) Method Base for Enterprise Architecture Integrated Architectural Design of Business Management. In: Ralyté J., Mirbel I., De- and Information Systems. In: Bosch J. (ed.) neckere R. (eds.) Engineering Methods in Proceedings of the Second Nordic Work- the Service-Oriented Context: 4th IFIP WG shop on Software Architecture (NOSA’99). 8.1 Working Conference on Method Confer- Research Report 13 Vol. 1999. University of ence on Method Engineering, ME 2011, Paris, Karlskrona/Ronneby, Ronneby, Sweden France, April 20-22, 2011. Springer, Berlin, Graves T. (2008) Real Enterprise Architecture: Germany, pp. 34–48 beyond IT to the whole enterprise. Tetradian Ciborra C. (1992) From Thinking to Tinkering: Books, Colchester, England, United Kingdom The Grassroots of Strategic Information Sys- http://tetradianbooks.com tems. In: The Information Society 8, pp. 297– Greefhorst D., Proper H. (2011) Architecture 309 Principles – The Cornerstones of Enterprise Cockburn A. (2002) Agile Software Develop- Architecture. Enterprise Engineering Series. ment: The Cooperative Game, 2nd. Addison Springer, Berlin, Germany Wesley Hammer & Company (1986) PRISM: Dispersion Davenport T., Hammer M., Metsisto T. (1989) and Interconnection: Approaches to Distrib- How executives can shape their company’s uted Systems Architecture, Final Report. CSC information systems. In: Harvard Business Index, Inc.. Cambridge MA Review 67(2), pp. 130–134 Hammer M. (1990) Re–engineering work: don‘t De Caluwé L., Vermaak H. (2003) Learning to automate, obliterate. In: Harvard Business Change: A Guide for Organization Change Review 68(4), pp. 104–112 Agents. Sage publications, London, United Henderson J., Venkatraman N. (1993) Strategic Kingdom alignment: Leveraging information techno- Dietz J. (2006) Enterprise Ontology – Theory logy for transforming organizations. In: IBM and Methodology. Springer, Berlin, Germany Systems Journal 32(1), pp. 4–16 Dietz J. (2008) Architecture – Building strategy Hoogervorst J. (2009) Enterprise Governance into design. Netherlands Architecture Forum, and Enterprise Engineering. Springer, Berlin, Academic Service – SDU, The Hague, The Germany, Berlin, Germany Netherlands http://www.naf.nl Humble J. (2010) Continuous delivery: reliable Erl T. (2005) Service-Oriented Architecture software releases through build, test, and de- (SOA): Concepts, Technology, and Design.. ployment automation. Addison Wesley Prentice Hall Hurwitz J., Nugent A., Halper F., Kaufman M. ESPRIT Consortium AMICE (1993) CIMOSA: (2013) Big Data For Dummies. For Dummies. Open System Architecture for CIM, 2nd. John Wiley & Sons Inc. Springer, Heidelberg, Germany Iacob M.-E., Jonkers H., Lankhorst M., Proper H., Evidence-Based Medicine. http://en.wikipedia. Quartel D. (2012) ArchiMate 2.0 Specification. org/wiki/Evidence-based_medicine. Last Ac- The Open Group cess: Last visited 02-10-2012 IEEE (2000) Recommended Practice for Architec- Fehskens L. (2008) Re-Thinking architecture. In: tural Description of Software Intensive Sys- 20th Enterprise Architecture Practitioners tems. IEEE P1471:2000, ISO/IEC 42010:2007. Conference. The Open Group The Architecture Working Group of the Soft- Fehskens L. (2010) What the “Architecture” in ware Engineering Committee, Standards De- Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Enterprise Architecture 19

partment, IEEE. Piscataway, New Jersey Enterprise Engineering Series. Springer, Ber- ISO (1996a) Information technology – Open Dis- lin, Germany tributed Processing – Reference model: Ar- Lee G., Xia W. (2010) Toward Agile: An integ- chitecture ISO/IEC 10746–3:1996(E) rated analysis of quantative and qualitative ISO (1996b) Information technology – Open field data on software development agility. Distributed Processing – Reference model: In: MIS Quarterly 34(1), pp. 87–114 Foundations ISO/IEC 10746–2:1996(E) Nederlandse Overheid Referentie Architectuur ISO (1998a) Information technology – Open Dis- (NORA). http : / / www . e - overheid . nl / tributed Processing – Reference model: Archi- onderwerpen/e-overheid/architectuur/nora- tectural semantics ISO/IEC 10746–4:1998(E) familie/nora. Last Access: Last checked: 02- ISO (1998b) Information technology – Open Dis- 10-2012 tributed Processing – Reference model: Over- Olle T., Sol H., Verrijn–Stuart A. (eds.) Informa- view ISO/IEC 10746–1:1998(E) tion Systems Design Methodologies: A Com- ISO (2011) Systems and software engineering - parative Review. North–Holland/IFIP WG8.1, Architecture description is an international Amsterdam, The Netherlands standard for architecture descriptions of sys- Olle T., Sol H., Tully C. (eds.) Information Sys- tems and software.. ISO/IEC 42010. ISO tems Design Methodologies: A feature ana- Jonkers H., Veldhuijzen van Zanten G., Buuren lysis. North–Holland/IFIP WG8.1, Amster- R. v., Arbab F., De Boer F., Bonsangue dam, The Netherlands M., Bosma H., Ter Doest H., Groenewegen Op ’t Land M. (2006) Applying Architecture and L., Guillen Scholten J., Hoppenbrouwers S., Ontology to the Splitting and Allying of En- Iacob M.-E., Janssen W., Lankhorst M., Van terprises: Problem Definition and Research Leeuwen D., Proper H., Stam A., Torre L. Approach. In: Meersman R., Tari Z., Herrero v. d. (2003) Towards a Language for Coher- P. (eds.) On the Move to Meaningful Inter- ent Enterprise Architecture Descriptions. In: net Systems 2006: OTM 2006 Workshops - Steen M., Bryant B. (eds.) 7th IEEE Interna- OTM Confederated International Workshops tional Enterprise Distributed Object Com- and Posters, AWESOMe, CAMS, COMINF, puting Conference (EDOC 2003), Brisbane, IS, KSinBIT, MIOS-CIAO, MONET, OnToCon- Queensland, Australia. IEEE, Los Alamitos, tent, ORM, PerSys, OTM Acadamy Doctoral California, pp. 28–39 Consortium, RDDS, SWWS, and SebGIS, Pro- Lahrmann G., Winter R., Fischer M. (2010) ceedings, Part II, Montpellier, France. Lec- Design and Engineering for Situational Trans- ture Notes in Computer Science Vol. 4278. formation. In: Harmsen A., Proper H., Schalk- Springer, Berlin, Germany, pp. 1419–1428 wijk F., Barjis J., Overbeek S. (eds.) Pro- Op ’t Land M. (2007) Towards Evidence Based ceedings of the 2nd Working Conference on Splitting of Organizations. In: Ralyté J., Practice-driven Research on Enterprise Trans- Brinkkemper S., Henderson-Sellers B. (eds.) formation, PRET 2010, Delft, The Netherlands. Proceedings of the IFIP TC8 / WG8.1 Working Lecture Notes in Business Information Pro- Conference on Situational Method Engineer- cessing Vol. 69. Springer, Berlin, Germany, ing: Fundamentals and Experiences (ME07), pp. 1–16 Geneva, Switzerland. IFIP Series Vol. 244. Lankhorst (ed.) Agile Service Development: Springer, Berlin, Germany, pp. 328–342 Combining Adaptive Methods and Flex- Op ’t Land M., Dietz J. (2008) Enterprise onto- ible Solutions. Enterprise Engineering Series. logy based splitting and contracting of or- Springer, Berlin, Germany ganizations. In: Proceedings of the 23rd An- Lankhorst (ed.) Enterprise Architecture at Work: nual ACM Symposium on Applied Comput- Modelling, Communication and Analysis, 3rd. ing (SAC’08), Fortaleza, Ceará, Brazil Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 20 Henderik A. Proper and Marc M. Lankhorst

Op ’t Land M., Proper H., Waage M., Cloo J., Scheer A.-W. (2000) ARIS – Business Process Steghuis C. (2008) Enterprise Architecture – Modeling. Springer, Berlin, Germany Creating Value by Informed Governance. En- Sowa J., Zachman J. (1992) Extending and form- terprise Engineering Series. Springer, Berlin, alizing the framework for information sys- Germany tems architecture. In: IBM Systems Journal Pfeffer J., Sutton R. (2006) Evidence-Based Man- 31(3), pp. 590–616 agement. In: Harvard Business Review Last Spewak S. (1993) Enterprise Architecture Plan- visited: 02-10-2012 http://hbr.org/2006/01/ ning: Developing a Blueprint for Data, Ap- evidence-based-management/ar/1 plications, and Technology. Wiley, New York, Pfeffer J., Sutton R. (2011) Trust the Evidence, New York Not Your Instincts. In: New York Times New TAFIM (1996) Department of Defence Tech- York edition, September 4th, BU8 http://www. nical Architecture Framework for Informa- nytimes.com/2011/09/04/jobs/04pre.html? tion Management – Overview. Defence In- _r=2&ref=business& formation Systems Agency Center for Stand- Proper H. (2012) Enterprise Architecture – ards, United States of America Growing up to evidence-based management? Tamura T., Yonemitsu S., Itoh A., Oikawa D., In: Baldinger F., Rijsenbrij D. (eds.) Nether- Kawakami A., Higashi Y., Fujimooto T., Na- lands Architecture Forum, The Netherlands kajima K. (2004) Is an entertainment robot chap. 8.1, pp. 317–328 useful in the care of elderly people with Richardson G., Jackson B., Dickson G. (1990) severe dementia? In: J Gerontol A Biol Sci A Principles-Based Enterprise Architecture: Med Sci 59(1), pp. 83–85 Lessons from Texaco and Star Enterprise. In: Tapscott D., Caston A. (1993) Paradigm Shift – MIS Quarterly 14(4), pp. 385–403 http://www. The New Promise of Information Technology. jstor.org/stable/249787 McGraw–Hill, New York, New York Rivera R. (2007) Am I Doing Architecture or Taylor J., Van Every E. (2010) The Situated Or- Design Work? In: It Professional 9(6), pp. 46– ganization: Case studies in the pragmatics of 48 communication research. Routledge Ross J., Weill P., Robertson D. (2006) Enterprise Teece D., Pisano G., Shuen A. (1997) Dynamic architecture as strategy: creating a founda- Capabilities and Strategic Management. In: tion for business execution. Harvard Business Strategic Management Journal 18(7), pp. 509– School Press, Boston, Massachusetts 533 Sacket D., Rosenberg W., Gray J., Haynes R., The Open Group (2005) TOGAF – The Open Richardson W. (1996) Evidence based medi- Group Architectural Framework – Version 8.1 cine: what it is and what it isn’t. In: Brit- Enterprise Edition. The Open Group – Van ish Medical Journal 312, pp. 71–72 http : Haren Publishing, Zaltbommel, The Nether- / / www. ncbi . nlm . nih . gov / pmc / articles / lands PMC2349778/pdf/bmj00524-0009.pdf The Open Group (2009) TOGAF Version 9. Van Scheer A.-W. (1986) Neue Architektur für Haren Publishing, Zaltbommel, The Nether- EDV-Systeme zur Produktionsplanung lands und -steuerung. In German. Institut für Treacy M., Wiersema F. (1997) The Discipline Wirtschaftsinformatik im Institut für of Market Leaders – Choose your customers, Empirische Wirtschaftsforschung an der narrow your focus, dominate your market. Universität des Saarlandes, Saarbrücken, Addison Wesley, Reading, Massachusetts Germany W3C Semantic Web Activity (2013). http://www. Scheer A.-W. (1988) Computer integrated manu- w3.org/2001/sw/. Last Access: Last visited 14- facturing : CIM. Springer, Berlin, Germany 06-2013 Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Enterprise Architecture 21

Wagter R. (2009) Sturen op samenhang op basis Marc M. Lankhorst van GEA – Permanent en event driven. In BiZZdesign Dutch. Van Haren Publishing, Zaltbommel, P.O. Box 321 The Netherlands 7500 AH Enschede Wagter R., Berg M. v. d., Luijpers J., Steenber- The Netherlands gen M. v. (2001) DYA: snelheid en samen- [email protected] hang in business en ICT architectuur. Tutein Nolthenius Wagter R., Berg M. v. d., Luijpers J., Steenbergen M. v. (2005) Dynamic Enterprise Architecture: How to Make It Work. Wiley, New York, New York Wagter R., Proper H., Witte D. (2011) Enter- prise Coherence Assessment. In: Harmsen A., Grahlmann K., Proper H. (eds.) Proceedings of the 2rd Working Conference on Practice- driven Research on Enterprise Transform- ation, PRET 2011, Luxembourg-Kirchberg, Luxembourg. Lecture Notes in Business In- formation Processing Vol. 89. Springer, Berlin, Germany, pp. 28–52 Wood–Harper A., Antill L., Avison D. (1985) In- formation Systems Definition: The Multiview Approach. Blackwell, Oxford, United King- dom Wout J. v., Waage M., Hartman H., Stahlecker M, Hofman A. (2010) The Integrated Architec- ture Framework Explained. Springer, Berlin, Germany Zachman J. (1987) A framework for information systems architecture. In: IBM Systems Journal 26(3)

Henderik A. Proper Public Research Centre Henri Tudor 29, avenue J.F. Kennedy L-1855 Luxembourg Luxembourg and Radboud University Nijmegen Faculty of Science Postbus 9010, 6500GL Nijmegen The Netherlands [email protected] Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 22 Ulrich Frank

Ulrich Frank

Enterprise Modelling: The Next Steps

Enterprise modelling is at the core of Information Systems and has been a subject of intensive research for about two decades. While the current state of the art shows signs of modest maturity, research is still facing substantial challenges. On the one hand, they relate to shortcomings of our current knowledge. On the other hand, they are related to opportunities of enterprise modelling that have not been sufficiently addressed so far. This paper presents a personal view of future research on enterprise modelling. It includes requests for solving underestimated problems and proposes additional topics that promise to promote enterprise models as more versatile tools for collaborative problem solving. In addition to that, the paper presents requests for (re-) organising research on enterprise modelling in order to increase the impact of the field.

1 Introduction at least one conceptual model of an organisational action system with at least one conceptual It has been a wide-spread conviction for long model of a corresponding information system. that the complexity of large information systems Usually, but not necessarily, the various mod- recommends the use of models. Information sys- els that constitute an enterprise model are cre- tems are aimed at representing domains through ated with domain-specific modelling languages data that is accessible by prospective users. Rep- (DSML). To emphasise that enterprise models are resenting a–factual or aspired–domain cannot be intended to provide a medium both for fostering accomplished by modelling it directly. Instead, analysis and design tasks and for communication it comprises a twofold abstraction: We perceive across traditional professional barriers, the term a domain primarily through language, which in “multi-perspective enterprise model” has been turn reflects an abstraction over “objective” fea- introduced (Frank 1994). A multi-perspective en- tures of a domain. At the same time, using an terprise model is an enterprise model that em- information system requires an interface that cor- phasises accounting for multiple perspectives. responds to the language spoken in the targeted A perspective represents a specific professional domain. Therefore, the construction of informa- background that corresponds to cognitive dis- tion systems recommends the design of concep- positions, technical languages, specific goals and tual models. They do not only promise to reduce capabilities of prospective users (Frank 2013b). complexity by abstracting from ever changing peculiarities of technical infrastructures; they also In recent years, the term “enterprise architec- allow for getting prospective users involved in ture” has gained remarkable attention (Buckl et the analysis and design process. Exploiting the al. 2010; Land et al. 2009; Lankhorst 2005). The potential of information systems will often re- differences between enterprise model and enter- quire reorganising existing patterns of action— prise architecture are mainly related to the inten- sometimes in a radical way. Therefore, analysis ded audience. Enterprise modelling is aimed at and design of information systems should usually various groups of stakeholders that are involved be done conjointly with analysing and designing in planning, implementing, using and maintain- the organisational action system. The conception ing information systems. Therefore, enterprise of an enterprise model was developed to address models are supposed to offer a variety of cor- this demand. An enterprise model integrates responding abstractions. These include models Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Enterprise Modelling: The Next Steps 23

that serve as a foundation of software develop- At first, it represents a renunciation of the ori- ment. Therefore, the development of respective ginal approach to enterprise modelling, which DSML is a particular characteristic of enterprise was aimed at developing information systems modelling. Different from that, enterprise archi- from scratch. With respect to the complexity of tecture mainly targets IT management. There- today’s IT infrastructures and the fact that most fore, it puts less emphasis on the specification organisations will not develop substantial parts of DSML. Nevertheless, there is no fundamental of their information system on their own any- difference between both approaches. Instead, the more, this additional focus is certainly reasonable. abstractions used in enterprise architectures can Secondly, it is not only aimed at supporting the be seen as an integral part of more comprehens- design of IT infrastructures that are in line with ive enterprise models. In Information Systems, the corporate action system, but also at providing enterprise modelling has been a pivotal field of an instrument for IT management. Figure1 illus- research that has evolved over a period of more trates the representation of an enterprise model than 20 years (CIMOSA: Open system architec- through a set of interrelated diagrams that cor- ture for CIM 1993; Ferstl and Sinz 1998; Group respond to the current state of the art. To ensure 2009; Scheer 1992; Zachman 1987). It has produced integration, the partial models that are repres- various modelling frameworks, DSML as well as ented by the diagrams should be created with corresponding tools. The field has achieved a modelling languages that were specified with the remarkable degree of maturity which is indicated same meta modelling language and that share by the fact that enterprise modelling is part of common concepts (Frank 2011). many IS curricula—even though to different ex- While the abstractions covered by today’s en- tent. Nevertheless, there is still need for further terprise modelling methods arguably represent research to exploit the potential of enterprise relevant perspectives on an enterprise, a compre- modelling. In this paper I will point at relevant hensive representation of all aspects that may be shortcomings of the current state of the art in relevant for analysing, designing and managing order identify core elements of a future research a company together with its information sys- agenda. tem requires accounting for more context. While such a demand may look like an exaggeration to 2 The Need for More Context some, it is actually the consequent continuation of current practice: All professional activities in At the beginning, approaches to enterprise mod- a company are characterised by the use of con- elling were mainly focussed on developing high- ceptual abstractions, i.e., by a specific technical level frameworks to provide a common structure terminology and corresponding language games. or architecture of an enterprise and its informa- Reconstructing these terminologies through ad- tion system (CIMOSA: Open system architecture ditional DSML would not only enable further use for CIM 1993; Scheer 1992; Zachman 1987). Apart scenarios, it would also enrich existing models from using general purpose modelling languages with additional context. Context does not only (GPML) like the ERM and the UML, the devel- refer to the topics that are represented in an en- opment of DSML was mainly aimed at business terprise model. It also refers to the context in process modelling. Later, DSML were created for which the development and use of models occur. modelling strategies, organisational structures On the one hand, this kind of context includes or generic resources. In recent years, some ap- specific methods for enterprise modelling. On proaches have evolved that are aimed at DSML the other hand, it refers to organisational and for modelling IT infrastructures and IT architec- managerial arrangements to foster an adequate tures. This focus is remarkable for two reasons. handling of enterprise models. Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 24 Ulrich Frank

Goal System Diagram Sustainable sales Labor costs 1 3 Value Chain Diagram Increase Keep sustainabl labor e sales costs Inbound Outbound Marketing stable Services Logistics Operations Logistics Sales Customer Number of satisfaction sales agents 2 4 Sustainable Sustainable sales sales 1 1 Increase Increase customer number of satisfactio sales n Increase Increase agents sust. sales sust. in region sales in A region B Organizational Chart

Business Process Map Board of Directors

Service> Marketing Finance Accounting Sales

Order Management Technical Support Complaint Product A Handling Product Asset Market Private A Manageme Systems Research Customers nt

Product Corporate Corporate Order Management Complaint Controlling Planning Credits Customers Product B Handling Product B Customer Akquisition Service> Promotion Relations Support Forecast

Order Complaint Management Handling Product Product C C

< Agent >

amount not Inform Customer Order < Agent > < Agent > available declined

Order Check Delivery Delivery Confirm Order < Fleet Manager > recieved Capacity possible Business Process Diagram amount Check Logistics < Agent > available

Delivery at Decline Order IT Resource Diagram desired date not possible

Object Model

Scan Order Invoice SAP CO SAP FI SAP R&D PLM 3.7 Customer Engine 5.1 created: Date issued: Date P10 6.0 P10 6.0 P20 custID: String issues 0,* no: String s total() : Decimal

1,1 0,* 1,1 e d

sum() : Decimal s u e

l 1,1

c Document 0,1 0,1 acts as total() : Decimal d n n

acts as u n i DMS l i o

c i

0,1 s t

Archive n t i

1.9 l a

r 0,1

1,* u c ABAP x Person 1,* e i 2.0 s l v Organisation e

p OrderItem InvoiceItem r lastName: String r p e

S name: String A units: Integer units: Integer firstName: String 0,* 0,1 P MW: SAP MW: SAP sum() : Decimal sum() : Decimal A born: Date

S BL: SAP Netweaver DB: t ERC ECC o 0,*

a sex: Gender t

s MDM PLM ADDS d r

e age() : Integer

e AccountEntry t f a t e c a 0,* r

created: Date o l

Address d 1,1 r e

r amount: Decimal 0,* t

e OS

e Product OS street: String a t v c s r o u

UNIX HP UX city: String l name: String e l UNIX HP

S 1,1

C 0,*

country: String id: String P

P UX 1,1 H

A 1,1 inStock: Integer Account S x i salesPrice: Money n Dell Power id: String Sun SPARC U retailPrice: Money Edge T110 balance() : Decimal

Figure 1: Diagrams Representing Example Enterprise Model Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Enterprise Modelling: The Next Steps 25

2.1 Further Modelling Topics of an enterprise and are usually not subject of design processes. Nevertheless, markets are of The variety of topics that are handled in enter- crucial importance for successful action in an en- prises is enormous. Among those that have not terprise. Furthermore, markets are getting more been sufficiently addressed in enterprise mod- complex and contingent: Often, they expand on elling are products, production processes, pro- an international scale and may be very dynamic jects, markets and logistic. While product mod- in the sense that products are displaced by in- elling is an issue on its own, integrating product novations or that customer preferences change models with enterprise models makes sense for quickly. Therefore, integrating models of mar- various reasons. Products can be very complex kets with enterprise models promises to gain a and may demand for quick adaptations. At the more differentiated understanding of relevant same time, developing, producing and handling market forces and to develop a better founda- products relates to various aspects of an enter- tion for decision making. Similar to production prise that are usually part of an enterprise model: processes, logistic networks have been subject of goals, business processes, organisational units or optimisation efforts for long. The respective mod- software systems. More and more, products com- els, often designed to satisfy the requirements prise software or are constituted by software. In of Operations Research methods, are mainly fo- addition to that products are often bundled—with cussed on optimisation with respect to certain services and/or other products. Therefore, integ- goals. Integrating the respective modelling con- rating product models with enterprise models cepts with languages for enterprise modelling would enable additional analysis scenarios such would enable to enrich both enterprise models as checking the effect of changing a product on and logistic models with more relevant context. required skills and on business processes. There are numerous approaches to model production In addition to traditional topics, organisations are processes. They aim at developing algorithms confronted with new phenomena that may de- and approximation procedures for production mand for appropriate action. They include social planning, process scheduling and process control. networks, virtual enterprises, nomad employees Integrating respective models with enterprise and many more. Extending enterprise models models will often be not trivial, because they with models of these phenomena would foster are based on different modelling paradigms. At analysing and handling them. This would, how- the same time, including elaborate models of pro- ever, require new modelling concepts. Finally, duction processes in enterprise models promises enterprise models can be supplemented with con- various advantages, such as supporting the concepts that are related to important further aspects joint analysis of production processes and related of managerial decision making. These include accounting concepts, e.g., specialised cost and business processes or generating software for benefit concepts, concepts to design and analyse controlling production processes from respective indicator systems (Strecker et al. 2012) as well models. In an increasing number of organisa- as concepts for modelling organisational know- tions, projects play a key role. Integrating project ledge. Adding these concepts would make enter- models with enterprise models would support prise models and corresponding tools a versatile project management by providing meaningful instrument for management—both on the oper- links to organisational resources. Also, project ational and strategic level. At the same time, it modelling could benefit from existing approaches could serve as a valuable extension of enterprise to business process modelling and would allow software systems (see Sect. 4.1). to take advantage of similarities between projects. Markets have not been part of enterprise Request: To make enterprise models a versatile models for an apparent reason: They are outside tool for supporting professional action in organ- Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 26 Ulrich Frank

isations, research needs to widen the scope of on the instantiation of methods from metamod- modelling by adding further topics that also com- els, while further approaches are based on a prise concepts to support managerial decision combination of composition and instantiation. making. It seems that the field has reached a stage of mod- erate maturity, which is also indicated by the spe- 2.2 Method Construction cification of a respective ISO standard (ISO/IEC 2007). Modelling languages are an important founda- Nevertheless, there are some aspects that have tion of enterprise modelling, since they provide been widely neglected so far. At first, current a purposeful structuring of a domain. However, approaches to method engineering are mostly they are not sufficient for designing and using en- generic in the sense that they are not restricted terprise models. In addition to languages, there to particular domains, nor do they account for the is need for processes that guide the purposeful peculiarities of enterprise models. That leaves development, interpretation and use of respect- prospective developers and users of enterprise ive models. In other words: There is need for models with the demanding task of adapting gen- modelling methods. Due to the diversity of pro- eric concepts to the idiosyncrasies of particular jects that can benefit from conceptual models, organisations. Second, current approaches to it is evident that a given set of modelling meth- method engineering focus on the design of pro- ods cannot not fit all demands—except for the cess models and take the modelling language as price of oversimplification. This insight shifted given. However, the diversity of topics that can the focus on approaches that guide the develop- be reasonable subjects of enterprise models may ment of customised methods. The only chance also require to adapt or even create modelling to provide support for the conceptualisation of languages. While a number of tools support the a range of methods is to increase the level of specification of DSML and the realisation of cor- abstraction by searching for essential character- responding model editors, prospective users can istics shared by all modelling methods. Against expect only little guidance with designing a lan- this background, the emergence of method en- guage that fits its purpose. At the same time gineering as a new field of research is a reas- the design of DSML is especially demanding. Of- onable consequence in two respects: First, it ten, prospective users will not have an idea of is aimed at rich abstractions that cover a wide what they might expect from a DSML. As a con- range of modelling projects. Second, it makes sequence, requirements analysis is a remarkable use of the same paradigm that it suggests for challenge. In addition to that, design conflicts the field of conceptual modelling, too: The con- need to be handled. Also, the creation of the construction of particular methods should follow crete syntax requires a specific competence that an engineering approach, which—among other many language designers do not have. There- things—recommends accounting for linguistic fore, there is need for substantiated guidance to rigour, consistency and coherence as well as for reduce the risk of poorly designed modelling lan- the development of supportive tools. During guages. Currently, there are only few approaches the last 15 years, a plethora of method engineer- that offer respective support (Frank 2013a; Moody ing approaches—originating mostly in Require- 2009). Finally, method engineering is often based ments Engineering and Software Engineering— on two assumptions: First, a method is an arte- has evolved (Brinkkemper 1996; Ralyté et al. 2005, fact that is created through an engineering act. 2007). For an overview see Henderson-Sellers and Second, applying the method appropriately is Ralyté (2010). Some emphasise the construction pivotal for successful action. However, with re- of methods from reusable elements, others focus spect to successfully using a method, it is not Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Enterprise Modelling: The Next Steps 27

sufficient to restrict it to its explicit definition, as linguistic artefacts and additional organisa- i.e., to take a mere technical perspective. This tional/managerial measures that promote the apis for two interrelated reasons. First, a method propriate use of methods in practice. will usually not be based on a pure formal specification. Instead, its conceptual and theoretical 3 The Need for More (Re) Use foundation as well as the process description re- The remarkable effort that is required to build quire interpretations that produce some degree elaborate enterprise models makes reuse of mod- of shared understanding. Second, for a method els and modelling concepts a pivotal issue for to work, it has to become an accepted orientation achieving higher productivity. At the same time, for individual and collaborative action. To sum- reuse can also contribute to model quality, if re- marise both aspects: A method needs to make usable artefacts are designed and evaluated with sense. From this point of view, a method can specific care. In addition to that reusable con- be regarded as a social construction that reflects cepts can serve to foster integration of those com- established patterns of professional action, ideas ponents that share them. Approaches to promote of professional values and aesthetics, organisa- reuse have been on the research agenda for long. tional culture, common beliefs as well as indi- The idea of reference enterprise models seems vidual interests. Against this background, we to be especially attractive. However, so far, re- can distinguish between a method as a linguistic use of enterprise modelling artefacts remained artefact, stressing a technical view, and a method on a modest level (Fettke and Loos 2007). There as an actual practice, stressing a more pragmatic are various reasons that contributed to this un- or organisational view. Therefore I intention- satisfactory situation. Two especially important ally avoid the term “method engineering” and reasons are related to modelling languages. On speak of “method construction” instead. This the one hand, current languages for enterprise is to express that a method is also constructed modelling lack concepts that enable reuse. On by those who use it, because it is shaped by ac- the other hand, the design of reusable DSML is tual interpretations and actions. A method as an facing a substantial challenge. artefact could be regarded as input or stimulus to trigger such a process. While for analytical 3.1 The Lack of Abstraction in Process reasons it may be useful to focus on methods Modelling mainly as linguistic artefacts, such a restricted view is certainly not sufficient with respect to a Taken the fact that business process modelling pragmatic objective such as improving efficiency has been a research subject for long, it seems sur- and quality of collaborative problem solving in prising that respective modelling languages are organisations. The benefit of methods for en- rather primitive in the sense that they do not al- terprise modelling will not only depend on the low for powerful abstractions. As a consequence, qualification of the involved stakeholders, but reuse of business process models remains on a also on certain aspects of the respective corpor- dramatically poor level. Since business process ate culture. It makes a clear difference, whether models play a pivotal role within enterprise mod- conceptual models are regarded as corporate as- els, this is a serious shortcoming. The follow- sets or as cost drivers with dubious outcome. ing scenario illustrates the problem. A company comprises a few tens of business process types Request: To promote the beneficial development including a core order management process type. and use of enterprise models it is required to A process type includes activity types. Various support the construction of respective modelling process types share similar activity types. Now methods that account for both, the conceptual two more specific order management process foundation of designing/customising methods types need to be implemented. For this purpose, Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 28 Ulrich Frank

standard order management

no chance for contract

calculate prepare contract term&conditions contract prepared order received terms&conditions satisfactory

order management for under-specified order

no chance for contract

determine options calculate prepare contract request options terms&conditions term&conditions contract prepared received determined satisfactory

Figure 2: Example of extending a business process type

it would be most helpful to specialise the existing ways effect the original control flow (Frank 2012). order management process type. The example in Fig.2 illustrates this problem.

This would not only allow reusing the respective There are a few approaches in Software Engineer- model and corresponding software implementa- ing and process modelling which are aimed at a tions, it would also promote safe and convenient relaxed conception of specialisation of behaviour maintenance: Future changes of the core process (Schrefl and Stumptner 2002) or of “workflow type would be immediately effective in the spe- inheritance” (Aalst and Basten 2002). Other ap- cialised types, too. To satisfy the demand for proaches focus on analysing structural similarit- integrity, a respective concept of process special- ies of control flows to promote reuse through pro- isation would have to satisfy the substitutabil- cess variants (Koschmider and Oberweis 2007). ity constraint: Every instance of a process type However, the restrictions these approaches imply can act as an instance of the corresponding su- remain unsatisfactory (Frank 2012). At the same per process type without causing harm (Liskov time, the still growing relevance of efficiently and Wing 1994). The substitutability constraint creating and maintaining business process mod- is satisfied, if the extensions defined for special- els demands for abstractions that allow taking ised concepts are monotonic. This can be accom- advantage of similarities. plished fairly easy for static abstractions. How- Request: Future research should aim at concepts ever, for dynamic abstractions such as business of relaxed process specialisation—which may be process models adding further activity or event combined with instantiation—that promote reuse types cannot be monotonic, because it will al- without unacceptable restrictions. Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Enterprise Modelling: The Next Steps 29

While the lack of a sound concept of process spe- same time, DSML foster model integrity, because cialisation creates a serious problem, the current they prevent the creation of inconsistent mod- state of business process modelling is even more els to a certain extent. By featuring a domain- dreary. The above scenario would suggest to re- specific concrete syntax, they also promote model use an activity type that was defined already for comprehensibility. Against this background, it a certain business process type in a new process does not come as a surprise that DSML are re- type. However, this is not possible: Every busi- garded by many as the silver bullet of conceptual ness process type has to be designed from scratch modelling and model-driven software develop- using the basic concepts provided by today’s pro- ment. However, their construction is facing a cess modelling languages. Hence, an activity dilemma. The more a DSML is tuned to a spe- such as “prepare contract” cannot be specified cific domain, the better is its contribution to pro- as a reusable type. Instead it is yet another in- ductivity and integrity. However, the more spe- stance of a basic (meta) type like “activity” or cific a DSML is, the more unlikely it can be used “automated activity” that is distinguished from in a wide range of particular domains. Figure3 other primarily through its label. There are ap- illustrates the conflicting effects of semantics on proaches that focus on analysing labels in order range of reuse and productivity. to detect similar activity types (Dijkman et al. e n

2011). However, their contribution to reuse is i s a u G e

y limited: Instead of removing the mess, they try R t

i f v o i

t e coping with it. c l u a d c o S r

Request P : There is need for extending business l a i t process modelling languages and tools with the n e t o possibility to define and reuse activity and event P types. This request is not easy to satisfy. An activity type is not only defined by its internal structure and behaviour, but also by its context such as Level of (domain-specific) Semantics the event that triggers it or the events it pro- duces. Reuse will be possible only, if the context Figure 3: DSML: Illustration of Essential Design Conflict can be adapted to some extent. Therefore, the required concept of an activity type—and of an Some authors suggest to design DSML to the event type respectively—must abstract from the needs of particular organisations or even pro- context without compromising reusability too jects only (Kelly and Tolvanen 2008; Völter 2013). much. This recommendation is based on two assumptions. First, the variety of organisations would 3.2 The Essential Conflict of Designing not allow for powerful DSML that fit all indi- DSML vidual requirements. Second, there is no need for further reuse, because creating and using a DSML are characterised by convincing advant- DSML in one particular project will usually pay ages (Kelly and Tolvanen 2008; Kleppe 2009; Völ- off already. Even though both assumptions may ter 2013). By providing domain-specific concepts, be valid to a certain extent, they are hardly con- they promote modelling (and programming) pro- vincing. There may be remarkable differences ductivity: Modellers are not forced anymore to between organisational actions systems and cor- reconstruct domain-level concepts from generic responding terminologies. However, it would be concepts such as “entity” or “attribute”. At the a sign of epistemological defeatism to deny the Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 30 Ulrich Frank

chance of finding substantial commonalities. Fur- to create models that include concepts on the M0 thermore, it can be assumed that actual variety level. is also a result of in part arbitrary processes of organisational evolution, i.e., it is not a reflection Universal DSML Organisational Unit M2 of inevitable differences. Also, there is evidence Committee Position that technical languages work in a wide range Language Designer of organisations of a certain kind: The terminology used in textbooks will often fit an entire Specific DSML (Local „Dialect“) Quality Circle industry in the sense that it provides a respected Department M1 linguistic structure and serves as as common ref- Team Market Analyst Organisation Analyst erence for professionals. Nevertheless, there are organisation-specific adaptations of textbook ter- Particular Organisation Quality Circle minology. They include extensions, refinements Model Marketing Department Product Group PG 1 M0 and modifications, some of which may be ques- Market Research Market tionable. The argument that a DSML will already Manager Team Analyst MA2 pay off in single use scenarios is fine, but it could Figure 4: Illustration of multi-level modelling languages still be much more profitable, since a wider range of reuse would allow for much better economies of scale. Therefore, it would be beneficial to Designing such language systems and corres- create hierarchies of DSML, where more specific ponding tools is far from trivial. It requires giv- ones are extensions and/or instantiations of more ing up prevalent architectures of modelling lan- general DSML. guages that feature a given set of classification layers (for respective approaches see Atkinson et Request : Research on DSML should aim at hier- al. (2009); Clark et al. (2008); Simonyi et al. (2006). archies of languages to enable both a wide range Instead, recursive language models such as the of reuse and customised languages for narrow “golden braid” architecture are more promising— domains. and more demanding at the same time, because Figure4 illustrates the idea of providing mod- they are not supported by most of today’s devel- elling languages on different classification lay- opment environments. Apart from that, designers. The highest level (“universal DSML”) corres- ing languages for enterprise modelling should ponds to textbook terminology. The concepts on account for a further issue. Current DSML are this level should be applicable to a wide range usually specified with metamodels. This is for of organisations, hence, promote economies of a good reason: On the one hand, this kind of scale. The universal DSML should be designed specification corresponds to a paradigm the mod- by experts that possess deep knowledge about elling community is familiar with. On the other the general domain as well as rich experience hand, it fosters the construction of corresponding with designing DSML. “Local” DSML represent tools, because a metamodel can be used as a con- more specific technical languages for organisa- ceptual foundation of a respective modelling tool. tion modelling that apply to a few organisations The semantics of DSML, e.g., the semantics of or to one only. They are designed by organisa- specialisation concepts, is usually based on the se- tion analysts that are familiar with the respective mantics of prevalent programming languages to domain. These local DSML that feature a graph- facilitate the transformation of models into code. ical notation much like the universal DSML can However, there are other language paradigms be used by authorised managers to specify par- and specification styles that might enrich enter- ticular organisational settings. The example also prise models. For instance, models designed with shows that there are cases where it makes sense logic-based languages allow for deduction could Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Enterprise Modelling: The Next Steps 31

enable more sophisticated approaches to analys- he could browse a graphical representation of the ing enterprise models. Since the semantics of corporate business process map, which shows all respective languages, which are typically found business process types, their interrelationships in Artificial Intelligence, is different from that and key performance indicators at a glance. He of DSML used for enterprise modelling today, could then select a business process type he is integrating them into enterprise modelling envir- interested in, study the model that describes its onments is a demanding task. execution and demand for further aggregate data that characterises it, such as the number of in- 4 The Need for Run-Time Use stances per month, average revenues etc. Also, he Originally, enterprise models like most other con- could select specific analysis views, e.g., a view ceptual models were intended for supporting the that associates a selected business process type creation of information systems only. However, with the IT resources it requires. If he was inter- it is obvious that they should be beneficial during ested in one particular business process type, he the entire life cycle of an information system. On could view the corresponding model. Then, he a more generic level, this issue is addressed by could leave the conceptual level and ask for the research on “models at runtime” (Blair et al. 2009). list of currently active business processes of this Multi-perspective enterprise models provide ab- type and inspect the state of the instances he is stractions of the enterprise that support decision interested in. In addition to that, advanced users making and other managerial tasks. Also, they could modify the enterprise system by changing can help people in organisations to develop a the enterprise model. The DSML, an enterprise deeper understanding of the action system, i.e., model is created with would help preventing ar- how their work is integrated into a bigger pic- bitrary modifications and hence contribute to sys- ture. In addition to that, enterprise models can tem integrity. An outline of a respective system, enable users to develop a better understanding referred to as “self-referential enterprise system” of the information system and its interplay with is presented in (Frank and Strecker 2009). Figure5 organisational patterns of action. illustrates the idea of integrating enterprise systems with enterprise modelling environments. 4.1 Integrating Enterprise Models with Such a system would allow realising the vision of Enterprise Systems interactive models propagated by Krogstie (2007, In an ideal case, an enterprise modelling environ- p. 306): “The use of interactive models is about ment would be integrated and synchronised with discovering, externalising, capturing, expressing, a corresponding enterprise (software) system. representing, sharing and managing enterprise On the one hand, this would enrich enterprise knowledge.” In other words: It would be a con- systems not only with their conceptual founda- tribution to empowering people who work in tion, but also with a representation of the context and interact with organisations. The realisation they are supposed to operate in. On the other of self-referential enterprise systems does not hand, enterprise models would be supplemented only require developing further DSML, but also with corresponding instance populations. redesigning enterprise software systems. This would enable users to navigate from con- Request: To further exploit the potential of both cepts on various classification levels to instances— enterprise software systems and enterprise mod- et vice versa. The following scenario illustrates elling environments, research should aim at de- the benefit of drilling down from an enterprise veloping the foundations for integrating both model to instances. A department manager who kind of systems into a versatile tool for managing is new to a firm wants to get a better understand- and adapting an organisation and its information ing of the way business is done. For this purpose, system. Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 32 Ulrich Frank

Is there a chance for outsourcing services?

Figure 5: Navigating an enterprise model and corresponding instances

4.2 Deficiencies of Prevalent classes in the enterprise system, from objects Programming Languages in the enterprise modelling environment. As a consequence, one would have to deal with the The integration of enterprise modelling envir- notorious problem of synchronising models and onments and enterprise systems does not only code. Figure6 illustrates how the M0 layer of require research on enterprise models and their representation. It also demands for system ar- modelling tools is overloaded and that concepts chitectures that cannot be satisfactorily accom- in modelling tools are located on a classification plished with prevalent programming languages. layer that is different from that of corresponding Integration implies common representations of concepts in an associated enterprise information shared concepts. In today’s modelling environ- system. ments, conceptual models are usually represented by objects on the M0 level—even though they Recent developments in research on program- belong to the M1 or even a higher level. Over- ming language has produced (meta) program- loading the M0 level happens for a good reason: ming languages that were designed for creating Prevalent programming languages are restricted domain-specific programming languages. Lan- to the dichotomy of objects and classes. Hence, guages like XMF (Clark and Willans 2012; Clark there are no meta classes that were required to et al. 2008) are especially promising, since they specify classes—and that would allow treating allow for an arbitrary number of classification classes as objects, too. Therefore, a common levels, which enables a common representation representation of classes in both systems is not of models and respective code. Hence, modify- possible. Instead, the only way to associate a ing an enterprise model implies modifying the modelling environment with a corresponding en- respective part of the enterprise software simul- terprise system would be to generate code, i.e., taneously. Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Enterprise Modelling: The Next Steps 33

Enterprise Modeling Environment

MetaEntity Metamodel Editor name: String ... concepts of metamodelling language M1 M3 Position name: String averageSalary: Money concepts of metamodel M0 M2 availability: Level ...

Model Editor(s) Position name: String averageSalary: Money Enterprise Information System concepts of modelling language M1 M2 availability: Level ... Position name: String Programmer concepts of model M0 M1 Implementation Layers Programmer averageSalary: Money id: String availability: Level staffed: Boolean ...... M1 schema (classes, types) average: € 48.300 availability: #low pos3: Position p1: Programmer M0 runtime (objects, data) name = 'Programmer' staffed = true averageSalary = 48.300 ... corresponds to availability = #low ...

Figure 6: Mismatch of Classification Levels

Request: To advance the state of current model- characteristic for research to abstract from single ling environments, research needs to focus on cases and aim at constructions that work for an tools that overcome the limitations of current entire class of cases. Furthermore, applied re- programming languages. search is motivated by improving existing practice with respect to certain goals. Unfortunately, 5 The Need for Collaboration the development of reference enterprise models and corresponding languages and tools requires Extending the scope of enterprise models and resources that are not available to a single re- developing them to an omnipresent represent- search institute. Furthermore, establishing and ation of organisations requires an amount of disseminating them in practice depends on eco- research and development that cannot be car- nomic and political aspects that are beyond the ried out by the current enterprise modelling com- abilities and intentions of academics. Against the munity alone. To advance the field, there is need background, it is obvious that there is need to for cross-disciplinary collaboration and for accu- bundle resources of research institutions. At the mulating resources. same time, it is necessary to get vendors of enterprise software and prospective users involved. 5.1 The Importance of Bundling On the one hand, they need to be involved to sup- Resources port requirements analysis. On the other hand, The development of comprehensive enterprise using reference artefacts in practice is the only models will overburden most organisations. This way to promote their dissemination. Unfortu- is the case, too, for respective DSML. Reference nately, there are serious obstacles that impede enterprise models and wide-spread DSML are both bundling of research resources and involve- suited to effectively address this problem. Fur- ment of companies. Research is based on compet- thermore they would provide a foundation for ition and the idea of scientific progress. Collab- cross-organisational integration of action sys- oration of research institutes implies to give up tems and information systems, which could en- competition to a large extent. At the same time, able a tremendous boost of productivity. At the for reference artefacts to be beneficial they need same time, the development of reference artefacts to be consolidated—which may jeopardise sci- that combine a descriptive and a prescriptive ap- entific progress. While there are probably many proach constitute an attractive research goal: It is vendors and client organisations that would be Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 34 Ulrich Frank

happy to use reference models, most of them will Enterprise modelling is not an end itself. Instead, likely be reluctant to participate in respective it is supposed to have a positive impact on a com- development projects, since the return on such pany’s economics and competitiveness. However, an investment would be hard to determine. Nev- assessing the costs of creating and maintaining ertheless, to promote the benefit of enterprise enterprise models, which may include the devel- models, reference artefacts that enable attractive opment of languages and tools, is not a trivial economies of scale are of pivotal relevance. task—and it is similarly challenging to determ- Request: There is need for initiatives to collabor- ine the benefits that can be contributed to the atively develop and disseminate reference arte- deployment of enterprise models. Apart from facts. They need to provide convincing incentives economic effects, the extensive use of enterprise both for academics and practitioners. models within an organisation may have an impact on how people perceive not only their tasks One of the prime examples of community-driven but also the entire organisation. The increase collaboration is free and open source software in transparency may have an effect on estab- (FOSS). Respective initiatives have successfully lished patterns of organisational power and may promoted collaboration of developers and users. require new approaches to managing organisa- Also, they led to software systems of surprising tions. Against this background it is obvious that quality, and, in some cases, to an impressive dis- enterprise modelling involves a wide range of semination. Inspired by the apparent success of demanding research questions that concern vari- some FOSS projects, corresponding “Open Mod- ous disciplines. Business and Administration in els” initiatives have been proposed (Frank and general is aimed at developing and improving ter- Strecker 2007) and inspired the creation of open minologies and methods that are suited to struc- model repositories (France et al. 2007), (www. ture and guide purposeful action in enterprises. openmodels.org, www.openmodels.at). While Various subfields, such as Financial Management, these repositories have triggered remarkable at- Accounting, Industrial Management, Logistics tention, there is still need for more active parti- could contribute to extend and deepen enterprise cipation. models. In Psychology, the interaction between cognitive models and external representations 5.2 Enterprise Models as Object and is a core research topic. Applied to enterprise Promoter of Cross-Disciplinary modelling this would include the question how Collaboration conceptual models effect individual and collect- Enterprise models are aimed at providing a me- ive decision making. That includes analysing the dium to foster communication between stake- impact of graphical notations on people’s ability holders with different professional background. to understand complex matters—and the develop- On the one hand that requires reconstructing ment of guidelines for designing notations that technical languages and professional patterns of fit certain cognitive styles. Both, from a psycho- problem solving. On the other hand it recom- logical and a sociological point of view, it would mends analysing how prospective users react be interesting to analyse how enterprise models upon the models they are presented with. That effect the social construction of reality, i.e., to includes concepts as well as their designation what extent people perceive the model as the en- and (graphical) representation. For using enter- terprise and what that means for the way they prise models effectively, software tools are man- (inter) act. Assuming that enterprise models may datory. Developing and integrating them with have a substantial effect on an organisation’s per- other enterprise software systems creates sub- formance implies challenging research questions stantial challenges for Software Engineering or— for economic studies that are not restricted to in other words—interesting research questions. enterprise models, but comprise the economics Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Enterprise Modelling: The Next Steps 35

of models and methods in general. Combining re- with more specific ones. Last but not least, enter- search results from various disciplines would not prise models provide a laboratory for learning, only contribute to advance our knowledge about because they convey a solid conceptual found- enterprise models and our ability to use them ef- ation of information systems and surrounding fectively, it is also suited to enrich the state of the action systems—and enable students to navigate art in the participating disciplines, since it would through an enterprise on different levels of de- integrate it with contributions from other fields. tail and abstraction. With respect to such a wide Therefore the following request could have a bet- and deep conception of enterprise modelling it is ter chance to succeed than yet another call for important not only to identify relevant steps of inter-disciplinary research. future research, but also to spread the word and Request: Advancing the field of enterprise model- encourage others to participate in joint projects. ling recommends to establish inter-disciplinary Further developing the field also requires to put research collaboration. more emphasis on assessing model artefacts. On the one hand that comprises the development of 6 Conclusion pragmatic criteria to evaluate models and modelling languages with respect to an intended prac- In the past, enterprise modelling, though argu- tical use. On the other hand, it relates to assess- ably pointing at a core topic of Information Sys- ing the epistemological quality of model artefacts tems, has been subject of a rather small, special- as research results. Developing and applying re- ised research community. In Business and Ad- spective criteria is an important prerequisite of ministration it is regarded as too much focussed scientific competition and progress. on technical aspects by some, while some tradi- tionalist colleagues in Computer Science suspect In this paper I gave a personal account of the it of lacking formal rigour. However, enterprise topics we should address in the next years to modelling is more than analysing and designing advance our field. It is needless to say that other information systems—and it is certainly much relevant topics exist, too. I would hope that the more than drawing “‘bubbles and arrows”. Enter- requests presented in this paper contribute to a prise modelling is about conceptualising an im- discourse on our future research agenda. portant part of the world—as it actually is and as it might be. Hence, it requires knowledge about References how people (inter) act in organisations, how information systems infrastructures are built—and der Aalst W., Basten T. (2002) Inheritance of the creativity to develop substantial images of Workflows – An Approach to Tackling Prob- attractive future worlds that comprise the pur- lems Related to Change. In: Theoretical Com- poseful construction and use of information sys- puter Science 270, p. 2002 tems. It is about how we perceive the world Atkinson C., Gutheil M., Kennel B. (2009) A Flex- we live and work in and how we think about it ible Infrastructure for Multilevel Language and might change it—alone and together with Engineering. In: IEEE Transactions on Soft- others. In addition to supporting collaboration ware Engineering 35(6), pp. 742–755 between stakeholders with different professional Blair G., Bencomo N., France R. B. (2009) backgrounds in organisations, enterprise models Models@ run.time: Computer. In: Computer may also serve as a medium and object of inter- 42(10), pp. 22–27 disciplinary research. At the same time, they Brinkkemper S. (1996) Method Engineering: En- are suited to foster the exchange between prac- gineering of Information Systems Develop- tice and academia, because they allow to integ- ment Methods and Tools. In: Information and rate more abstract representations of enterprises Software Technology 38(4), pp. 275–280 Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 36 Ulrich Frank

Buckl S., Matthes F., Roth S., Schulz C., Schweda 2nd Edition. ICB-Research Report 43 C. (2010) A Conceptual Framework for Enter- Frank U. (2012) Specialisation in Business Pro- prise Architecture Design. In: Proper E., Lank- cess Modelling: Motivation, Approaches and horst M. M., Schönherr M., Barjis J., Over- Limitations. 51. University of Duisburg-Essen. beek S. (eds.) Trends in Enterprise Architec- Essen ture Research. Lecture Notes in Business In- Frank U. (2013a) Domain-Specific Modeling Lan- formation Processing Vol. 70. Springer, Ber- guages - Requirements Analysis and Design lin, Heidelberg and New York, pp. 44–56 Guidelines. In: Iris Reinhartz-Berger, Aron CIMOSA: Open system architecture for CIM. Sturm, Tony Clark, Yair Wand, Sholom Co- Springer, Berlin, Heidelberg and New York hen, Jorn Bettin (eds.) Domain Engineering: Clark T., Willans J. (2012) Software Language Product Lines, Conceptual Models, and Lan- Engineering with XMF and XModeler. In: guages. Springer, pp. 133–157 Mernik M. (ed.) Formal and Practical Aspects Frank U. (2013b) Multi-Perspective Enterprise of Domain-Specific Languages. Information Modeling: Foundational Concepts, Prospects Science Reference, pp. 311–340 and Future Research Challenges. In: Software Clark T., Sammut P., Willans J. (2008) Superlan- and Systems Modeling (in print) guages: Developing Languages and Applica- Frank U., Strecker S. (2007) Open Reference Mod- tions with XMF. Ceteva els – Community-driven Collaboration to Dijkman R., Dumas M., van Dongen B., Käärik Promote Development and Dissemination of R., Mendling J. (2011) Similarity of business Reference Models. In: Enterprise Modelling process models: Metrics and evaluation. In: and Information Systems Architectures 2(2), Inf. Syst. 36(2), pp. 498–516 pp. 32–41 Ferstl O. K., Sinz E. J. (1998) Modeling of Frank U., Strecker S. (2009) Beyond ERP Systems: business systems using the Semantic Object An Outline of Self-Referential Enterprise Sys- Model (SOM): A methodological framework. tems: Requirements, Conceptual Foundation In: Bernus P., Mertins K., Schmidt G. (eds.) and Design Options Handbook on Architectures of Information Group T. O. (2009) TOGAF Version 9. Van Haren, Systems. International Handbooks on Inform- Zaltbommel ation Systems Vol. 1. Springer, Berlin, Heidel- Henderson-Sellers B., Ralyté J. (2010) Situational berg and New York, pp. 339–358 Method Engineering: State-of-the-Art Re- Fettke P., Loos P. (eds.) Reference Modeling for view. In: Journal of Universal Computer Sci- Business Systems Analysis. Idea Group, Her- ence 16(3), pp. 424–478 shey Kelly S., Tolvanen J.-P. (2008) Domain-specific France R., Bieman J., Cheng B. (2007) Repository modeling: Enabling full code generation. for Model Driven Development (ReMoDD). Wiley-Interscience and IEEE Computer Soci- In: Kühne T. (ed.) Models in Software En- ety, Hoboken and N.J gineering. Lecture Notes in Computer Sci- Kleppe A. G. (2009) Software language engin- ence Vol. 4364. Springer Berlin Heidelberg, eering: Creating domain-specific languages pp. 311–317 using metamodels. Addison-Wesley, Upper Frank U. (1994) Multiperspektivische Un- Saddle River and NJ ternehmensmodellierung: Theoretischer Hin- Koschmider A., Oberweis A. (2007) How to de- tergrund und Entwurf einer objektorientier- tect semantic business process model vari- ten Entwicklungsumgebung. Oldenbourg, ants? In: Proceedings of the 2007 ACM sym- München posium on Applied computing. SAC ’07. Frank U. (2011) The MEMO Meta Modelling ACM, New York, NY and USA, pp. 1263–1264 Language (MML) and Language Architecture. Krogstie J. (2007) Modelling of the People, by the Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Enterprise Modelling: The Next Steps 37

People, for the People. In: Krogstie J., Opdahl ACM, pp. 451–464 A., Brinkkemper S. (eds.) Conceptual Mod- Strecker S., Frank U., Heise D., Kattenstroth H. elling in Information Systems Engineering. (2012) MetricM: a modeling method in sup- Springer Berlin Heidelberg, pp. 305–318 port of the reflective design and use of per- Land M. O., Waage M., Proper E., Cloo J., formance measurement systems. In: Inform- Steghuis C. (2009) Enterprise architecture: ation Systems and e-Business Management Creating value by informed governance. 10(2), pp. 241–276 Springer, Berlin Völter M. (2013) DSL Engineering: Designing, Lankhorst M. M. (2005) Enterprise architecture Implementing and Using Domain-Specific at work: Modelling, communication, and ana- Languages. dslbooks.org lysis. Springer, Berlin, Heidelberg and New Zachman J. A. (1987) A framework for inform- York ation systems architecture. In: IBM Systems Liskov B. H., Wing J. M. (1994) A Behavioral Journal 26(3), pp. 276–292 Notion of Subtyping. In: ACM Transactions on Programming Languages and Systems 16, Ulrich Frank pp. 1811–1841 Moody D. L. (2009) The “Physics” of Notations: Chair of Information Systems and Enterprise Toward a Scientific Basis for Constructing Modelling Visual Notations in Software Engineering. In: Institute for Computer Science and Business IEEE Transactions on Software Engineering Information Systems 35(6), pp. 756–779 University of Duisburg–Essen Ralyté J., Agerfalk P. J., Kraiem N. (2005) Universitätsstraße 9 Methods, Techniques and Tools to Support 45141 Essen Situation-Specific Requirements Engineering Germany Processes: SREP’05. University of Limerick, [email protected] Limerick Ralyté J., Brinkkemper S., Henderson-Sellers B. (2007) Situational method engineering: Fun- damentals and experiences. Proceedings of the IFIP WG 8.1 Working Conference, 12-14 September 2007, Geneva, Switzerland. IFIP - the International Federation for Information Processing Vol. 244. Springer, New York Scheer A.-W. (1992) Architecture of Integrated Information Systems: Foundations of Enter- prise Modelling. Springer, Berlin and New York Schrefl M., Stumptner M. (2002) Behavior- consistent specialization of object life cycles. In: ACM Transactions on Software Engineer- ing Methodologies 11(1), pp. 92–148 Simonyi C., Christerson M., Clifford S. (2006) Intentional software. In: Proceedings of the 21th Annual ACM SIGPLAN Conference on Object-Oriented Programming, Systems, Lan- guages, and Applications (OOPSLA 2006). Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 38 José Tribolet and Pedro Sousa and Artur Caetano

José Tribolet and Pedro Sousa and Artur Caetano

The Role of Enterprise Governance and Cartography in Enterprise Engineering

Enterprise artography is fundamental to govern the transformation processes of an organisation. The artefacts of enterprise cartography represent the structure and dynamics of an organisation from three temporal views: as-was (past), as- is (present), and to-be (future). These views are dynamically generated from a continuous process that collects operational data from an organisation. This paper defines a set of enterprise cartography principles and provides an account of its role in understanding the dynamics of an organisation. The principles are grounded on control theory and are defined as a realisation of the observer and modeller components of the feedback control loop found on dynamic systems. As a result, an organisation can be abstracted as a dynamic system where a network of actors collaborate and produce results that can be depicted using cartographic maps.

1 Introduction isation during such transition. This is important because during each transformation initiative an This paper explores the role played by enterprise organisation has to react to events. Some of these cartography and enterprise governance within events may be unrelated to the transformation the enterprise engineering discipline. Enterprise initiative but may impact the transformation pro- governance relates to enterprise transformation cess and therefore deviate the organisation from since the change of operational processes, resources and business rules define new manage- achieving the planned future state. ment boundaries (Hoogervorst 2009). Enterprise This paper presents two contributions. The first architecture contributes to enterprise transform- is defining enterprise cartography as a function ation as it enables modelling the organisation’s of the observer and modeller roles as defined by structure and dynamics along with the underly- the enterprise’s dynamic feedback control loop. ing restrictions and design principles (Lankhorst Enterprise cartography is not associated with 2013; Op’t Land 2009). Transformation is often the enterprise design, but with the abstraction seen as the set of initiatives that change the or- and representation of the enterprise reality. Al- as-is ganisation’s domain from the current state though this differentiates enterprise cartography to-be to an intended state. These two states de- from enterprise architecture, it may be correctly scribe organisational variables at different mo- pointed out that cartography is part of enterprise as-is ments in time. The state is defined by the architecture. But given the relevance of carto- variables that changed due to past events, while graphy to understand the dynamics of the feed- the to-be state specifies an expected state config- back control loop of an organisation, we opted to uration of the organisational variables. Between discuss the concerns of cartography separately these two events, the organisation reacts to other from those of enterprise architecture. The second events that are triggered by the operation of the contribution of the paper is stating the empirical transformation processes. principles that ground the design of the carto- The issues we address in this position paper focus graphy process to play the role of the observer in the ability to observe and govern the organ- and modeller in the enterprise dynamic feedback Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 The Role of Enterprise Governance and Cartography in Enterprise Engineering 39

control loop. Dynamic systems and enterprise 3 Enterprise Governance governance are described in Sect.2 and Sect.3. An enterprise is a network of independent act- Section4 presents enterprise cartography. ors. Actors collaborate with other actors along time and thus create a dynamic collaborative net- 2 Dynamic Systems work. Actors also produce autonomous behaviour that may change the overall state of the system. Actors can be classified ascarbon-based The application of systems theory to systems actors, i.e. humans, and silicon-based actors, i.e. engineering has been discussed since the 1970s computers. This network runs within a domain (Eriksson 1997; Moigne 1977). Systems theory where the independent actors behave towards a relates to organisational systems mainly through future state of affairs, and thus produce events, the principles of dynamic systems, especially con- some of which may be unexpected. Therefore, trol feedback loops (Abraham et al. 2013; San- all enterprise domain state changes are a con- tos et al. 2008). These concepts can be further sequence of the individual behaviour of an actor combined with classic management theories as or of the composite behaviour that derives from a means to clarify how feedback loops interact the actor collaborations. These collaborations with different organisational views, such as gov- may occur between actors that are enclosed by ernance, management, and operations (Fig.1). the organisation’s boundary, or between an actor that is external to the organisation and one in- modeller In control theory, the presents a system ternal actor. So, the behaviour of an enterprise as-is view that specifies its current state (Levine “is” a result of what “it does”. An enterprise can 1996). The current state makes possible to estim- therefore be regarded as a large “bionic” distrib- ate a future state of the system in the absence of uted network of carbon-based and silicon-based unexpected events. To handle the potential devi- actors that are continuously interacting and pro- ations that occur from such events, control the- ducing behaviour. ory introduces the concept of controller. The controller analyses the continuous stream of events The current technological advances make pos- and modifies the system’s controllable variables sible near real-time, transparent and ubiquitous as a means of keeping the system behaving as interaction between people and systems. As such, planned (Fig.2). This is similar to the control of a the boundary between manual, semi-automated physical body moving toward a target: the mod- and even some automated operations becomes blurred. This means that the actions performed eller determines the current position and speed by people cannot be easily separated from those of the object and feeds it to the controller; if an of people supported by a network of computers, unexpected event occurs, then the controller cor- and from those of networks of computers. These rects the movement of the object by applying the collaborations can be abstracted as the result of a necessary forces and thereby ensuring that the single network that operates in (near) real-time. target is reached. The actors that interact within this network act We argue that the relationships between enter- autonomously. prise governance and enterprise cartography can Autonomous behaviour is evident from how a be established using the principles of dynamic person acts within an organisation since the state systems feedback control, where cartography change produced by a human actor can only be plays the aforementioned roles of observer and observed after the action is concluded. But the modeller. These relationships are explained in same phenomenon is also observed on informa- the next two sections. tion systems because one can only assert what Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 40 José Tribolet and Pedro Sousa and Artur Caetano

Figure 1: Organisational views and feedback loop, adapted from (Abraham et al. 2013).

This reasoning supports the conclusion that enterprises are dynamic systems. Enterprises are actually a system of systems, composed of and part of other dynamic systems. As such, there Figure 2: A single-input, single-output feedback loop. is an opportunity to try to understand an enterprise as a complex system through the lenses of systems theory, in particular through the body a computer actor has produced after the actual of knowledge of systems theory and dynamic action is performed. The degree of predictabil- systems control. However, this application must ity of automated computer actions is potentially always consider the intrinsic bionic nature of an higher than that of humans. But achieving cer- enterprise, as people cannot be dissociated from tainty is not feasible due to a number of factors. its essence. We defend that all this body of know- On the one hand, a system may not behave as ledge is directly applicable to enterprises through expected due to faults or failures. And even in enterprise engineering. The fundamental pur- the absence of faults of failures, the system may pose of engineering is to provide humans with be misaligned with the business. On the other artefacts that augment their individual and col- hand, the interaction between multiple systems lective capability to deal with specific situations. can produce emergent behaviour, meaning that Engineering helps humans to understand reality the overall behaviour of the system may not be and to pro-actively and purposefully transform it as idealised by individual and collective goals. the linear sum of each individual behaviour unit. This is the primary purpose of enterprise engin- As a result, there is a potential gap between the eering (Dietz et al. 2013). results that derive from planned actions and the actual results. This makes it impossible to fully How do systems theory and dynamic systems estimate the outcome of the interactions in such control relate to enterprise engineering? Well, a network. let us start with the “bionic state machine” meta- Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 The Role of Enterprise Governance and Cartography in Enterprise Engineering 41

phor presented earlier. According to systems systems control model. The actions of a per- theory, this model can be abstracted as two sep- son are the result of a combinatorial proced- arate subsystems: a feed-forward action system, ure: a person observes the world, attempts which is combinatorial in nature and transforms to contextualise and understand its meaning, inputs into outputs, and a feed-back cybernetic and then performs an action. This procedure system, which uses as input the state observa- corresponds to the role of controller. By act- tions and results provided by the feed-forward ing as a controller, the person can correct the action system. The feed-back system uses this deviations between the current state and the information to continuous estimate the current intended state. As such, to achieve goals an state of the system. This is accomplished by con- human actor operates his own local feed-back textualising the observations, i.e. by situating the subsystem. These actions do not occur at the observations into the semantic model of the sys- operational layer but at an higher layer that tem. Based on these observations, the feed-back plans and controls the operations (Abraham cybernetic system then decides on the actions et al. 2013). that all the actors of the system must perform Principle 3 An enterprise is more than the sum in order to keep the system on a trajectory that of its actors and resources. Organisational achieves its goals. This process is continuously factors such as culture, values, power, and hier- performed. These concepts have been extens- archical structures are elements in defining an ively applied to most engineering areas for at enterprise. We abstract these “soft” factors as least half a century (Andrei 2005). quality requirements that constrain and para- In this paper we hypothesise that the application metrise the operating system of an human of control theory is useful to help understanding actor. They are key determinants to the way enterprise engineering. The next hypothetical a human interprets the observations of reality, principles characterise enterprise governance as as well as he reads these observations through a dynamic systems theory problem. his own models of the world, based on which his own sense making operates. These factors Principle 1 Actions performed by people are have impact on the actions of a human actor enacted by the feed-forward action system. since they change how it plays the controller People play multiple actor roles within an role. enterprise such as operational, middle man- Principle 4 Enterprise self-awareness requires agement, knowledge work, auditing, advisory, the specification of the domain of action. governance or executive roles. If an enter- This prise is abstracted as a layered system, all is the realms of enterprise governance. Gov- these actions occur at the operational layer, ernance actions are distinct from executive, where actual operations are performed by act- managerial, and operational actions, because ors. People are abstracted as actors playing they are geared towards the preservation of roles within well specified semantic domains the enterprise self-awareness. Hence, gov- that uniquely define their contexts of indi- ernance focuses on the design rules and prin- vidual action and interaction (Caetano et al. ciples that constrain the enterprise actors, along 2009; Zacarias et al. 2005). An actor is capable with their actions and interactions. of playing several roles simultaneously. Principle 5 Maintaining the enterprise as a single Principle 2 A person can be abstracted as a sys- entity requires actors to dynamically main- tem of systems whenever its actions and in- tain a view of the actual state of the enter- teractions occur within the enterprise net- prise. work. This means that the roles played by The previous principles state the relationships people are subject to the rules of the dynamic between an individual actor and its own dynamic Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 42 José Tribolet and Pedro Sousa and Artur Caetano

control system. But how do the multiple actors, reality can be modelled in ways that commu- either carbon or silicon, interact and produce nicate information effectively. Enterprise carto- composite behaviour? Using a metaphor: what graphy deals with providing up-to-date model- makes a group of heterogeneous and autonomous based views of an enterprise architecture and its musicians become a musical ensemble? Why is goal is facilitating its communication and ana- this collective entity more than the linear sum of lysis. We have been successfully applying enter- its individual parts? So, what defines the bound- prise cartography concepts to enterprise architec- ary of an enterprise? What forces bind together ture projects (Caetano and Tribolet 2006; Caetano its autonomous actors as a single entity? We et al. 2009, 2012b; Sousa et al. 2007, 2009) and de- believe that the answer to this question lies in veloping computer-based tools to support enter- the enterprise’s “semantic model of itself”. We prise cartography (Caetano et al. 2012b; Filipe call this enterprise self-awareness (Abraham et et al. 2011; Sousa et al. 2011). Currently, the prin- al. 2013; Potgieter and Bishop 2003; Santos et al. ciples described here are implemented in a com- 2008). This means that if an enterprise has a commercial tool that is being used in several medium 1 mon semantic model of its actors then in becomes and large scale enterprise architecture projects . a single collective entity. If there is no common This section describes some empirical findings semantic model then the actors are unable to that we have observed in these cases. be self-aware of their context and as a result no The concept of abstracting reality through repres- single collective entity can be defined. This se- entations is not limited to engineering disciplines. mantic model is a shared dynamic model that is Cartography itself is an established discipline constantly updated by all its active components. that has played a major role in the development It is precisely this shared semantic model that of mankind. Cartography is an abstraction pro- defines a musical ensemble: each musician has cess that systematically and consistently trans- its own role, but both individually and as a whole forms an observation of reality into a map or a they are self-aware that they share the goal of graphical representation. The production of a playing the same piece of music according to a map embraces many different concerns, includ- set of rules. ing scientific, technical, and purely aesthetic. En- terprise cartography denotes the discipline that The systemic nature of an enterprise and its cy- deals with the conception, production, dissemin- bernetic attributes stress the need for having en- ation and study of the maps of an enterprise to gineering artefacts to support the collective un- support its analysis and collective understanding. derstanding of its changing reality. Enterprise cartography is fundamental to support this task. Classic cartography is usually associated with Furthermore, this enterprise engineered augmen- the representation of static objects, as in the case ted capability is essential to support the increas- of geographic maps. Modern cartography deals ing challenges of enterprise governance, which with the representation of both static and dy- are essential to preserve the integrity of an en- namic objects and is commonly grounded in in- terprise as a collective entity. The next section formation science, geographic information sci- describes the goals of principles of enterprise ence and geographic information systems. Car- cartography. tography must also provide multiple consistent views of the same system. For example, geo- 4 Enterprise Cartography graphical maps often combine different views, such as political boundaries, topographic features and several other features. This entails defining Cartography is the practice of designing and creating maps. It is based on the premise that 1http://www.link.pt/eams/ Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 The Role of Enterprise Governance and Cartography in Enterprise Engineering 43

Figure 3: Relationships between meta-model, views, viewpoints, diagrams, and stakeholders, adapted from (The Open Group 2009). abstraction rules and classification mechanisms organisation. The process of organisational so that all of views are consistent. The carto- data collection is a core concern of enterprise graphy of dynamic objects also requires to ab- cartography. Data collection is not a concern stract the rules that constrain how objects change of the mainstream approaches to enterprise and relate to each other over time. architecture. Enterprise cartography deals with the dynamic Principle 2 Enterprise cartography focus on the design and production of architectural views that depict the components of an organisation and dynamic description of an organisation. It their dependencies. It shares its constructs with does not deal with the processes or governance enterprise architecture, such as meta-models, mod- of organisational transformation. The pur- els, views, repositories, frameworks, and design poseful transformation of organisations is ad- rules. However, its goal is descriptive. A view ex- dressed by enterprise architecture. presses the architecture of a system from the perspective of concerns defined by its stakeholders. Principle 3 Enterprise cartography keeps up- Views are defined by viewpoints, which establish to-date architectural views. This implies auto- the conventions for the construction, interpreta- mated or supervised data collection and view tion and use of architecture views (ISO/IEC/IEEE creation. Ideally, these tasks should be per- 2011; The Open Group 2009). Figure3), taken formed at the same frequency as that of or- from the TOGAF 9 specification, illustrates the ganisational change. Enterprise architecture basic relationships between these concepts. The following principles distinguish cartography from techniques do not aim to provide systematic enterprise architecture. support for data collection nor the automated Principle 1 Enterprise cartography uses obser- design and creation of views, meaning these vations to produce the representations of an tasks are usually manual and creative. Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 44 José Tribolet and Pedro Sousa and Artur Caetano

4.1 Approaches to Enterprise the feed-back control loop by providing man- Cartography agers with a model of the organisation that allows them to ground their actions and decisions. There are several approaches to generate organisational models from the data extracted from Enterprise cartography is already a reality in sev- enterprise systems. Configuration Management eral domains. However, handling dynamic ob- Databases (CMDB), as defined by ITIL (Adams jects, time and change is not explicitly addressed 2009), manage the configurations and relation- by most approaches. We aim at a generic and ships of information systems and technological systemic approach, very much in line with the concept of ”Enterprise Architecture Dashboard” infrastructure. To populate a CMDB, some solu- (Op’t Land 2009), that displays the enterprise cur- tions provide auto-discovery techniques that de- rent and future states, its performance and the tect nodes, virtual machines and network devices directions of the organisation transformation pro- to create infrastructural views. Auto-discovery cess. is actually a cartographic process and requires that the type of the concepts to be discovered 4.2 Principles of Enterprise is specified in advance (Filipe et al. 2011). The Cartography resulting CMDB instance will contain a partial model of the organisation’s infra-structure. This This section describes a set of principles that model can be communicated through different define Enterprise Cartography. These principles but consistent visualisation mechanisms, such use the following definitions. as textual reports or graphical models that are Project is an transformation process designed to designed according to a symbolic notation and achieve a goal specified by a to-be state. design rules (Lankhorst 2013). Organisation variable references specific inform- At the business and organisational layer there ation or a value associated to an organisational are several cartographic techniques defined by artefact. business process management (Dumas et al. 2013) Organisation state contains the values of a sub- and process mining (Aalst et al. 2012). These tech- set of organisation variables at a given point in niques make use of event logs to discover process time. activities, control and data flows, as well as or- As-was state is the set of all organisation states ganisational structures (Aalst 2011; Aalst et al. observed in a specific point in the past. 2012; Agrawal et al. 1998). In this case, discovered As-is state is the set of organisation states as processes correspond to actual instances of pro- observed in the current point in time. cesses, not to the designed processes. Model state organisation states analysis can also be used to assess the conform- To-be is the set of that ance of processes against constraints (Caetano et are predicted to occur in a specific point in the al. 2012a; Molka et al. 2014). Another example of future. enterprise cartography is the inference of inter- Principle 4 The as-is state is defined by the as- organisational processes based on EDI event logs was and to-be states. (Engel et al. 2012). Semantic technologies, such Memory of the past state (as-was) and the fu- as ontologies, can also be used to analyse enter- ture state (to-be) define the behaviour of an or- prise models (Antunes et al. 2013, 2014). Business ganisation. The to-be state specifies the goals intelligence techniques that collect data from or- of transformation projects. Without the to-be ganisation systems to produce reports and dash- state the transformation processes cannot be boards are another example of cartography (Neg- executed or measured since no project goals ash 2004). Business intelligence actually supports are defined. Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 The Role of Enterprise Governance and Cartography in Enterprise Engineering 45

Principle 5 The definition of the to-be state al- ating state. They remain in this state until the ways precedes the definition of the as-is project successfully completes. After that the state. artefact becomes alive. An alive artefact dies Organisational artefacts must be always de- when a decommissioning project completes. A fined as goals in the to-be state before being gestating artefact can also die if the project is captured in the as-is state. This means that cancelled or not completed. A dead artefact the organisational artefacts are not created is retired when a retirement project explicitly incidentally but always as the result of a trans- removes the artefact from the organisational formation project. structure. Therefore, all state changes apply- Principle 6 All organisational artefacts can be ing to an artefact are the result of a transform- to-be classified as being in one of four invariant ation project. As such, the state always as-is states. precedes the state (Sousa et al. 2009). Gestating is the state that describes an organ- Principle 7 Organisation models and projects isation artefact after it is conceived, i.e. after plans are fundamental artefacts. it starts being planned, designed or produced. Organisation models and project plans must At this state, the artefact does not yet exist as be observed as variables whose values are cap- as-is an active element of the organisation in the tured during the state assessment. This sense it is not yet able to produce behaviour also means that architectural views, viewpoints, but can be passively used by organisational models and other architectural artefacts should transactions and processes. be regarded as organisation variables. For ex- Alive is the state that an artefact enters after ample, the repository of a UML modelling tool birth. Birth is the event that signals the mo- holding the specification of a system under ment when a gestating artefact enters the alive development must be an organisation artefact because it contributes to the specification of state. This means that the artefact is now able the to-be state. In contrast, a project is often to produce behaviour as part of the organisa- regarded as an organisation artefact. For in- tional transactions and processes. stance, both TOGAF and ArchiMate explicitly Dead is when a gestating or alive artefact is in- consider the concept of project Work Pack- active in the sense it is no longer able to play a age. However, organisational models, view- role in the organisational transactions and pro- points and views are not explicitly regarded cesses. This state is the opposite of gestation as artefacts by enterprise architecture model- that brought the artefact into existence. How- ling languages. Nonetheless, system architec- ever, a dead artefact may still have impact on ture guidelines such as ISO 42010 point out the the organisation. For example, an application importance of considering these elements as or server enter the dead state when they stop system artefacts (ISO/IEC/IEEE 2011). operating and will remain in that state until Principle 8 The state is sufficient to plan they are fully retired from the organisational to-be a transformation project. infrastructure. For the purpose of planning a transformation Retired represents the post-death state where project the current as-is state is not required the artefact is unable to further interact with because the to-be state must fully specify the other artefacts. organisational goals. Organisational artefacts exist first in the to-be state and only then in the as-is state. This ap- 4.3 Discussion plies to each state transition of the artefact’s Figure4 depicts a time line and a series of events life-cycle. Artefacts are conceived as the future in time (T0-T5). T0 represents the current mo- result of a project, thereby entering the gest- ment, therefore indicating the instant the as-is Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 46 José Tribolet and Pedro Sousa and Artur Caetano

Figure 4: Project planning and execution.

state was captured. At T0 the project P is con- time and starts today the corresponding project ceived and enters the gestating state: this project plan. The project planning phase must have an is planned to start at T3 and to be completed at understanding of the dependencies between that T5. Events T1, T2, T4 signal the completion of system and other systems, as well as to the busi- projects X, Y and Z, respectively. Therefore, T1, ness processes it supports. If no state changes T2, T4 also indicate that the artefacts that were occur in the next 6 months, then the organisa- produced by these three projects became alive. tion can indeed rely on the as-is state to plan Since project P is planned to start at T3 the or- the replacement project. But if the organisation ganisation requires knowing about its state at is performing a set of additional transformation state to-be(T3) and not at state as-is(T0) although projects that will change the organisation’s state planning is actually taking place at T0. This hap- during that period, then planning the system re- pens because the completion of projects X and placement project will require knowing about Y at T1 and T2 may interfere with the execution the sequence of to-be states during the next 6 of P at T3. Furthermore, the organisation also months and during the actual execution of the requires knowledge about its state at T4 because replacement project. Otherwise, it will not be the changes resulting from project Z may also possible to plan according to the actual network interfere with project P. of dependencies between the system to replaced To plan a transformation initiative an organisa- and other organisational artefacts. Therefore, for tion needs to be aware of the set of to-be states the purpose of project planning and execution, while the project is being executed. A descrip- the current as-is state will often not mirror the tion of the as-is state for planning purposes is organisation’s reality. In fact, the relevance of actually of limited use because there is often a the as-is state is inversely proportional to the temporal gap between project planning and pro- number of projects being completed per unit of ject execution. On the other hand, other projects time. At the limit, all dependencies of the system conclude and change the organisation state while to be replaced may change between the planning the project stands between planning and execut- and execution phases, meaning that all as-is state ing. These observations minimise the relevance variables will become irrelevant for planning pur- of the as-is state as a means to design the trans- poses. formation processes of the organisation. Nevertheless, the knowledge about an organisa- As an example, consider an organisation that tion’s current state is a fundamental asset for its plans the replacement of a system in 6 months operational management. At operational level, Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 The Role of Enterprise Governance and Cartography in Enterprise Engineering 47

actions and reactions are based on near real-time enterprise architecture projects that aim captur- observations and events, meaning that planning ing the as-is state as an enabler of transformation and execution occur in close sequence. However, planning. the requirements of the near real-time opera- This dilemma is found in many organisations: tional level of an organisation should not be inter- the contrast between the notion that an as-is as- twined with the medium to long-range require- sessment is a valuable asset for organisational ments required for organisational transformation transformation, and knowing at the same time and governance. that achieving such continuous task is demanding. This paper defends that an organisation does 5 Conclusions not need to have a full and accurate depiction of the as-is state but of its to-be state. The to-be Organisations do plan and execute projects, re- state is specified according to the specific goals gardless of not having a full or accurate rep- of projects, that are required for planning pur- resentation of the as-is or to-be states. Such poses. This contrasts with the as-is state that an accomplishment implies that projects include requires observing the variables of all organisa- to some degree an assessment of the impact of tional artefacts that are not retired. Consider a change between and during planning and execu- project that aims creating a new system that will tion. interact with an existing legacy system. Planning this project requires collecting information about An organisation that does not have a represent- the legacy system as well about the design of the ation of its to-be state will be unable to create a new system. However, the task of collecting in- detailed plan of project P as depicted in Fig.4. formation about a legacy system for the purpose This means that parts of the plan must be post- of project planning is actually contributing to poned until T3 to minimise the gap between the extending the knowledge about the current state planning and execution of P. This reality is com- of the organisation. This is a potential avenue monly observed in many organisations despite to sort out the dilemma stated earlier because a having impact on the project costs and risk, and representation of the as-is state can be built in- staff assignment. It also interferes with the plan- crementally by specifying the to-be state(s) that ning of other projects, thereby having negative are required to plan the multiple projects of an impact on the organisation’s agility. To remedy organisation. this issue, enterprise architecture projects often This position paper has presented a general frame- attempt to obtain a complete and accurate rep- work that provides representations of dynamic resentation of the as-is state. As a result, the organisations in the context of enterprise engin- primary goal of these projects is an attempt to eering. It specifically describes a set of prin- keep an organisational repository updated with ciples grounded on dynamic systems theory that as-is an observation of the state. This approach is provide guidelines on how to represent a carto- often justified by statements such as “knowing in graphic representation of an organisation. Such detail where we stand today is a pre-requirement representations facilitate the planning of organ- to any transformation project.” Although this isational transformation. sounds wise, this is a demanding task in terms of effort and time. Moreover, and as discussed be- References fore, the rate of organisational change will make the as-is state obsolete for the purpose of trans- van der Aalst W. (2011) Process mining: discov- formation planning. Therefore, we posit that ery, conformance and enhancement of busi- organisations should reassess the actual value of ness processes. Springer Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 48 José Tribolet and Pedro Sousa and Artur Caetano

van der Aalst W., Adriansyah A., de Medeiros A. Applied Computing, pp. 253–258 K. A., Arcieri F., Baier T., Blickle T., Bose J. C., Caetano A., Pereira C., Sousa P. (2012b) Gener- van den Brand P., Brandtjen R., Buijs J. et al. ation of Business Process Model Views. In: (2012) Process mining manifesto. In: Business Procedia Technology 5, pp. 378–387 process management workshops. Springer, Dietz J. L., Hoogervorst J. A., Albani A., Aveiro pp. 169–194 D., Babkin E., Barjis J., Caetano A., Huysmans Abraham R., Tribolet J., Winter R. (2013) Trans- P., Iijima J., Kervel S. J. V. et al. (2013) The formation of multi-level systems–theoretical discipline of enterprise engineering. In: In- grounding and consequences for enterprise ternational Journal of Organisational Design architecture management. In: Advances in and Engineering 3(1), pp. 86–114 Enterprise Engineering VII. Springer, pp. 73– Dumas M., Rosa M. L., Mendling J., Reijers 87 H. A. (2013) Fundamentals of Business Pro- Adams S. (2009) ITIL V3 foundation handbook cess Management, 2013 Edition. ISBN 978- Vol. 1. itSMF International, Office of Govern- 3642331428. Springer, p. 350 ment Commerce, ISBN: 978-0113311972 Engel R., van der Aalst W. M., Zapletal M., Agrawal R., Gunopulos D., Leymann F. (1998) Pichler C., Werthner H. (2012) Mining inter- Mining process models from workflow logs. organizational business process models from In: Lecture Notes in Computer Science EDI messages: A case study from the auto- Volume 1377. Springer, pp. 467–483 motive sector. In: Advanced Information Sys- Andrei N. (2005) Modern Control Theory: A tems Engineering. Springer, pp. 222–237 historical perspective. Center for Advanced Eriksson D. M. (1997) A Principal Exposition of Modeling and Optimization, Romania Jean-Louis Le Moigne Systemic Theory. In: Antunes G., Caetano A., Bakhshandeh M., Cybernetics & Human Knowing 4(2), p. 42 Mayer R., Borbinha J. (2013) Using Ontolo- Filipe J. A. C. B., Coelho M. F. P., Ferreira M. A. gies to Integrate Multiple Enterprise Archi- M., Figueredo I. C. (2011) Social Responsib- tecture Domains. In: Abramowicz W. (ed.) ility and Environmental Sustainability: The Business Information Systems. Lecture Notes Case of Caixa Geral de Depósitos (Portugal) in Business Information Processing Vol. 160. and Vale (Brazil). In: Editorial Board Mem- Springer, pp. 61–72 bers 10(5), pp. 352–375 Antunes G., Bakhshandeh M., Mayer R., Borb- Hoogervorst J. A. (2009) Enterprise governance inha J., Caetano A. (2014) Using Ontologies and enterprise engineering. Springer for Enterprise Architecture Integration and ISO/IEC/IEEE (2011) 42010. Systems and Soft- Analysis. In: Complex Systems Informatics ware Engineering: Architecture Description and Modeling Quarterly 1 Lankhorst M. (2013) Enterprise architecture at Caetano A., Assis A., Tribolet J. (2012a) Using work: Modelling, communication and ana- Business Transactions to Analyse the Con- lysis. Springer sistency of Business Process Models. In: 45th Levine W. S. (1996) The Control Handbook Lev- Hawaii International Conference on System ine W. S. (ed.). CRC Press Science, pp. 4277–4285 le Moigne J.-L. (1977) La theorie du systeme Caetano A., Tribolet J. (2006) Modeling organ- general. Theorie de la modelisation. Collec- izational actors and business processes. In: tion les classiques du reseau intelligence de Proceedings of the 2006 ACM Symposium on la complexite Applied computing. ACM, pp. 1565–1566 Molka T., Redlich D., Drobek M., Caetano A., Caetano A., Silva A. R., Tribolet J. (2009) A role- Zeng X.-J., Gilani W. (2014) Conformance based enterprise architecture framework. In: Checking for BPMN-Based Process Models. Proceedings of the 2009 ACM Symposium on In: Proceeedings of the 29th ACM Symposium Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 The Role of Enterprise Governance and Cartography in Enterprise Engineering 49

On Applied Computing José Tribolet Negash S (2004) Business Intelligence. In: Com- Department of Computer Science and munications of the Association of Informa- Engineering tion Systems 13, 177â195 Instituto Superior Técnico Op’t Land M. (2009) Enterprise architecture: cre- University of Lisbon ating value by informed governance Springer Av. Rovisco Pais (ed.). Springer 1049-001 Lisboa Potgieter A., Bishop J. (2003) The Complex Portugal Adaptive Enterprise: Self-Awareness and [email protected] Sustainable Competitive Advantage using Bayesian Agencies, In: The Journal of Con- Pedro Sousa vergence, UCT Graduate School of Business Department of Computer Science and Santos C., Sousa P., Ferreira C., Tribolet J. (2008) Engineering Conceptual model for continuous organiza- Instituto Superior Técnico tional auditing with real time analysis and University of Lisbon modern control theory. In: Journal of Emer- Av. Rovisco Pais ging Technologies in Accounting 5(1), pp. 37– 1049-001 Lisboa 63 Portugal Sousa P., Pereira C., Vendeirinho R., Caetano and A., Tribolet J. (2007) Applying the Zachman Link Consulting Framework Dimensions to Support Business Avenida Duque de Avila 23 Process Modeling. In: Digital Enterprise Tech- 1000 Lisboa nology. Springer, pp. 359–366 Portugal Sousa P., Lima J., Sampaio A., Pereira C. (2009) [email protected] An approach for creating and managing enterprise blueprints: A case for it blueprints. Artur Caetano In: Advances in Enterprise Engineering III. Department of Computer Science and Springer, pp. 70–84 Engineering Sousa P., Gabriel R., Tadao G., Carvalho R., Instituto Superior Técnico Sousa P. M., Sampaio A. (2011) Enterprise University of Lisbon Transformation: The Serasa Experian Case. Av. Rovisco Pais In: Practice-Driven Research on Enterprise 1049-001 Lisboa Transformation. Springer, pp. 134–145 Portugal The Open Group (2009) TOGAF, the open group and architecture framework, version 9. Van Haren INESC-ID Publishing Information Systems Group Zacarias M., Caetano A., Pinto H., Tribolet J. Rua Alves Redol 9 (2005) Modeling Contexts for Business Pro- 1000-029 Lisboa cess Oriented Knowledge Support. In: Althoff Portugal K.-D., Dengel A., Bergmann R., Nick M., Roth- [email protected] Berghofer T. (eds.) Professional Knowledge Management. Lecture Notes in Computer Sci- ence Vol. 3782. Springer Berlin Heidelberg, pp. 431–442 Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 50 Jorge L. Sanz

Jorge L. Sanz

Enabling Front-Office Transformation and Customer Experience through Business Process Engineering

The scope of business processes has been traditionally circumscribed to the industrialisation of enterprise operations. Indeed, Business Process Management (BPM) has focused on relatively mature operations, with the goal of improving performance through automation. However, in today’s world of customer-centricity and individualised services, the richest source of economic value-creation comes from enterprise-customer contacts beyond transactions. The need to make sense of a mass of such touch-points makes process a prevalent and emerging concept in the Front- Office of enterprises, including organisational competences such as marketing operations, customer-relationship management, campaign creation and monitoring, brand management, sales and advisory services, multichannel management, service innovation and management life-cycle, among others. While BPM will continue to make important contributions to the factory of enterprises, the engineering of customer-centric business processes defines a new field of multi-disciplinary work focused on serving customers and improving their experiences. This new domain has been dubbed Business Process Engineering (BPE) in the concert of IEEE Business Informatics. This paper addresses the main characteristics of BPE in comparison with traditional BPM, highlights the importance of process in customer experience as a key goal in Front-Office transformation and suggests a number of new research directions. In particular, the domains of process and information remain today disconnected. Business Informatics is about the study of the information process in organisations and thus, reuniting business process and information in enterprises is a central task in a Business Informatics approach to engineering processes. Among other activities, BPE is chartered to close this gap and to create a suitable business architecture for Front-Office where organisational and customer behaviour should guide and benefit from emerging data analytics techniques.

1 Process is out of the Industrialisation methods, techniques, and software to design, enact, Box control and analyse operational processes involving humans, organisations, applications, documents Business process has been at the center of the and other sources of information."1 stage in both research and industry for several decades. Under the brand of Business Process While the above definitions are quite compre- Management (BPM), business process has attrac- hensive and broad, in reality most BPM research ted a great deal of attention from many practi- and industry activity has grown upon the motiv- tioners and scholars. BPM has been defined as ation of reducing operating costs through auto- the analysis, design, implementation, optimisa- mation, optimisation and outsourcing. There are tion and monitoring of business processes (Du- a several Schools of thought and practice (such mas et al. 2013; Franz and Kirchmer 2012; Rosen- as lean, lean sixsigma, and others (Andjelkovic- berg et al. 2011; Schönthaler et al. 2012; Sidorova 1Aalst et al. (2003) exclude strategy processes from BPM, and Isik 2010). Aalst et al. (2003) defined some tar- a remarkable pointthat will be revisited in more depth later gets of BPM: ". . . supports business processes using in this paper. Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Enabling Front-Office Transformation and Customer Experience through Business Process Engineering 51

Pesic 2007; Andjelković Pešić 2004, 2006; Näslund Furthermore, the examples provided usually deal 2008)) and a myriad of related literature in the last with rather idealised operations, probably offered 40 years that serve to illustrate the focus on cost as simple examples with the purpose of illustrat- contention. Around the middle of the past dec- ing theoretical or foundational research results ade, T. H. Davenport (2005) stated in a celebrated (Aalst 2004; Aalst and Hee 2002; Aalst et al. 2003; Harvard Business Review paper that processes Yan et al. 2012). Thus, radically simplified ver- were being "analyzed, standardized, and quality sions of "managing an order", "approving a form", checked", and that this phenomenon was happen- "processing a claim", "paying a provider", "deliv- ing for all sort of activities, stated in Davenport’s ering an order" etc. are among the most popular own terms: "from making a mouse trap to hiring examples of processes found in the literature. a CEO". The actual situation is that industry investment and consequential research have stayed The lack of public documentation of substantial much more on "trapping the mouse" than in dif- collections of real-world processes is remarkable. ferentiating customer services through innovat- Houy et al. (2010) both confirmed the dominant ive and more intelligent processes, let alone hir- focus on simple business processes and also sug- ing CEOs. This may be explained partly from gested potential practical consequences of related Davenport’s own statements in 2005: "Process research: ". . . there is a growing and very active standards could revolutionize how businesses work. research community looking at process modelling They could dramatically increase the level and and analysis, reference models, workflow flexib- breadth of outsourcing and reduce the number ility, process mining and process-centric service- of processes that organizations decide to per- oriented architecture (SOA). However, it is clear form for themselves" (bold face is added here that existing approaches have problems dealing for emphasis). with the enormous challenges real-life BPM projects are facing [. . . ] Conventional BPM research With the advent of different technologies such seems to focus on situations with just a few isol- as mobile, cloud, social media, and other digital ated processes . . . ". Of course, the list of available capabilities that have empowered consumers, the real-world processes would be a lot richer if one classical approach and scope of business process included the set defined by enterprise packaged have begun to change quickly. Organisations are applications (Rosenberg et al. 2011). However, adopting new operating models (Hastings and this comprehensive collection is proprietary be- Saperstein 2007) that will drastically affect the cause it constitutes a key piece of intellectual way processes are conceived and deployed. As capital coming from software vendors or integ- stated by many authors in the last four decades, rators in the industry. business process work is supposed to cover all competences in an organisation, irrespective of The traditional focus on process has also raised the specific skills from human beings participat- much controversy. At the S-BPM ONE Confer- ing in such operations. However, in an unpub- ence in 2010, a keynote speaker (Olbrich 2011) lished inspection of about 1,300 papers conduc- remarked: "Let me be as undiplomatic as I pos- ted by the author and some of his collaborators2, sibly can be without being offensive [. . . ] The aca- most process examples shown in the literature demic community is as much to blame [. . . ] as the deal with rather simple forms of coordination of vendors of BPM systems, who continue to reduce work, mostly exhibiting a flow structure and ad- the task of managing business processes to a dressing administrative tasks (like those captured purely technological and automation- office information systems in early works on ). oriented level". While other authors in the same 2The co-authors are L. Flores and V. Becker both from conference debated "who is to blame" very an- IBM Corporation. imatedly (Fleischmann 2011; Singer and Zinser Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 52 Jorge L. Sanz

2011) it is important to highlight that the state- (2) Resulting process models have typically yiel- ment from Olbrich (in bold face above for em- ded the form of a "workflow" (Sharp and McDer- phasis) reinforces that BPM has mostly followed mott 2009; White 2004). This means that the activ- the obsession of automation and optimisation by ation of a task in the assembly line only occurs means of Information Technology. when certain predefined events take place, one or more previous tasks are completed and their pro- A detailed inspection of the extant literature con- duced artifacts transferred to the next task in the firms that business process work has been de- pipeline for continuing "the assembly". In fully voted to a rather small fraction of the actual automated systems, like software applications, variety and complexity found in enterprise be- this is a good abstraction (see Fig.1). On the haviour. This behaviour enacts many valuegen- other hand, in actual business processes where erating capabilities that organisations cultivate humans participate or supervise the individual based on skills provided by their own workforces tasks, workflows do not always capture the ac- and through rich interactions with other enter- tual pattern of work, including the contractual prise stakeholders, particularly customers. The commitments made across role-players. following points offer a simplified summary: Consequently, IT systems used to implement (1) Business process research in Computer Sci- such workflows, called "Business Process Man- ence has been traditionally focused on certain agement Systems" (BPMS) in IT jargon3, are not classes of enterprise operations, mostly involving suitable to communicate the nature of work to simple coordination mechanisms across tasks. business stakeholders. This point has been ex- This type of coordination and the overall beha- tensively addressed in recent Enterprise Engin- viour represented in underlying models reflect eering work (Dietz et al. 2013), such as DEMO very much an "assembly line" where work is lin- and related contributions (Albani and Dietz 2011; early synchronised to deliver a desired artifact Aveiro et al. 2011; Barjis et al. 2009; Proper et or outcome. BPMN, emerged from OMG as the al. 2013). The issue of clarity was brought up industry standard for business process modelling by Dietz eloquently during a key-note entitled is a good illustration of this point. Simplicity of "Processes are more than Workflows" in the 2011 the choreography is ensured by removing any KEOD Conference: "With modelling techniques form of overhead in communication when mov- like Flowchart, BPMN, Petri Net, ARIS/EPC, UML ing from one stage to the next. Unlike other more and IDEF you get easily hundreds of pages of pro- complex business processes, many software ap- cess diagrams. Nobody is able to understand such models fully. Consequently, nobody is able to re- plications do have this simplified structure. In design and re-engineer a process on that basis fact, a trend since the early 2000’s is to separ- ". ate the specific application logic from the co- Beyond communication issues, the distinction ordination / choreography needed across mod- of contexts between an organisational design ules, and both of them from the actual data contained in a data-base management system. Dif- 3The term BPMS is somewhat questionable because it ferent foundations and a plethora of languages implies that these IT systems implement processes while they actually do so only for very special types of processes, have been created to capture this semantics of i.e., workflows. Thus, the earliest denomination of Work- coordination such as Business Process Modelling flow Management Systems (WMS) is more adequate. As Notation (BPMN), Business Process Execution an example, Cases emerged later in the software industry Language (BPEL), Unified Modelling Language and model more complex processes. The term Case Man- agement Systems (CMS) has been used to distinguish them (UML), Event Process Chain (EPC), Petri Nets, from BPMS. This incorrectly implies "cases are not business etc. processes". Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Enabling Front-Office Transformation and Customer Experience through Business Process Engineering 53

concern and an IT concern should also be care- or even millions of interventions done with the fully addressed. In the workflow abstraction, the same process is adversarial to the business need potential role-players assigned to the execution of introducing modifications. As organisations or supervision of the individual tasks will be have been progressively more affected by sudden "idling" unless they get activated through the change or involved in operations where change pipeline. This model of reality is well-suited to is a common requirement this type of factory fully automated tasks (like those realised by soft- optimisation does not work. In fact, rigidity of ware) but unsuited to other situations in organ- process models has been a long-standing and bit- isations where humans take part of the process ter finding. More recently, the broader issue of execution. process evolvability in the presence of continuous change has been the subject of solid research, user user including a recent PhD thesis (Nuffel 2011) and interface interface references therein. (4) Implicitly or explicitly in the traditional approaches to business process, it lies the Taylorian principle of replacing individuals by applying

pplication pplication

A pplication pplication pplication automation whenever possible. As in other busi- A A A BPM system ness theories that build on a "dehumanisation" of

database database database enterprises, the consequence is that the role of system system system humans as sources of value-creation in processes 1960 1975 1985 2000 is ignored. The connection of this foundation and BPM work has been openly recognised by Figure 1: The evolution of information systems devel- Van der Aalst in his recent review of a decade of opment and the role of BPM systems in the newest generations of software (from Aalst et al. 2003). Business Process Management conferences Aalst (2012): "Adam Smith showed the advantages of the division of labor. Frederick Taylor introduced the Indeed, the factory model of operations captured initial principles of scientific management. Henry into a workflow implies that people are actually Ford introduced the production line for the mass "doing nothing" unless their "activation" occurs production of black T-Fords. It is easy to see that by the preceding tasks in the pipeline. The latter these ideas are used in today’s BPM systems". is far from modelling accurately the reality of work in most enterprise processes. In close connection to this moral coming from certain economics and business schools, it also (3) The tradition of business process management resides the goal of avoiding variation of the pro- works on the assumption that the investment cess by all possible means. This good idea origin- made in optimally designing a process will be ally coming from manufacturing practices (i.e., recovered through the repeated application of reducing variation as a means to controlling qual- the process for a long-enough period of time. ity and cost of the resulting production) has been The principle is that economic benefits will ac- translated to other forms of operations (such as crue from accumulated cost reduction obtained services) where variation is inevitable when inter- by the application of the optimised process over action with non-automated agents becomes an and over again. This approach reflects a true integral part of the actual production process.4 ’factory’ in the conception and modelling of or- 4 ganisational behaviour. Furthermore, the idea of Most call centers begin all their interaction with customers by following pre-established routines. In some cases, perfecting the process with such an effort pay- this may disgrace the effectiveness of the service and satis- ing off through hundreds of thousand repetitions faction of the caller. A known example is when reasonably Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 54 Jorge L. Sanz

Inevitable process variation is a significant sign be easily illustrated by using some of the Process of ’lost control’, as organisational capabilities go Classification Framework (PCF) content. from the tangible to the less tangible. As said in While some people may argue that this frame- the less tangible the capability, the Le Clair (2012), work may arguably be called a process architec- more control will be ceded to the customer . The ture (Eid-Sabbagh et al. 2012; Miers 2009; M. A. tradition of BPM work contrasts sharply with En- Ould 1997) it still provides a solid clue of many terprise Engineering (Dietz et al. 2013), a theory operations that are either common across indus- in which humans are seen as a precious source of tries or unique to specific industry segments value, particularly for achieving improvements such as retail banking or consumer packaged and differentiation. In particular, all processes goods. None of these enterprise operations can involving interaction with customers offer this be modeled by workflows. opportunity (services researchers often call this concept "co-creation"). In addition, the componentised business architecture and its resulting industry models addressed (5) It is important to recall that existing process in Sanz et al. (2012) are also very useful to illus- classifications such as the Process Classification trate the same points. In these approaches, there Framework (Process Classification Framework is no functional decomposition at the heart of the (PCF)) reveal common areas of work in organisa- modelling, unlike in PCF, and thus the resulting tions that do not follow the BPM tradition in the construction follows more closely some of the sense that they do not represent work amenable core principles of Enterprise Engineering (Dietz to workflows. Indeed, PCF is a standardisation et al. 2013). This will be addressed briefly in the effort in different industries that includes many next section. non-factory areas of an enterprise. Consequently, (6) Another important evidence that process has these operations are not adequately addressed by moved out of the industrialisation box is Case the application of existing BPM research, meth- Management (more recently also called Adaptive ods and tools. Case Management by the authors in (Swenson The clarification from Van der Aalst and his col- et al. 2010) and Dynamic Case Management by laborators when excluding strategy processes from analysts in Forrester). The need for Case Man- the scope of their work was an excellent and agement has been illustrated with different en- very early sign, although "strategy" should not terprise operations such as claim processing in have been the only area excluded from the scope Property and Casualty Insurance, customer ap- of their contributions. Indeed, there are other plications in Social Services, Health Care claim critical business processes in enterprises beyond processing, Judicial Cases, and so on. Van der "strategy" that do not fit workflow models, Petri Aalst and others (Aalst and Berens 2001; Aalst Nets, BPMN, or related instruments popular in et al. 2005) presented Case Handling as a new Computer Science (Sanz et al. 2012). Specifically, paradigm for supporting flexible and knowledge these other forms of organisational behaviour intensive business processes. In his work on case beyond ’the factory’ involve complex activities management, De Man (2009) states that ’work- carried our by humans in collaboration with one flow’ is an adequate representation for factory- another and with the support of technology in type, highly predictable behaviour admitting for ways that are observable and may also be cap- little or no deviation from pre-established models. tured into process models. This point can also In recent literature (Khoyi 2010), the argument in support of the need for Case Management hinged educated customers are asked first whether their obviously around the fact that "Case Management allows nonfunctioning product is plugged to the power supply, to unplug and plug it again, try to turn it on once more, and the business to be described in known terms rather so on. than artificially fitting it into a process diagram". Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Enabling Front-Office Transformation and Customer Experience through Business Process Engineering 55

Online

One seamless Email entity behind all the channels Mobile

Call center

Branches CUSTOMERS & Enterprise PROSPECTS Kiosk, ATM, etc.

Point of sale

Postal mail

process data Online SILOS strategy tech

process data Email SILOS strategy tech

process data Mobile SILOS strategy tech

process data Call center SILOS strategy tech

process data Branches SILOS CUSTOMERS & strategy tech PROSPECTS process data Kiosk, ATM, etc. SILOS strategy tech

process data Point of sale SILOS strategy tech

process data Postal mail SILOS strategy tech

2 © 2009 IBM Corporation Figure 2: Customers and prospects deal with an enterprise through a number of channels by following patterns or Journeys that vary according to individuals’ goals and behaviour. The picture on the upper side represents the expected experience of meeting the enterprise as a single and well-integrated entity. However, reality is very different as channels are not well-integrated, represented visually by the horizontal silos of the lower side picture. Each silo has their own processes, data, strategy (incentives) and IT. Thus, a Customer Journey is the integration of individual customer-enterprise touch-points to realise a specific customer outcome. These Journeys are essential processes deeply related to loyalty and other significant measures of customer experience, unlike traditional customer satisfaction metrics. Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 56 Jorge L. Sanz

2 Process and the Broken Customer teractions6 taking place across different enter- Experience prise channels (upper side of the Fig.). These patterns are typical for a single customer pursu- In the context of this paper, customer experience ing a specific outcome. In most organisations, is the conjunction of all experiences a consumer each channel behaves as a silo (lower side of has with an enterprise over the duration of their the Fig.) thus having its own strategy, goals, relationship (Harrison-Broninski 2005). Customer processes, data and technology. This discon- experience is critical for enterprises because it nect across channels impacts customer experi- has been widely understood as a key factor driv- ence quite negatively. In summary, a much more ing customer loyalty (Propp 1968). Poor customer engaged consumer through multiple channels experience in business-toconsumer enterprises is making the already disrupted customer experience for large enterprises has been a top concern in organisations for longer unmanageable . than five years. The main reason is the pro- All these challenges lead organisations to revisit found lack of loyalty that customers exhibit in some of their core competences related to cus- the business-toconsumer (b-to-c) industries (Cap- tomer experience. In fact, a number of key capab- 5 gemini 2012). While this challenge has been com- ilities have been emerging over the last decade, monplace in many industry segments, the prob- starting to yield best-practices for front-office op- lem is particularly acute in most b-to-c services erations (Hastings and Saperstein 2007). However, organisations where many initiatives have been it is the lack of understanding, modelling and taken to address the problem, even to the point instrumenting critical customer journeys the of introducing a new role at the top management main reason why customer experience contin- level called Customer Experience Officer (Bliss ues to be disrupted and has got worse with the 2006). advent of more channels. Furthermore, aligning these customer journeys with back-office opera- The advent of multiple channels of engagement tions yielding end-to-end business processes is es- for the same enterprise exposes deeper gaps in sential to enable customer experience. Business the way organisations deal with their custom- analysts characterised this new process trend dir- ers. Indeed, multiple channels have generated ectly affecting customer experience under dif- even more disconnects with customers as these ferent names and also alerted practitioners, re- channels are generally managed by different or- searchers and process professionals about dif- ganisational units and have isolated measures ferent shifts taking place along the entire "hype of performance. Traditional customer satisfac- cycle" of process evolution. In particular, Forres- tion measures tend to focus on individual cus- ter used the name "tamed processes" and char- tomer interactions on a specific channel but these acterised them as follows: "Tamed processes are do not seem to correlate positively with cus- designed from the outside in, can be driven by tomer loyalty (Rawson et al. 2013; Stone and big data and advanced analytics, support social Devine 2013). Figure2 illustrates customer in- and mobile technology, provide end-to-end support

5In North America, 80% of clients are "happy" with their 6The set of customer-enterprise interactions followed to bank service but only 50% say they will remain with their achieve a specific outcome for an individual costumer has current bank over the next 6 months. This reflects the been named customer journey (Rawson et al. 2013). This term finding that globally, only 42% of bank customers have has probably been coined by some technical and business rate their experience as being positive. Furthermore, sat- people with the goal of implying that the concept should not isfaction levels with branches, despite being the most ex- be made part of the classical "process grinding" experienced pensive and most developed channel, averages 40% world- though four decades of BPM, lean six-sigma and the like. wide with highest being 60% in North America (Capgemini Beyond communication intent, journeys are processes and 2012). this is a well-supported fact in Social Science work. Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Enabling Front-Office Transformation and Customer Experience through Business Process Engineering 57

across systems of record and functional areas, and ditional loss of loyalty and consequently, cus- link on-premises and cloudbased services" (Le Clair tomer equity or profitability issues (Villanueva 2012). and Hanssens 2007). Engineering (i.e., designing and running) these Probably to the surprise of many data analyt- customer journeys is a very different problem ics advocates, if customer-centric processes are from those BPM has been focusing in four dec- not engineered to reflect the demands from the ades. These needs around modelling and archi- new economy, the emphasis on individualising tecting for customer experience are in sharp con- customers and "inferring their behaviour" will trast to applying Customer Relationship Man- just make customer experience even worse. The agement (CRM) packaged applications used to reason is that customers will increase their ex- monitor sales, manage customer center calls or pectations for personalised services while the design optimised workflows for efficient backof- ability for organisations to address this expecta- fice processes. In fact, there is a risk that soft- tion remains far from the current state-of-the-art. ware may be used precipitately for supporting This issue will become particularly challenging enterprise capabilities related to customer exper- for some services industries because (i) such perience. Indeed, some of these emerging practices sonalisation may not be viable due to the nature are being made into software without adequate of the service being delivered; (ii) personalisation exposure of the underlying business processes. requires in many cases a co-created design and This should constitute a warning to management delivery, a pursuit that many enterprises are not as these software applications bury rich busi- yet in a position to address; (iii) regulatory lim- ness processes into their packaged software, itations may prevail thus limiting the enterprise thus signaling the same issues experienced in to discriminate across customers; or (iv) scalab- mature back-office operations. This warning is ility of good quality customer service may be at a significant call for the adequate research and odds with profitability targets. This remark is an practice necessary to surface the key processes attempt to warn "data scientist" approaches to before they are fully embedded into "concrete", front-office operations, as the main disconnects a fact that will impact agility as the frequency will only be widened by "data-only" insights. of change in these processes is a lot higher than 3 Process in critical areas of the in those modeled in conventional enterprise re- Front-Office source planning. Traditional approaches to business process instrumentation based on packaged The term "Front-Office" is used here to denote applications in conjunction with custom BPM the set of enterprise activities and resources ded- systems come to memory after four decades of icated to the support of customer experience. In cost-take out and efficiency improvements. In this category, they fall many customer service part, this rigidity has created fragmented cus- management operations. But other Front-Office tomer experience as a consequence of the lack areas in organisations also go beyond the pur- of flexibility and long time-to-value for desired pose of dealing directly with customers. Some changes in the information technology systems examples are brand monitoring, campaign design deployed across the enterprise. This is an obser- and deployment, enterprise marketing operations, vation coming from direct practice in the field product and service innovation, customer loyalty and can also be corroborated by exploring a very and advocacy management and others among extensive business literature. In short, if front- the top areas where organisations have been in- office processes are not addressed according to vesting in the last decade or so. These enterprise the new business and societal needs, the on- capabilities and related competences support cus- going fragmented experience will result in ad- tomers indirectly, although boundaries may blur Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 58 Jorge L. Sanz

in some cases (for example, a campaign design Notice that the hierarchy of resources represen- may involve realtime intervention based on cus- ted in Fig.3 does not mean the same as the clas- tomer interactions). These capabilities are be- sical management concept of "control". Instead, ginning to have more visible best-practices and it only represents an arrangement in which dif- thus, corresponding business processes are emer- ferent skills, information, assets (intangible and ging. Consequently, their study is at the realm of capital) and derivative entangled capabilities are Business Process Engineering because they en- bundled together to produce one or more relev- compass key work-practices. These operations in- ant outcomes in the enterprise. Likewise, these volve humans and collaborative activities deeply components are not necessarily aligned with tra- interrelated with technology and information, ditional Lines-of-Business and do not intend to and their patterns of work are also emerging, map departmental capabilities or other conven- become more and more visible, being subjected tional "reporting structures" in enterprises. Revis- to white box modelling rather than remaining as iting Penrose (2009), the components highlighted black boxes. In these new process areas, Inform- on the right may be thought as the formalised ation Technology will still be essential but in grouping of resources whose entanglement pro- radically different ways from "the factory" of en- duces those core services (internal or external) terprises. Actually, translating those experiences that the organisation needs to serve all stake- from Information Systems in the Back-Office to holders. Some enterprises may be endowed with the Front-Office is a sure recipe for disaster. This some of these resources in unique ways, being inadequate translation would also add significant also more idiosyncratic for some industries than longterm strategic and cost-centric consequences others. to the ongoing broken customer experience. Concrete models recently built for many industry Searching for further practical evidence on the segments by following the modularisation prin- emergence of non-traditional enterprise areas ciples reveal that there are hundreds of busi- needing process study, it is important to revisit ness components that the business process tra- in depth some theories of organisational design dition has failed to address. In fact, most pro- and related work by different business research cesses available from the research literature fall schools (Penrose 2009). Figure3 shows an or- in the category of operations involved in the ganisation of the resource-base of a typical en- last row of business components, i.e., production terprise into four distinct types and the corres- and maintenance processes. As the level of in- ponding bundling of such resources into disjoint volved resources moves into oversight and man- business components. Each column on the right agement, several interesting examples of cases hand side of the Fig. represents one typical com- may be found and used to illustrate the type petence whose organisation is described by the of operations at play. Going further into learn- generic concepts of the column on the left, as ing and innovation, traditional contributions fade presented in Sanz et al. (2012). Although a dif- quickly or disappear entirely. Interestingly, the ferent language was used, the foundations of top row of Fig.3 includes the ’strategy processe’ the structure of a generic competence should be that Van der Aalst and collaborators explicitly honored to Brumagin in Brumagim (1994), among excluded from their foundational work in the other more recent business researchers.7 early 2000’s. A diversity of processes like those 7This is probably the only known actionable model de- needed for controlling the quality of a cartoon in rived from the general and powerful concepts running un- an entertainment industry enterprise, managing der the denomination of Resource-Based View (RBV) in the pipeline of compounds in a pharmaceutical the theory of the firm. Business process researchers are strongly encouraged to delve into RBV, search for cross- business research topics such as those addressed in Organ- pollination with related Social Sciences work, and revisit isational Behaviour schools.

Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Enabling Front-Office Transformation and Customer Experience through Business Process Engineering 59

Figure 3: The four types of resources defined according to the different forms of behaviour that are observed in a generic enterprise (left). Componentised organisation of such resources based on different competences (right). Each of these components deals with a number of core subjects (Nandi and Sanz 2013) whose evolution is key for the definition of corresponding competences (columns in the picture) company, and disseminating the learning harves- and (ii) new languages are needed to close the ted from a specific family of consulting practices remarkable communication gap left in the cross- throughout a services enterprise should not be in- enterprise process space. It would be impossible cluded under the term ’strategy’. However, these to address these statements in full detail here oversight and management processes have not but it should suffice to say that loss of visibility been addressed by the BPM tradition. in cross-enterprise processes is a proven pain- point (Nandi and Sanz 2013) still yielding well- There may be still an argument that processes in identified performance and communication prob- classical BPM work aim at modelling operations lems in many firms. In other words, the "hundred across the components and not inside them , i.e., of pages" alluded by Dietz (2011) are real and the value streams end-to-end processes also called ’ ’ insight that these many pages have unraveled is in some business literature. However, this argu- minimal. ment does not necessarily follow from inspecting From a research perspective and practical point the work reported in more than one thousand papers in the last twelve years. The BPM tradition of view, the reader is referred to the recent work has adequately responded to the need of min- in Nandi and Sanz (2013) for evidence that the imising transaction costs across the enterprise main ’value streams’ across an enterprise are progressions of core subjects life- and builds upon existing governance mechan- in fact and not cycle of objects isms defined as true systems of control aligned , at least when the latter is un- statemachines with functions (Le Clair 2012). In that sense the derstood in the tradition of , i.e., traditional approach has followed closely the en- artifacts evolving through a number of micro- terprise disconnection and rigidity leading to states that separate the initiation and completion of "tasks". This fact goes back to the funda- the present state-of-the-art in customer exper- mental way metaphysics of processes has been ience. Moving the foundational basis to address approached in Social Sciences (Rescher 1996) and the next generation of business process (called the conceptual duality between process and sub- "hybrid connected processes" in Le Clair (2012)), jects8 in the organisation of the world of a gen- crossfunctional and complex processes (i) cannot be made or realised into workflow structures 8The word "subject" here means "theme" or "topic". This Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 60 Jorge L. Sanz

eric enterprise. Indeed, subjects are higher-level business process focus has to shift in order to abstractions than conventional objects and their contribute to other urgent goals in organisations. evolution is thus subjected to lots of asynchron- Business process is called to play as a key in- ous activity taking place across the enterprise. strument for achieving the customer experience The delivery of outcomes produced by these asyn- needed in front-office operations and deep end- chronous activities signals the completion of ne- toend integration of the latter with the back-office cessary results as agreed in pre-determined cross- in enterprises. The main motivation for the new functional commitments. These commitments are, work needed does not hinge around cost reduc- in fact, a form of organisational contracts and tion, industrialising routine operations or build- may be regarded as quite granular macro-states ing better software with BPM systems. in the evolution of an individual subject. These ’states’ are called milestones in Nandi and Sanz 4 Back to Process Foundations (2013). The need for aligning the research agenda in pro- The evolution of business process has not cess to the main challenges faced by industry happened without significant divergence and to was also called out in the closing recommenda- some extent, also confusion. The state-of-the- tions from the BPM study in Indulska et al. (2009): art is plagued by language chasms, cultural silos ". . . despite being an actively researched field, anec- and idiosyncratic viewpoints. Some of these chal- dotal evidence and experiences suggest that the lenges were documented in De Man (2009); Indul- focus of the research community is not always ska et al. (2009); Recker et al. (2009); Reijers et al. aligned with the needs of industry". A couple of (2010) and others. Reijers et al. (2010), state the years have elapsed since related papers were pub- challenge in clear terms: "Considerable confusion lished but the situation has not changed much. exists about what Business Process Management Reijers et al. (2010) also addressed the import- entails . . . ". Indeed, the definition of business ance of rooting BPM activities in industrial prac- process is still troubled by ambiguity and adding tice and correctly questioned the understanding the term "management" has done little to clarify of the actual adoption of BPM by organisations: the confusion. A plea for this clarity has been it may come as a surprise that contemporary ". . . articulated by Olbrich (2011): "It seems a pity that insights are missing into which categories of organ- a lot of current research fails to provide a basic izations are adopting BPM and which type of BPM definition of what underlying understanding of projects they are carrying out ". Actually, Aalst ’process’ and ’BPM’ it bases its work on". In fur- (2012) did some justice in his recent review of ther exchanges in the same S-BPM conference, research in the last decade of BPM Conferences other authors such as Fleischmann (2011); Singer and highlighted that this work mostly addressed and Zinser (2011) agreed that the problem goes automation concerns. In particular, Van der Aalst further into a lack of clarity on the very defin- revisited BPM systems as an opportunity to fur- ition of BPM. A review of the literature shows ther position BPM tools as valuable instruments that there is not a single and agreed definition to build better software applications. of these terms. While ". . . a scientific foundation While this traditional BPM research work and is missing" was clearly stated by Van der Aalst practices should definitely continue, new market back in 2003, the review of BPM Conferences pub- trends and needs from new enterprise capabil- lished by the author a decade later confirms that ities in the Front-Office strongly suggest that the fundamental shortfalls have not been overcome yet (Aalst 2012). The underlying reason differs from the interpretation of subject as an actor carrying out an activity, and thus, it should not be confused with is deeply related to the nature of business pro- related semantics in S-BPM. cess being a socio-technical system and thus, its Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Enabling Front-Office Transformation and Customer Experience through Business Process Engineering 61

complexity cannot be approached by a narrow fo- process. The depth of Van de Ven’s classification cus on technology dimensions. In Fleischmann’s reveals the foundations underlying many busi- own words: ". . . sociological systems like organ- ness process definitions. In spite of having been izations are combined with technical systems like published two decades ago, this work has gone information and communication technology. For a unnoticed in most of the BPM literature (Aalst holistic view of business process management we 2012; Aguilar-Savén 2004; Klein and Petti 2006; have to consider all aspects" (Fleischmann 2011). Ko et al. 2009; Lu and Sadiq 2007; Ould 1995b; Weske (2012) also highlights the deep nature of Propp 1968; Toussaint et al. 1998; Trkman 2010 process: "a business process consists of a set of and many others). activities that are performed in coordination in an organizational environment. These activities Another language chasm across different schools jointly realize a business goal." While using differ- of thought or communities of practice is the un- ent language, other authors also defined business clear relationship between the concept of busi- organisational routine processes (Davenport 1993; Debevoise 2007; Du- ness process and that of . mas et al. 2013; Indulska et al. 2009; Krogstie et al. Rich literature is available on the study of routines 2006; Ould 1995a; Smith and Fingar 2007 and the (Becker 2004), the significance of routines as a list goes on). unit of analysis for organisations (Levin 2002; Pentland and Feldman 2008; Pentland et al. 2012) The Object Management Group recognised the the collectivist meaning of routines and the need foundational problem with the definition of pro- for establishing solid micro-foundations (Felin cess. Siegel (2008), the leader of the BPM group and Foss 2004) among others. It is very likely that stated: "there is no agreed-upon industry defini- business process and routine address identical tion of Business Process. Instead, there are multiple concerns in organisation theory; however, in definitions, each looking at the field from its own spite of the prolific technical production in the unique point of view, concentrating on its own set two subjects during decades, their formal rela- of concerns". Certainly, it is not a matter of one tionship and the reasons for keeping two differ- definition being right and the others being wrong. ent terms remain unclear. Rather, the issue is about the varying points of view used. As a consequence, the main efforts in More recently, there has been a fundamental process modelling standardisation have not yet piece of work in process that builds upon a recon- yielded the expected outcomes, as discussed in ciled view of process and information available Recker et al. (2009), more broadly exposed in In- since the early days of the Information Engineer- dulska et al. (2009) and highlighted in Aalst (2012). ing schools in Europe. This approach to business Unquestionably, most people do have a similar process goes under the brand of entity-centric and informal notion of "business process". But operations modelling (Sanz 2011) and offers a hol- this intuitive agreement does not mean a conver- istic approach that reunites different types of gence across viewpoints. In fact, the variations processes under the same conceptual understand- in the definition of process may suggest that the ing. This entity-centric concept has been used term is a boundary object across disciplines, in- intensively by (Ould 1995a,b) and although the dividuals from different units of an organisation notion of life-cycle is from the early 1980’s, sev- or communities of practice. Other researchers eral important contributions has been made in in Social Sciences and Philosophy have also fo- different industries and software to merit a de- cused extensively on the concept of process and tailed inspection in Business Process Engineering its definition. Ven (1992) addressed the topic in (Bhattacharya et al. 2009; Cohn and Hull 2009; the context of one of the most complex types Nandi 2010; Nigam and Caswell 2003; Robinson of processes in organisations, i.e., the strategy 1979; Rosenquist 1982). Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 62 Jorge L. Sanz

Quite interestingly, another related approach was (A) Establish a foundation for understanding recently presented to model cross-functional end- and modelling the journeys that customers fol- to-end processes in enterprises based on the no- low in their multiple touchpoints when inter- tion of subjects and nexus of commitments (Nandi acting with enterprises across different chan- and Sanz 2013). The foundation for all this work nels. These journeys are probably the most appears as an important step toward the design looselycoupled type of processes, i.e., they are and construction of different process types, in- highly unstructured but they are not "random cluding the so-called value streams, by using a walks" at all as customers seek for specific common approach in which information does outcomes. This type of interactions is also not take back seat as a mere "after-thought" in found in other collaborative work in enter- the modelling of behaviour or becomes confused prises (Harrison-Broninski 2005). In addition, with "state model", being the latter a common as involved interactions combine and altern- misunderstanding incurred by most computer ate human-to-human and human-todigital con- scientists as Van der Aalst remarkably noted. The tacts, these journeys are rich in information point of reunion of these seemingly related mod- and behaviour. Then, their adequate under- elling techniques does not reside in "artifacts" standing is imperative for the next generation or "object life-cycle" but instead, it goes back to of customer experience. Some work has been the Social Sciences in the sense that the unifying done on this topic but there are no foundations concept is the very epistemology of process, i.e., yet with a theory that explicates the journeys "things in the making" (Tsoukas 2001; Tsoukas and how behaviour of the actors should be and Chia 2002). Consequently, process design is guided from footprints of customer contacts the evolution of a core subject about describing . and previous experiences. This is one of the While the roots of this approach come from sev- most fundamental research problems that dif- eral decades of work and different schools of ferent industries need to benefit from as its thought, not all process researchers and practi- value is directly related to customer loyalty. tioners seem familiar with these concepts and (B) Discover customer-enterprise co-creation related literature sources. mechanisms and have them reach a massive scale through innovative processes. This will 5 Research topics in Business Process support the social transformation necessary Engineering for the information coming from social data It would be difficult to propose here a complete to become a trustable source of actual beha- agenda of research and practice in Business Pro- viour and intent of individuals. While so- cess Engineering. Like in any other emerging cial media means a flood of useful data, in- field of work, only the pass of the time, com- ferring human intention and behaviour from munity activities and market consolidation will these sources remains illusory. Co-creation pro- determine its boundaries and shape its ultimate cesses deploying collaborative and mutually priorities. However, based on current work and beneficial practices appear essential for the ongoing industry needs, it would be safe to high- next generation of customer experience. Ex- light some important areas with the purpose of plicit provision of knowledge on an individual stimulating further research. could be then done in exchange for personalised services or some other form of tangible This is a first pass through such a list. Topics are value-propositions. This will lead individuals classified according to four basic categories: to provide trustworthy evidence of their beha- Customer-Enterprise Behaviour: Foundations and viour and intent. Designing and implement- Models ing the necessary processes to reach the scale Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Enabling Front-Office Transformation and Customer Experience through Business Process Engineering 63

needed requires deep socio-technical innova- (E) Reconcile the ever-deepening silos of Inform- tion. These processes will also help encourage ation and Process. As suggested by the differ- full transparency from consumers and enforce ent levels shown in Fig.4, the information and accountability from companies. The latter will process domains have traditionally evolved in help replace today’s legal disclaimers in which almost complete isolation from each other. As consumers are asked to resign their privacy damaging as this disconnection may result for rights under terms-and-conditions that prob- the well-being of any organisation, the probably few consumers read and even a fewer lem has stayed unresolved throughout several number of them understand. decades. In fact, the gaps have widened and (C) Create a "sociology of the customer" that got deeper as the new "business analytics" helps understand the effect of using mass pro- trend has been getting momentum in enterprises and gathering the attention of the Chief cesses even with individualised clients in the Marketing Officer. The introduction of "big pursuit of ’profitability’. If economic analysis data" and other marketing concepts in Inform- renders it viable, data footprints left by con- ation Management technology has continued sumers will not be the only hint to infer cus- to widen the chasm. Hopefully, by building on tomer behaviour (which is an erroneous ap- a new foundation where the Information Pro- proach to understand people’s needs and true cess in organisations and society is repurposed expectations anyway). Furthermore, the integ- as a single phenomenon through Business In- ration of process and big data will allow for formatics, new bridges will be built across the full operationalisation of "insight", thus mak- two silos. This reunion is dubbed "Deep Pro- ing the latter move from "interesting discov- cess meets Business Analytics" on Fig.4. The ery" to a Social Science-supported theory to need for this integration will reposition "pro- enhance services and provide enterprises with cess analytics" as the integration of on-line higher customer equity. (real-time) analytics and customer journeys. Front-Office Business Architecture (F) Provide data-only analytics and related statistical modelling with a better foundation (D) Propose complete Front-Office operational through behaviour-based causation. This models that represent the actual work enter- should help foster a blended approach through prises do with and for their own customers. "white box" Enterprise Engineering modelling This should include process and performance for today’s decision-making techniques based frameworks for all those key competences and on "black-box" statistics. Among other areas of capabilities in the enterprise that belong to the critical enterprise value, this topic should also Front- Office operations. In particular, the cre- help define an enterprise business perform- ation of solid Process Reference Architectures ance framework that integrates behaviour and for emerging operational areas in marketing, data in organisations. This goal corresponds brand management, campaign management, to achieving the important integration shown etc. would be critical for accelerating industry in the top level of Fig.4. value of new research. As suggested earlier in (G) Develop a theory of Process Modularisation this paper, surfacing and documenting these that is consistent and evolvable with change. new workpractices is essential. Software pack- This work has been initiated by different col- ages are already in the market and these ap- leagues in (Nuffel 2011). As the "unit of change" plications bury important processes whose fre- in Process gets progressively more clear, the quent change is imperative for flexibility of topic of Process Evolvability will also become Front-Office operations. connected to modularisation, thus addressing Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 64 Jorge L. Sanz

Process and INFORMATION PROCESS Business Analycs SILO SILO Integrated

Value-Driven Enterprise Scorecards Process Performance Management Management Intelligent Business Big Data Operaonal Operaons and Massive Analycs Process Innovaon Systems

Analycs Capabilies Connued Process Adapve / Agile for LOBs Improvements and (Transaconal) Flexibility Operaons

Business Intelligence Cross-Funconal Business Visibility (from ERPs) Processes Modeling (BVM)

Business Enes Process Flows Data Modeling & and ERPs Execuon (BELA)

Deep Integraon

Figure 4: Silos in information and process management have deepened with the evolution of each domain. This gap is more notorious after the advent of business analytics, scorecards, performance management and value-driven process BPM

the need for managing combinatorial effects (I) Benefit from Enterprise Engineering princip- (as already addressed by the general principles les to reposition the role of humans in the of Normalised Systems Theory in (Mannaert value-creation of Front-Office business areas. and Verelst 2009) for the case of software sys- This topic has several deep social connota- tems). tions and should include the provisioning of (H) Clarify the distinction, if any, between the economic evidence of the scalability (or lack Social Science concept of organisational routine thereof) of human-centric methods for under- (Pentland et al. 2012) and the broader meaning standing individual behaviour of customers. of process coming from Business Process En- Industry-Oriented Content gineering. This will help reconcile work across the different schools of research in Social and (J) Create industry-specific multi-channel cus- Computer Sciences. While practitioners sel- tomer journeys for key services industries such dom use the word "routine" (and when they as banking, insurance, retail and telecommu- do, they imply repetitive or boring tasks which nications. Link to and support these customer is not the meaning in Social Sciences), it is journeys with knowledge-based representa- important to benefit from cross-insemination tions that bridge process and knowledge man- between Enterprise Engineering and Social Sci- agement. This is a significant area of work ence research for better understanding of or- that will pave new integration of Process with ganisational design through deep behaviour Knowledge Management by creating a cus- research. tomer-centric knowledge-based organisation of Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Enabling Front-Office Transformation and Customer Experience through Business Process Engineering 65

the enterprise. The meaning of the latter state- 7 Acknowledgements ment is about making all pertinent information from an enterprise to be organised and be The author would like to express his gratitude made available to customers in new, intelligent to many colleagues for inspiring discussions on ways in which "process footprints" serve as a historical base to reorganise and find informa- business process in the context of Enterprise En- tion personalised to individual customers (this gineering, Business Informatics and practical con- comment comes from a private communica- notations of all these domains. A few names are tion with P. Nandi). mentioned here: J. Verelst, H, Mannaert, J. Di- Tooling etz, E. Proper, A. Albani, P. Nandi, M. Indulska, J. Sapperstein, J. Sphorer, H. Hastings, M. Cefkin, (K) Propose new tools that further the current J. Barjis, B. Hofreiter, C. Huemer, J. Tribolet, V. state-of-the-art of Information Technology for Becker, L. Flores and many others. Sincere thanks process design and construction in the concert are also due to two referees for their invaluable of a Business Process Engineering approach comments. (in this connection, the generation of code is a secondary concern but flexible and open endto- end integrated capabilities would be a break- References through). These process tools will be the car- rier of data analytics in real-time while sup- van der Aalst W. M. P., Berens P. J. S. (2001) Bey- porting the delivery of personalised services ond Workflow Management: Product-driven to individual customers. Case Handling. In: Proceedings of the 2001 In- 6 Conclusions ternational ACM SIGGROUP Conference on Business Process has left the productivity corner Supporting Group Work. GROUP ’01. ACM, where it has been confined by "scientific manage- Boulder, Colorado, USA, pp. 42–51 http://doi. ment". With the advent of customer-enterprise acm.org/10.1145/500286.500296 interactions of all forms and exercised through van der Aalst W. M. (2004) Business Process multiple channels, the need for a significantly Management Demystified: A Tutorial on improved customer experience is an imperative Models, Systems and Standards for Workflow in transforming front-office operations. Conven- Management. In: Desel J., Reisig W., Rozen- tional approaches to process have proven to have berg G. (eds.) Lectures on Concurrency and a devastating effect on loyalty. Renewed research Petri Nets. Lecture Notes in Computer Sci- and professional efforts to approach process as ence Vol. 3098. Springer, Berlin, Heidelberg, part of complex social systems are a must to cope pp. 1–65 http://dx.doi.org/10.1007/978-3-540- Fig.4. Silos in information and process manage- 27755-2_1 ment have deepened with the evolution of each van der Aalst W. M. (2012) A Decade of Busi- domain. This gap is more notorious after the ad- ness Process Management Conferences: Per- vent of business analytics, scorecards, perform- sonal Reflections on a Developing Discipline. ance management and value-driven process BPM In: Barros A., Gal A., Kindler E. (eds.) Busi- with the challenges faced in those competences of ness Process Management. Lecture Notes in enterprises dealing with customers, particularly Computer Science Vol. 7481. Springer, Berlin, in the business-toconsumer industries. Business Heidelberg, pp. 1–16 http://dx.doi.org/10. Process Engineering is a new domain of work 1007/978-3-642-32885-5_1 that attempts to make the past IT-centric view van der Aalst W. M., van Hee K. M. (2002) Work- of process into a multidisciplinary area of both flow Management: Models, Methods, and Sys- institution and practice knowledge. tems. MIT Press, Cambridge, MA, USA Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 66 Jorge L. Sanz

van der Aalst W. M., ter Hofstede A., cesses. In: Cardoso J., van der Aalst W. (eds.) Kiepuszewski B., Barros A. (2003) Workflow Handbook of Research on Business Process Patterns English. In: Distributed and Parallel Modeling. IGI Global Snippet, Hershey, PA, Databases 14(1), pp. 5–51 http://dx.doi.org/10. USA, pp. 503–531 1023/A%3A1022883727209 Bliss J. (2006) Chief Customer Officer: Getting van der Aalst W. M., Weske M., Grünbauer D. Past Lip Service to Passionate Action. John (2005) Case handling: a new paradigm for Wiley & Sons business process support. In: Data & Know- Brumagim A. L. (1994) A hierarchy of corporate ledge Engineering 53(2), pp. 129–162 resources. In: Advances in strategic manage- Aguilar-Savén R. S. (2004) Business process mod- ment 10(Part A), pp. 81–112 elling: Review and framework. In: Interna- Capgemini (2012) World Retail Banking Report tional Journal of Production Economics 90(2) Center) A. A. P. Q. Process Classification Frame- Production Planning and Control, pp. 129 – work (PCF). www.apqc.org 149 http://www.sciencedirect.com/science/ Cohn D., Hull R. (2009) Business Artifacts: A article/pii/S0925527303001026 Data-centric Approach to Modeling Business Albani A., Dietz J. L. (2011) Enterprise ontology Operations and Processes. In: IEEE Data Eng. based development of information systems. Bull. 32(3), pp. 3–9 In: International Journal of Internet and En- Davenport (1993) Process Innovation: Reengin- terprise Management 7(1), pp. 41–63 eering Work Through Information Techno- Andjelkovic-Pesic M. (2007) Six Sigma philo- logy. Harvard Business School Press, Boston, sophy and resource-based theory of compet- MA, USA itiveness: An integrative approach. In: Facta Davenport T. H. (2005) The coming commod- universitatis-series: Economics and Organiz- itization of processes. In: Harvard Business ation 4(2), pp. 199–208 Review 83(6), pp. 100–108 Andjelković Pešić M. (2004) Six Sigma project De Man H. (2009) Case Management: A Re- implementation. In: YUPMA – VIII Interna- view of Modeling Approaches. http://www. tional Symposium in Project Management, bptrends.com pp. 191–197 Debevoise T. (2007) Business Process Manage- Andjelković Pešić M. (2006) Enterprises of New ment with a Business Rules Approach: Imple- Economy – High Performance Organizations. menting the Service Oriented Architecture. In: Business Politics, pp. 54–58 Tipping Point Solutions Aveiro D., Silva A. R., Tribolet J. M. (2011) To- Dietz J. L., Hoogervorst J. A., Albani A., Aveiro wards a G.O.D. theory for Organisational En- D., Babkin E., Barjis J., Caetano A., Huysmans gineering: modelling the (re)Generation, Op- P., Iijima J., Kervel S. J. V. et al. (2013) The eration and Discontinuation of the enterprise. discipline of enterprise engineering. In: In- In: International Journal of Internet and En- ternational Journal of Organisational Design terprise Management 7(1), pp. 64–83 and Engineering 3(1), pp. 86–114 Albani A., Barjis J., Dietz J. (eds.) Advances Dietz J. (2011) A Business Process is More Than in Enterprise Engineering III. Lecture Notes a Workflow. Tutorial at KEOD, Paris in Business Information Processing Vol. 34. Dumas M., Rosa M., Mendling J., Reijers H. Springer (2013) Fundamentals of Business Process Becker M. C. (2004) Organizational routines: a Management. Springer review of the literature. In: Industrial and Eid-Sabbagh R.-H., Dijkman R., Weske M. (2012) Corporate Change 13(4), pp. 643–678 Business Process Architecture: Use and Cor- Bhattacharya K., Hull R., Su J. (2009) A data- rectness. In: Barros A., Gal A., Kindler E. centric design methodology for business pro- (eds.) Business Process Management. Lec- Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Enabling Front-Office Transformation and Customer Experience through Business Process Engineering 67

ture Notes in Computer Science Vol. 7481. Springer, Berlin, Heidelberg, pp. 501–514 Springer, Berlin, Heidelberg, pp. 65–81 http: http://dx.doi.org/10.1007/978-3-642-02144- //dx.doi.org/10.1007/978-3-642-32885-5_5 2_39 Felin T., Foss N. (2004) Organizational Routines: Khoyi D. (2010) Data Orientation. In: Swen- A Sceptical Look. DRUID Working Paper 04- son K. D. (ed.) Mastering the Unpredictable. 13. Danish Research Unit for Industrial Dy- Meghan-Kiffer, Tampa, FL, USA namics (DRUID) Klein M., Petti C. (2006) A handbook-based Fleischmann A. (2011) Do We Need to Re- methodology for redesigning business pro- think Current BPM Research Issues? In: cesses. In: Knowledge and Process Manage- Fleischmann A., Schmidt W., Singer R., Seese ment 13(2), pp. 108–119 http://dx.doi.org/10. D. (eds.) Subject-Oriented Business Process 1002/kpm.248 Management. Communications in Computer Ko R., Lee S., Lee E. (2009) Business process and Information Science Vol. 138. Springer, management (BPM) standards: a survey. In: Berlin, Heidelberg, pp. 216–219 Business Process Management Journal 15(5), Franz P., Kirchmer M. (2012) Value-Driven Busi- pp. 744–791 ness Process Management: The Value-Switch Krogstie J., Sindre G., Jørgensen H. (2006) for Lasting Competitive Advantage. McGraw- Process models representing knowledge for Hill Education http://books.google.at/books? action: a revised quality framework. In: id=DYrOiwThP4UC European Journal of Information Systems Harrison-Broninski K. (2005) Human Interac- 15(1), pp. 91–102 tions: The Heart and Soul of Business Process Le Clair C. (2012) Stuck In Cement: When Pack- Management: how People Really Work and aged Apps Create Barriers To Innovation. how They Can be Helpful to Work Better. http://blogs.forrester.com Meghan-Kiffer, Tampa, FL, USA Levin D. Z. (2002) When Do Organizational Hastings H., Saperstein J. (2007) Improve Your Routines Work Well? A New Approach to Marketing to Grow Your Business: Insights Knowledge Management. Rutgers University, and Innovation That Drive Business and Newark, USA Brand Growth, First. Wharton School Pub- Lu R., Sadiq S. (2007) A Survey of Comparat- lishing ive Business Process Modeling Approaches. Houy C., Fettke P., Loos P., Aalst W., In: Abramowicz W. (ed.) Business Informa- Krogstie J. (2010) BPM-in-the-Large To- tion Systems. Lecture Notes in Computer Sci- wards a Higher Level of Abstraction in Busi- ence Vol. 4439. Springer, Berlin, Heidelberg, ness Process Management. In: Janssen M., pp. 82–94 http://dx.doi.org/10.1007/978-3- Lamersdorf W., Pries-Heje J., Rosemann M. 540-72035-5_7 (eds.) E-Government, E-Services and Global Mannaert H., Verelst J. (2009) Normalized sys- Processes. IFIP Advances in Information tems: re-creating information technology and Communication Technology Vol. 334. based on laws for software evolvability. Springer, Berlin, Heidelberg, pp. 233–244 Koppa Digitale Media BVBA, Kermt, Belgium http://dx.doi.org/10.1007/978-3-642-15346- Miers D. (2009) Process Architecture: The Key to 4_19 Effective Process Relationships. Presentation Indulska M., Recker J., Rosemann M., Green from BPM Focus P. (2009) Business Process Modeling: Cur- Nandi P. (2010) Introducing Business Entities rent Issues and Future Challenges. In: and the Business Entity Definition Language Eck P., Gordijn J., Wieringa R. (eds.) Ad- (BEDL): A first class representation of data vanced Information Systems Engineering. for BPM applications. http://www.ibm.com Lecture Notes in Computer Science Vol. 5565. Nandi P., Sanz J. (2013) Cross-Functional Op- Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 68 Jorge L. Sanz

erations Modeling as a Nexus of Commit- EEWC2013, Luxembourg, May 13-14, 2013, ments: A New Approach for Improving Busi- Proceedings. Lecture Notes in Business In- ness Performance and Value-Creation. In: formation Processing Vol. 146. Springer, Ber- IEEE 15th Conference on Business Inform- lin, Heidelberg atics (CBI 2013). IEEE, pp. 234–241 Propp V. (1968) Morphology of the Folktale Scott Näslund D. (2008) Lean, six sigma and lean L., Wagner L. (eds.), Second. University of sigma: fads or real process improvement Texas Press, p. 184 methods? In: Business Process Management Rawson A., Duncan E., Jones C. (2013) The truth Journal 14(3), pp. 269–287 about customer experience. In: Harvard Busi- Nigam A., Caswell N. S. (2003) Business artifacts: ness Review 91(9), pp. 90–98 An approach to operational specification. In: Recker J., Rosemann M., Indulska M., Green P. IBM Systems Journal 42(3), pp. 428–445 (2009) Business Process Modeling - A Com- van Nuffel D. (2011) Towards designing modular parative Analysis.. In: Journal of the Associ- and evolvable business processes. PhD thesis, ation for Information Systems 10(4) Universiteit Antwerpen Reijers H. A., Wijk S., Mutschler B., Leurs M. Olbrich T. J. (2011) Why We Need to Re- (2010) BPM in Practice: Who Is Doing What? think Current BPM Research Issues. In: In: Hull R., Mendling J., Tai S. (eds.) Busi- Fleischmann A., Schmidt W., Singer R., Seese ness Process Management. Lecture Notes in D. (eds.) Subject-Oriented Business Process Computer Science Vol. 6336. Springer, Berlin, Management. Communications in Computer Heidelberg, pp. 45–60 http://dx.doi.org/10. and Information Science Vol. 138. Springer, 1007/978-3-642-15618-2_6 Berlin, Heidelberg, pp. 209–215 http://dx.doi. Rescher N. (1996) Process Metaphysics: An Intro- org/10.1007/978-3-642-23135-3_12 duction to Process Philosophy. SUNY series Ould (1995a) Business Processes: Modelling and in philosophy. State University of New York Analysis for Re-Engineering and Improve- Press ment. John Wiley & Sons Robinson K. (1979) An entity/event data model- Ould (1995b) Business processes: modelling and ling method. In: The Computer Journal 22(3), analysis for re-engineering and improvement. pp. 270–281 Wiley, p. 224 Rosenberg A., von Rosing M., Chase G., Omar Ould M. A. (1997) Designing a re-engineering R., Taylor J. (2011) Applying Real-World BPM proof process architecture. In: Business Pro- in an SAP Environment, 1st. SAP Press cess Management Journal 3(3), pp. 232–247 Rosenquist C. (1982) Entity Life Cycle Models Penrose E. (2009) The Theory of the Growth of and their Applicability to Information Sys- the Firm. Oxford University Press tems Development Life Cycles: A Framework Pentland B. T., Feldman M. S. (2008) Designing for Information Systems Design and Imple- routines: On the folly of designing artifacts, mentation. In: The Computer Journal 25(3), while hoping for patterns of action. In: In- pp. 307–315 formation and Organization 18(4), pp. 235– Sanz J. L. (2011) Entity-centric operations mod- 250 eling for business process management - A Pentland B. T., Feldman M. S., Becker M. C., Liu multidisciplinary review of the state-of-the- P. (2012) Dynamics of organizational routines: art. In: IEEE 6th International Symposium on a generative model. In: Journal of Manage- Service Oriented System Engineering (SOSE ment Studies 49(8), pp. 1484–1508 2011). IEEE, pp. 152–163 Proper H., Aveiro D., Gaaloul K. (2013) Ad- Sanz J. L., Leung Y., Terrizzano I., Becker V., vances in Enterprise Engineering VII: Third Glissmann S., Kramer J., Ren G.-J. (2012) Enterprise Engineering Working Conference, Industry Operations Architecture for Busi- Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Enabling Front-Office Transformation and Customer Experience through Business Process Engineering 69

ness Process Model Collections. In: Daniel F., Trkman P. (2010) The critical success factors of Barkaoui K., Dustdar S. (eds.) Business Pro- business process management. In: Interna- cess Management Workshops. Lecture Notes tional Journal of Information Management in Business Information Processing Vol. 100. 30(2), pp. 125–134 Springer, Berlin, Heidelberg, pp. 62–74 http: Tsoukas H. (2001) What Does it Mean to be Sens- //dx.doi.org/10.1007/978-3-642-28115-0_7 itive to Process? Keynote address European Schönthaler F., Vossen G., Oberweis A., Karle Group Organization Studies Colloquium 17, T. (2012) Business Processes for Business Lyon, France Communities: Modeling Languages, Methods, Tsoukas H., Chia R. (2002) On organizational Tools. Springer becoming: Rethinking organizational change. Sharp A., McDermott P. (2009) Workflow Model- In: Organization Science 13(5), pp. 567–582 ing: Tools for Process Improvement and Ap- Van de Ven A. H. (1992) Suggestions for study- plications Development, First. Artech House, ing strategy process: a research note. In: Stra- Inc., Norwood, MA, USA tegic management journal 13(5), pp. 169–188 Sidorova A., Isik O. (2010) Business process Villanueva J., Hanssens D. M. (2007) Cus- research: a cross-disciplinary review. In: tomer equity: Measurement, management Business Process Management Journal 16(4), and research opportunities. Foundations and pp. 566–597 Trends in Marketing 1 Vol. 1. Now Publishers Siegel J. (2008) In OMG’s OCEB Certification Inc, Hanover, MA, USA, pp. 1–95 Program, What is the Definition of Business Weske M. (2012) Business Process Management: Process? Certification Whitepaper, OMG Concepts, Languages, Architectures, Second. Singer R., Zinser E. (2011) Business Process Springer Management Do We Need a New Research White S. (2004) Process modeling notations and Agenda? In: Fleischmann A., Schmidt W., workflow patterns. http://www.bptrends. Singer R., Seese D. (eds.) Subject-Oriented com Business Process Management. Communic- Yan Z., Dijkman R., Grefen P. (2012) Business ations in Computer and Information Sci- process model repositories–Framework and ence Vol. 138. Springer, Berlin, Heidelberg, survey. In: Information and Software Techno- pp. 220–226 http://dx.doi.org/10.1007/978-3- logy 54(4), pp. 380–395 642-23135-3_14 Smith H., Fingar P. (2007) Business Process Man- Jorge L. Sanz agement: The Third Wave. Meghan-Kiffer Press Business Analytics Center Stone D., Devine J. (2013) From Moments to School of Computing – Business School Journeys: A Paradigm Shift in Customer Ex- National University of Singapore perience Excellence Singapore Swenson K. D. et al. (2010) Mastering the Unpre- dictable: How Adaptive Case Management Will Revolutionize the Way that Knowledge Workers Get Things Done. Meghan-Kiffer, Tampa, FL, USA Toussaint P. J., R Bakker A., Groenewegen L. P. (1998) Constructing an enterprise viewpoint: evaluation of four business modelling techniques. In: Computer methods and programs in biomedicine 55(1), pp. 11–30 Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 70 Eng K. Chew

Eng K. Chew

Service Innovation for the Digital World

The foundational principles and conceptual building blocks of customer-centric service innovation (SI) practice are explained, and a resultant integrated framework of SI design practices for customer value co-creation is synthesised. The nexus of service strategy, service concept and business model is identified to assure SI commercialisation. The requisite SI models and processes to systematise the innovation practice are reviewed. The emergent practices of customer and community participation, in a digital world, across the firm’s entire SI lifecycle are explicated, together with the requisite strategic management practices for successful service innovation.

1 Introduction (2009); Hastings and Saperstein (2013); Ordanini and Parasuraman (2011). It focuses, in particu- Service innovation – the art and science of cre- lar, on the various critical roles of customers in ating innovative services that customers value value co-creation for themselves in conjunction and are willing to pay for – in the digital world with the service provider and their network of exemplifies many of the fundamental challenges partners. of business informatics. Recent studies of service innovation have focused on the effective The paper is structured as follows. Section2 de- management of service innovation to enhance scribes in detail the fundamental building blocks firm performance – such as the importance of of service innovation: service dominant logic, ser- managing inter-organisational relationships and vice systems, operant resources and dynamic cap- commitments (Eisingerich et al. 2009), the ante- abilities, value networks, and finally, customer cedents and consequences of service innovation value co-creation – the ultimate purpose of ser- (Ordanini and Parasuraman 2011), and a prelim- vice innovation. Section3 synthesises from the inary service-thinking framework for value cre- extant literature a framework of design practices ation (Hastings and Saperstein 2013). These stud- for service innovation, comprising four business ies have shown that success in service innova- strategy-aligned interrelated practices of service tion requires "service thinking" (and attendant conceptualisation, service design, customer ex- service culture) and is contingent on effective perience and value creation, and service architec- collaboration with the firm’s customers and part- ture which, collectively, are typically pursued by ners in the overall innovation process. These designers iteratively (experimentally) and hol- authors also concur that service innovation is istically. Section4 links the design practices about the creation of customer value (Grawe et framework to service strategy on one hand and al. 2009). However, the art and science of design- business model design on the other to address ing and managing service innovations, especially the commercialisation aspect of service innova- for the digital world, remains an under-explored tions. Section5 reviews, individually, the com- research area. This paper seeks to contribute mon and foundational service innovation func- to filling this void by exploring the extant liter- tional models (in terms of the ’scope’ of and the ature to identify the critical constitutive theor- ’competence-based approach’ to service innov- ies and practices that would lead to successful ation) and processes (in terms of new service service innovation in line with Eisingerich et al. development) for the creation of customer value. Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Service Innovation for the Digital World 71

Section6 and Sect.7, respectively, review the in- view constitutes the service-dominant logic (S- creasingly important ’open innovation’ practices DL) which defines service as a process of apply- of involving customers and online community ing the competencies and skills of a provider for in the end-to-end service innovation process in the benefit of, and in conjunction with, the cus- the digital world, while Sect.8 addresses the re- tomer (Vargo and Lusch 2004, 2008). A service quisite strategic management capability to en- offering is produced using the firm’s resources sure service innovation success. Finally, Sect.9 including both tangible (such as goods) and in- concludes the paper by summarising the requis- tangible (such as knowledge, competence and ite principles (theories) and service design and relationship) assets (Arnould 2008). The value innovation management practices for service in- characteristics of the service provisioned, how- novation excellence. ever, are co-created through the interactions of the client’s competences with that of the service 2 Conceptual Building Blocks provider (Gallouj 2002). Thus the client is active in a service interaction; it co-creates value 2.1 Service Dominant Logic (for itself) with the provider (Fitzsimmons and Fitzsimmons 2010; Gadrey and Gallouj 2002; Gal- In the early days (pre-1980) of services marketing, louj 2002). The central idea of S-D logic is that services were seen as a special kind of products. "exchange is about the process of parties doing Seen as a unit of production output, services were things for and with each other, rather than trad- defined as residues of, and thus subordinate to, ing units of output, tangible or intangible" (Vargo products (Lovelock and Gummesson 2004; Vargo and Lusch 2008). and Lusch 2004). From this goods production perspective, services as an output are seen to possess 2.2 Service Systems four so-called IHIP characteristics which are dis- tinctly different from physical products: Intangib- Service systems are the basic unit of analysis of ility, Heterogeneity, Inseparability and Perishabil- (the customer-centric view of) service (Maglio ity (Lovelock and Gummesson 2004). Intangibility and Spohrer 2008). A service system is defined as refers to the services output as being intangible. a complex adaptive system of people, and tech- Heterogeneity refers to the services possessing nologies working together to create value for its variable input resources and performance out- constituents (Spohrer et al. 2007). For example, comes. Inseparability refers to the production a telecom company is a complex market-facing and consumption of services occurring simultan- technology-based service system. The study of eously. Perishability refers to the services output service systems is focused on creating a basis as being non-durable and non-storable. How- for systematic service innovation (University of ever, these services characteristics were actually Cambridge and IBM 2007). It requires a mul- shown to be not generally applicable to all ser- tidisciplinary integrative understanding of the vices (Lovelock and Gummesson 2004). Leading ways organisation, human, business and tech- service scholars around the globe also regard nology resources and shared information may the production-oriented IHIP view as outdated be combined to create different types of service (Edvardsson et al. 2005), because it fails to cap- systems; and how the service systems may inter- ture the processual, interactive and relational act and evolve to co-create value (Maglio et al. nature of service co-creation and consumption as 2009). A service system has a service provider seen from the customer perspective (Edvardsson and a service client or beneficiary (Maglio et al. et al. 2005; Fitzsimmons and Fitzsimmons 2010). 2006). Service systems are connected by value This alternative customer-centric and relational propositions (Maglio et al. 2009). IT or business Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 72 Eng K. Chew

process outsourcing service configurations nego- an operant resource (i.e., intangible knowledge- tiated and agreed to between service providers based capability) "which acts on other operand and clients are examples of service systems. Con- or operant resource to produce an effect" (Vargo sistent with S-DL, value-cocreation interactions and Lusch 2004). Operant resources are dynamic, between service systems are service interactions, which include competences or capabilities that each comprising three main activities: propos- can be nurtured and grown in some unique ways ing a value-cocreation interaction to another ser- to provide competitive advantage to firms vice system (proposal); agreeing to the proposal (Madhavaram and Hunt 2008). Operant resources (agreement); and realising the proposal (realisa- that are valuable, rare, inimitable and not sub- tion) (Maglio et al. 2009). stitutable will generate sustainable competitive advantage for firms. For example, market orient- Service systems are dynamic, constantly compos- ation – i.e., market sensing and customer linking ing, recomposing and decomposing over time. A capabilities – is an operant resource that would service system operates as an open system cap- create that advantage (Arnould 2008). This mo- able of improving the state of another system tivates firms to create and use dynamic operant through sharing or applying its resources (in- resources to sustain the competitive advantage. cluding competences/capabilities), and improving its own state by acquiring external resources Highly innovative firms possess "masterfully de- (Maglio et al. 2009). Thus, service systems engage veloped" operant resources accumulated over a in knowledge-based interactions to co-create long period from institutionalised learning prac- value, where value is derived and determined tices (Madhavaram and Hunt 2008). These re- in use – the integration and application of resources allow the firm to effectively manage co- sources in a specific context embedded in firm’s evolution of knowledge, capabilities, and products output – and captured by price (Vargo et al. 2008). or services to sustain its competitive advantage. Consequently, advances in service innovation are Collaborative competence is identified as a pivotal only possible when a service system has inform- operant resource for sustained service innovation ation about the capabilities and the needs of its (Lusch et al. 2007) – one that assists in the devel- clients, its competitors and itself (Maglio et al. opment of two additional meta-competences: ab- 2009). sorptive competence, and adaptive competence (also collectively known as dynamic capabilit- Integral to and as a consequence of service innov- ies (Teece 2007)) which enable the firm to, re- ation, service systems co-create value through spectively, absorb new knowledge and inform- collaboration and adaptation, and establish a bal- ation from partners, and adapt to the complex anced and interdependent framework for sys- and turbulent environments by reconfiguring its tems of reciprocal service provision. Service resources (and organisational capabilities) with systems survive, adapt, and evolve through ex- those of the external partners. These operant re- change and application of resources (especially sources are key components of a service system knowledge and skills -operant resources as ex- which is conceptualised as a resource integrator plained below) with other systems (Vargo et al. (Spohrer et al. 2007). It is the people’s unique 2008). knowledge and skills and dynamic capabilities that make service systems adaptive to and sus- 2.3 Operant Resources & Dynamic tainable with the changing market environments Capabilities (Spohrer et al. 2007; Teece 2007; Vargo et al. 2008).

A resource is called an operand resource (i.e., 2.4 Value Networks of Digital World tangible physical resource) "on which an opera- In the increasingly digital world, information tion or act is performed to produce an effect", or technologies are "liquefying" physical assets into Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Service Innovation for the Digital World 73

information resources, and transform a service ates as an open system (Maglio et al. 2009). Firms firm into a value-creating service system in which must practice open innovation (Chesbrough 2003) a constellation of economic actors (customers, and develop systems integration capability (Ches- suppliers, business partners and the like) are able brough and Davies 2010) as part of its dynamic to seamlessly collaborate to co-create value (Nor- capabilities (Teece 2007) to integrate the requisite mann and Ramirez 1993). This reflects the S-D competences and resources from external sources logic’s commitment to explicating the firm’s col- with their own to co-create value; e.g., Apple’s laborative processes with customers, partners, creation of the iPod/iTune music service. and employees to engage in the co-creation of Value co-creation and innovation in the digital value through reciprocal service provision (Lusch world would require firms to institute individu- et al. 2007). And the customer is regarded as an alised and immediate customer feedback (to and operant resource – a dynamic resource that is from the customers) to engender customer and capable of acting on other resources to create organisational learning (Johannessen and Olsen value for itself (Vargo and Lusch 2008). 2010). This requires a new IT-enabled organisa- With the ubiquitous digitalisation, goods are in- tional logic which encompasses modular (multi- creasingly being embedded with microprocessors sourcing) flexibility, front-line (customer learn- and intelligence and becoming versatile platforms ing) focus, IT-enabled individualisation and "con- for service provision with enhanced customer nect and develop" innovation practices (Ches- and supplier insights and superior self-service brough and Davies 2010; Johannessen and Olsen ability. It also reduces transport and commu- 2010). In addition, the firm needs new cooper- nications costs, enhances the ability to interact ation structures by partaking in global compet- directly with customers and suppliers and con- ence clusters and practising coopetition (Johan- sequently coordination between firms becomes nessen and Olsen 2010). Above all, to be agile more efficient and responsive (Lusch et al. 2009). and adaptable as they learn of changing customer Thus, the firm will become an essential service needs, firms need to develop dynamic operant provisioning agent (actor) in a complex and ad- resources – the dynamic capabilities (Teece 2007). aptive value network comprising a spatial and The dynamic capabilities allow firms to continu- temporal structure of loosely coupled value-pro- ally align their competences to create, build and posing social and economic actors. The actors maintain relationships with (thus the value pro- interact through institutions and technology cap- positions to) customers (the ultimate source of able of spontaneously sensing and responding revenue) and suppliers (the source of resource inputs). via their dynamic capabilities to co-produce service offerings, exchange service offerings, and 2.5 Customer Value Co-creation finally co-create value. They are linked together by means of competences, relationships, and in- Customer is at the heart of value creation and ser- formation (Lusch et al. 2009). The relationships vice is about relationship with the customer (Ed- are collaborative and guided by non-coercive gov- vardsson et al. 2005). The customer interacts with ernance. This implies voluntary, reciprocal use the service provider via the interface through of resources for mutual value creation by two or which information /knowledge, emotions and ci- more interacting actors, through the symmetric vilities are exchanged to co-create value (Gallouj exchange of information and resources (compet- 2002). Value is wholly determined by the cus- ences) (Vargo et al. 2008). So in the value network, tomer upon, and in the context of, service usage customers and suppliers become partners, and (and customer experience), in which the compet- competitors become collaborators as well (Ches- ence (operant resource) of the provider is integ- brough and Davies 2010). Each firm (actor) oper- rated with the competence (operant resource) of Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 74 Eng K. Chew

the customer to (perform ’a job’ to) create (busi- 2005). From a service system viewpoint, value, ness) value with the customer (Edvardsson et al. created as a result of integrating the provider’s 2005; Vargo and Lusch 2008). The service pro- resources with the client’s, increases the client vider cannot deliver value, but only offer value system’s adaptability and survivability to fit with propositions (Vargo and Lusch 2008). To win the its changing environment (Vargo et al. 2008). service game, the value proposition must consistently meet the customer expectations and beha- 3 Framework of Design Practices vioural needs (Schneider and Bowen 2010). This can be assured by co-opting the customer com- To create innovative services that sustainably co- petence in co-creating the service offering with create superior customer value in the constantly the provider (Prahalad and Ramaswamy 2000)– evolving value networks of the digital world, a e.g., user toolkits for innovation (Hippel 2001). design framework is synthesised from the ex- However, the customer would collaborate with tant literature consistent with the preceding con- the provider in co-creation of core service offer- ceptual building blocks. The design framework ings only if they would gain benefits, such as: for service innovation consists of closely inter- expertise, control, physical capital, risk taking, related practices of: (a) service concept which psychic benefits, and economic benefits (Lusch defines what the service is and how it satisfies et al. 2007). customer needs, (b) service design which defines the service delivery mechanisms to consistently Service innovation must therefore be concerned satisfy customer needs, (c) customer experience with effectiveness of value co-creation between and value creation which guides service design the provider and beneficiary. It recognises the to align the provider’s competences and learn- principle that a proposed value by the provider, ing regime to those of the customers to ensure in the context of the client, is actually a compos- superior experience, and (d) service architecture ite of benefits and burdens (or costs), which can which systematises service design and innova- be evaluated using a customer value equation tion. These four interrelated practices are de- (Fitzsimmons and Fitzsimmons 2010). Burdens re- tailed below individually, but are typically prac- late to the service’s usability (or its relative ease- tised in the real-world iteratively and holistically. of-integration with the client’s resources or activities to "perform the job the service is hired to do") – the more user-friendly it is the less the bur- 3.1 Service Concept den and the greater the user experience; and the A service concept defines the conceptual model greater the customer efficiency (Xue and Harker of the service. It describes what the service is 2002). Thus, the most compelling service with and how it satisfies customer needs (Bettencourt the best "value for money" to the client is one 2010). Service concept is the most critical com- that has the largest "benefit-to-costs" ratio. This ponent of service strategy, and reflects the align- suggests that user involvement in co-creating the ment of the customer needs (job and outcome op- service offerings (or co-designing the value pro- portunities) with the company capabilities. Ser- positions) with the provider would more likely vice concept also forms the fundamental part create ’fit-for-purpose’ service for the client and of service design, service development and ser- thereby maximising the benefit. vice innovation (Fynes and Lally 2008). It is de- Service firms must therefore "consider not only veloped as the end-result of the activities of stra- the employees’ productivity but also the ’pro- tegic positioning, idea generation and concept ductivity’ and experience of the customer" (Fitz- development/refinement. The conceptual model simmons and Fitzsimmons 2010; Lusch et al. 2007; of a service consists of seven components which Schneider and Bowen 2010; Womack and Jones together define the desired customer outcomes Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Service Innovation for the Digital World 75

(value propositions) of the service: service bene- customer benefits minus the service cost (Clark fits, participation activities, emotional compon- et al. 2000). Service concept and the correspond- ent, perception component, service process, phys- ing service design (described below) are intended ical environment, and people/employee (Fynes to reflect the service firm’s business strategy and and Lally 2008). To define an innovative service therefore directly impact the firm’s financial per- concept, Bettencourt (Bettencourt 2010) recom- formance. From the perspective of service innov- mends that a service firm should: focus creative ation (or new service development) process (de- energies on specific job and outcome opportunit- tailed in Sect. 5.2) service concept is developed in ies; identify where the key problems lie in satis- the "Create Ideas" phase and selected for design fying high-opportunity jobs and outcomes; sys- in the "Evaluate and Select Ideas" phase (after ex- tematically consider a diverse set of new service perimentation), while the corresponding service ideas to satisfy the opportunities; and build a de- design is developed in the "Plan, Design Develop tailed concept with service strategy and service and Implement Ideas" phase. However, in the delivery in mind. digital world, the innovation process would tend to be circularly iterative akin to "agile (emergent) Service concept is the principal driver of service development" as opposed to a purely linear (pre- design decisions at all levels of planning and im- dictive) manner. plementation. It relates to service architecture or service blueprint which guides service design, 3.2 Service Design and to service governance which defines the decision rights and the decision making process Service design starts with the customer/user and for service design, planning and implementation defines how the service will be performed using (Goldstein et al. 2002). For example, at the stra- human-centred and user-participatory methods tegic planning level, the service concept drives to model the service performance (Holmlid and design decision for new or redesigned services. Evenson 2008). A service is conceptualised as an At the operational level it defines how the service open system with customers being present every- delivery system implements the service strategy where. Service design must address strategic and how to determine appropriate performance service issues such as marketing positioning and measures for evaluating service design. At the the preferred type of customer relationship, in service recovery level, it defines how to design line with the strategic intent of the service or- and enhance service encounter interactions. Thus ganisation. Service governance is also required service concept is the common foundation for to monitor the service qualities and financial per- new service development, service design and ser- formance against the design outputs. The frame- vice innovation. For instance, service concept work for designing the service delivery system development and testing is at the heart of service must address multiple interrelated factors: stand- design in new service development. Central to ardisation; transaction volume per time period; service conceptualisation is declaring what the locus of profit control; types of operating person- customer value proposition is in relation to the nel; types of customer contacts; quality control; firm’s strategic intent, how it meets the customer orientation of facilities; and motivational char- needs and what is the service logic required in acteristics of management and operating person- delivering the value proposition (Goldstein et al. nel (Goldstein et al. 2002). The service delivery 2002). Service concept articulates the service op- system fulfills the firm’s strategic service vision eration – why and how the service is delivered; and is designed/specified by means of service the service experience – i.e., customer’s experi- blueprinting (Bitner et al. 2008; Fitzsimmons and ence; the service outcome – i.e., customer bene- Fitzsimmons 2010). Service blueprinting is a map fits; and the service value – i.e., the perceived or flowchart of all the transactions constituting Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 76 Eng K. Chew

the service delivery process. The map identifies: to the technology; (b) technology-facilitated ser- the potential ’fail-points’; the line of interaction vice encounter – both the customer and the con- between client and provider known as service en- tact personnel have access to the technology; counters; the line of visibility – above it employ- (c) technology-mediated service encounter – the ees actions are visible to the customer (directly customer and contact personnel are not physic- affecting customer experience); below it is the ally co-located and their interaction is mediated ’back-stage’ ; and the internal line of interactions through the (online) technology; (d) technology- below the line of visibility (Bitner et al. 2008; generated service encounter – i.e., self-service, Fitzsimmons and Fitzsimmons 2010). The service the contact personnel is completely replaced by encounter design is a critical element of service technology (Fitzsimmons and Fitzsimmons 2010; design, because from the customer’s viewpoint Froehle and Roth 2004). Thus technological innov- "these encounters ARE the service" (Bitner et al. ation in services could require a change in cus- 2008). The design focuses on maximising the qual- tomer role in the service delivery process. There- fore it is critical to take into account the potential ity of ’service experience’ by the customer. How- customer (as well as employee) reaction to the ever, service experience is the result of the com- new technology in the design phase to avoid fu- bined efforts of the ’back stage’ information and ture problems of acceptance (Fitzsimmons and processes and the ’front stage’ customer handling Fitzsimmons 2010). – both must work seamlessly in unison in satis- fying the customer request (Glushko and Tabas Service design must include strategies for hand- 2009). ling service variability to ensure sustained level of service quality expected by customers (Glushko Taking an end-to-end view of service process and Tabas 2009). For instance, to manage an unex- allows designers to analyse the stakeholders’ re- pected deviation from normal service encounter, quirements, pain points and performance met- the service design (per service strategy and gov- rics from which service design (or redesign for ernance) may incorporate the notion of service an existing service) could be developed in col- personnel ’empowerment’ which grants them laboration with the stakeholders incorporating the discretion to recover from service deviation a combination of changes across process, organ- (failure) by offering ’compensations’ or altern- isation, technology, and tool in an integrative ative solutions to the customer to minimise ad- manner (Maglio et al. 2006). verse impacts to the customer (Glushko and Ta- bas 2009). Moreover, where multichannel services Service encounter design is guided by the pos- are provided, the design must ensure consistent sible relationships between the three parties in service experience across all channels. Finally, the service encounter: the service organisation service design needs to incorporate the require- (whether to pursue a service strategy of efficiency ments of lean consumption (Womack and Jones (cost leadership) or effective (customer satisfac- 2005) and achieve the objectives of service profit tion) or both); the contact personnel (following chain (Heskett et al. 2008). strict rules/order or empowered with autonomy Design of a service system (which offers the ser- and discretion); and the interaction between con- vice) similarly must address the roles of people, tact personnel and the customer (balancing con- technology, shared information, as well as the flicting "perceived control" by both parties) (Fitz- role of customer input in production processes simmons and Fitzsimmons 2010). Technology and the application of competences to benefit oth- could be designed into the service encounter in ers. The design must also address the service sys- four ways: (a) technology-assisted service en- tems’ requirements for agility and adaptability counter – only the contact personnel has access in alignment with their environments (Spohrer Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Service Innovation for the Digital World 77

et al. 2007). A learning framework is necessary to design embodying the customer experience re- sustain the firm’s creative design ability, and im- quirements (Patricio et al. 2008). prove and scale the service systems. The framework is designed to achieve three critical require- Service organisations are increasingly managing ments: effectiveness – the right things get done; customer experiences to promote differentiation efficiency – things are done in the right way; and customer loyalty. The experience-centric ser- sustainability – the right relationships exist with vice providers design the activity and context other service systems to ensure the system’s long of the experience to engage customers in a per- term sustainability (Maglio et al. 2009; Spohrer sonal, memorable way. The experience design et al. 2007). Sustainability is achieved through must address the dynamic and ongoing engage- the service system’s (brand) reputation, because ment process between customers and the service excellent reputations naturally attract value pro- organisation. The engagement can be emotional, positions from other service systems wanting physical, intellectual, or even spiritual, depend- to co-create value. It also requires appropriate ing on the level of customer participation and amount of shared information to be available to all service systems (the principle of information the connection with the environment (Zomerdijk symmetry) to enhance coordination and mutual and Voss 2010). sustainability within the service ecosystem. The design is however inherently challenged by the Customer value creation process is a dynamic, in- people factor, as people are complex and adapt- teractive, non-linear and often unconscious pro- ive. cess (Payne et al. 2008). Value is in the context of the performance outcome of the customer’s In sum, service system design, broadly, must ad- resource integration practice. To ensure optimal dress four variables: physical setting; process value co-creation, the three contiguous processes: design – the service blueprinting or mapping the customer value-creating processes; the sup- which designs ’quality’ into the service deliv- plier value-creating processes and the interfacing ery system; job design – the social technical job design which include addressing the employee service encounter processes must all be aligned motivational requirements; and people – the staff (Payne et al. 2008). The customer experience is a (competence) selection (Goldstein et al. 2002). culmination of the customer’s cognitions, emotions and behaviour during the relationship with 3.3 CustomerExperience & Value the supplier. These elements are interdependent Creation and involve the customer in thinking, feeling and doing – leading to customer learning – in the Customer experience requirements of each ser- process of value co-creation (Payne et al. 2008). vice type are usually analysed using use-case Indeed, a recent study by (Helkkula et al. 2012) scenarios similar to that of service blueprint (Bit- showed that "value in the [customer] experience ner et al. 2008; Patricio et al. 2008). Customer [is characterised] as an ongoing, iterative circu- experience is influenced by the service intens- ity, which is defined in terms of the number of lar process of individual, and collective customer actions initiated by the service provider, or the sense making, as opposed to a linear, cognitive amount of information exchanged in a service en- process restricted to isolated service encounters." counter or the duration of the service encounter (p.59) More research is required on "the need for (Glushko and Tabas 2009). The service design of appropriate metrics for the cognitive and emo- multi-interface system must unify service man- tional demands" of customer experience imposed agement, human computer interface, and soft- by different service interaction designs (Glushko ware engineering perspectives into an integrated and Tabas 2009). Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 78 Eng K. Chew

3.4 Service Architecture loosely coupled than product components (Roth and Menor 2003). Service architecture is conceptualised to systematise service design and innovation. Leveraging A service system can be analysed, for the pur- concepts from product architecture, service ar- poses of service architecture, in terms of four chitecture aims to create a common language levels of increasing details in specification: in- (comprised of nodes and linkages) across differ- dustry level, service company/supply chain level, ent views on service design and a systematic way service bundle level, and service package/com- to operationalise and measure the degree of ser- ponent level (Voss and Hsuan 2009). At level 0, vice architecture modularity (Voss and Hsuan the industry architectural template defines the 2009). value creation and the division of labour as well as value appropriation and the division of surplus Service architecture is constituted in accordance or revenue among the different players. At level with the principle of modularity, which in turn 1, the service company and its supply chain(s) is characterised by five dimensions: compon- are modelled both upstream and downstream. ents and systems as the basic modular units, the Both shared (internal cross-functional) and out- interfaces, degree of coupling, and commonal- sourcing of service components are important ity sharing between components, and platform consideration for the service company level for as the overarching configuration of components economic and resource flexibility reasons, in line and interfaces that makes up the product/service with its business strategy. At levels 2 and 3, the architecture (Fixson 2005). Modularity refers to service concept and service design activities of the degrees by which interfaces between com- service innovation practice are harmonised and ponents are standardised and specified to allow integrated to assure service agility. At level 2, for greater re-usability and sharing of (common) the individual service bundles of the service of- components among product/service families. It fering at the company level are analysed – each provides the basis for mixing and matching of bundle is viewed as a set of modules of service components to meet the mass-customisation re- delivery, comprising the front- and back-office quirements; yields economies of scale and scope, functions (and associated capabilities). At level 3, and can help structure products/services to fa- the service package and component level, the cilitate outsourcing. Platform strategies are the characteristics of the building blocks (compon- vehicles for realisation of mass customisation ents) are specified that contribute to the overall (Fixson 2005). As platform decisions often cut systems architecture, namely: standardisation, across several product/service lines or divisional uniqueness, degree of coupling and replicability boundaries, platform strategic decisions must (Voss and Hsuan 2009). Thus, service architecture belong in the top management team who need enables service agility as new services can be to and can resolve cross-functional conflicts to provisioned with minimal cost and little internal jointly-achieve the firm overall strategy. change, and the architecture can be dynamically adapted in response to external stimuli. But An important and challenging aspect of service this would require support by a corresponding architecture is the interface. Interfaces in ser- modular organisational architecture as well as IS vices can include people, information, and rules architecture (Voss and Hsuan 2009). governing the flow of information. Service interface can also include the flow of people. In 4 Service Strategy & Business Model general, an active role in service customisation There is a four-step approach to developing a would be played by both the front-end employ- successful service strategy: (1) Select the innova- ees and the customers themselves. This would tion focus, such as new service innovation or ser- suggest the service components need to be more vice delivery innovation, and the target customer Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Service Innovation for the Digital World 79

group(s); (2) Uncover customer needs in terms Strategy defines the choice as to which business of jobs to get done and outcomes expected; (3) model among many options to adopt for competi- Prioritise customer needs; (4) Develop a service tion in the marketplace. Thus the chosen business strategy (and attendant service concept) to fulfil model is a reflection of the service strategy – it the high priority customer needs (Bettencourt represents the logic of the firm, the way it oper- 2010). A successful service strategy fits what the ates and how it creates value for its stakehold- customer will value with what the company can ers (Casadesus-Masanell and Ricart 2010; Oster- deliver. This means aligning the service concept walder and Pigneur 2005). Service business model (what it would take to deliver on the customer defines the end-to-end service delivery activities, value propositions), and hence service architec- in accordance with the service concept, by which ture, with firm’s capabilities, resources, culture firms deliver value to customers, entice custom- and strategy. ers to pay for value, and convert those payments to profit (Osterwalder and Pigneur 2005; Teece Experiences of leading companies, such as South- 2010). It articulates the logic, the data, and other west Airlines, show that successful strategies evidence that support a value proposition for would include: (1) close coordination of the mar- the customer, and a viable structure of reven- keting and operations relationship; (2) a strategy ues and costs for the enterprise delivering that built around elements of a strategic service vis- value. Business model embodies the organisa- ion; (3) an ability to redirect the strategic service tional and financial ’architectures’ of a business inward to focus on vital employee groups; (4) an (Osterwalder and Pigneur 2005; Teece 2010). A appraisal of the effects of scale on both efficiency business model can be conceptualised as a sys- and effectiveness; (5) the substitution of informa- tem of interdependent (service delivery) activit- tion for other assets; and (6) the exploitation of ies that transcends the focal firm and spans its information to generate new business (Heskett et boundaries, and enables the firm, in concert with al. 2008). In addition, six successful strategic prac- its partners, to create value and also to appropri- tices have been identified for service commer- ate a share of that value. The service business cialisation: (1) leveraging fundamental sources model is composed of two building blocks: (a) of value that influence shareholder wealth, (2) design elements – content, structure and gov- managing customers’ perceptions of the service ernance that describe the architecture of a service value proposition, (3) creating an attractive fin- delivery activity system (Level 2 and Level 3 of ancial architecture for customising pricing for service architecture); (b) design themes – novelty, profitability, (4) ensuring service excellence in im- lock-in, complementarities and efficiency that de- plementation, (5) planning for service recovery, scribe the sources of the service delivery activity and (6) managing the holistic service experience system’s value creation (Zott and Amit 2010). (including the servicescape) (Bolton et al. 2007). These successful strategic practices mirror the In sum, a service firm’s customer value propos- design of corresponding business model design ition crystallised by the service concept serves considerations below and require superior collab- as the bridge connecting its service strategy and orative competence. This is because it leverages business model. The former defines the service the firm’s dynamic capability to absorb informa- concept and service delivery mechanisms (con- tion and knowledge from the environment, cus- sistent with the service architecture) while the tomers, and its value networks, and adapt the latter defines the revenue and cost models (finan- service to respond to dynamic and complex en- cial architecture) of the selected activity system vironments, while ensuring consistent superior (in accordance with the service delivery archi- customer experience at each service encounter tecture) designed to serve the targeted customer point. segments. Both practices tend to be pursued in Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 80 Eng K. Chew

parallel and interactively due to their close inter- concept – a new idea of concept of how to organ- relationship. And both practices are required to ise a solution to a job/problem in a given mar- create and sustain the competitive advantage for ket; (b) new client interface – new information- the firm. centric (often online) personalised interface (Gal- louj 2002) to facilitate service offering co-design, 5 Service Innovation Models and co-production and value co-creation with the cli- Process ents; (c) new service delivery system and organisation in line with the firm’s strategic service Service innovation is about the creation of cus- vision and new service concept; and (d) techno- tomer value (Grawe et al. 2009). The source of logy options – the specific role of technology 1 service innovation opportunities is from discov- selected (Gallouj 2002) in the service innova- ering how customers define value – for instance, tion (Hertog 2002). Thus service innovation is a customers hire products and services or solutions multi-dimensional phenomenon. A completely to get a job done; or use outcomes to evaluate new service (radical innovation) usually means innovations in all the above four dimensions. success in getting a job done; and have distinct On the other hand, incremental service innov- needs that arise related to the "consumption" of a ation means innovation in one or more of the solution (Bettencourt 2010). Four types of service above four dimensions. Equally important is innovation can be identified from the customer the need to address the linkages between these viewpoint: (1) New service innovation – discov- dimensions in order to implement the service ery of new or related jobs to get done; (2) Core innovation, as they represent the requisite mar- service innovation – helping the customer get a keting, organisational development and learning core job done better; (3) Service delivery innov- processes (human resource) (Gallouj 2002; Maglio ation – improving the ways a core job get done; et al. 2009; Spohrer et al. 2007) and distribution (4) Supplementary service innovation – helping (supply chain/logistics) capabilities to realise the the customer get jobs done related to product us- innovation. For example, launching a new ser- age or consumption done (Fynes and Lally 2008). vice concept requires marketing expertise. The Service innovation can also be characterised by decision as to whether to develop new services the degree of interaction with the customer and requires organisational knowledge: the organisa- the degree of information asymmetry within the tional capabilities required versus available and service relationship (Gallouj 2002). This section suitability of existing organisational structure to reviews the common, foundational service innov- deliver the service (Gallouj 2002; Hertog 2002). ation (functional and competence-based) models Thus while service innovation may arise from and processes for creating all types of innovative changing one of the above four dimensions, it services that help customers get their jobs done. requires interdisciplinary collaboration between marketing, human resource, distribution and IT 5.1 Functional Model of Service to bring about the change and take the innova- Innovation tion to market. In sum, each particular (type of) service innovation is characterised by the com- Service innovation is often a result of a combin- bination of the four dimensions: the weight of ation of conceptual, technological and organisa- the individual dimensions and the relative sig- tional innovations combined with new ways of nificance of the various linkages between them relating to the consumer (Hertog 2002). A com- 1 monly used functional model for identifying the Use of technologies in service firms tends to follow the so-called "Barras reverse product cycle RPC" model – focus or vector of a service innovation consists start with back-end then front-end process innovations and of four dimensions of service: (a) new service finally product/service innovation (Gallouj 2002). Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Service Innovation for the Digital World 81

(Hertog 2002). To co-create and capture value for 5.3 Service Innovation Process and the innovative firm, a new business model must Management be designed that reflects the operating and finan- Service innovation competence is a crucial oper- cial model of the service concept and associated ant resource for the firm’s competitive advantage. linkages to the other dimensions (Teece 2010). Service innovation practice depends critically on a streamlined and flexible process for internal 5.2 Competence-based Model of Service and external resource coordination and integra- Innovation tion to achieve effective and efficient customer value co-creation. Service innovation process, There are three different approaches to defining also known as new service development, generally (Engel et al. 2006; Thomke 2003) consists of and studying service innovation (Gallouj 2002): five phases: an assimilation or technologist approach, which treats services as similar to manufacturing; a de- • Create ideas – this phase defines the idea, its marcation or service-oriented approach, which scope and business benefits distinguishes services (possessing the aforemen- • Evaluate and select ideas – this phase prior- tioned IHIP characteristics) from manufacturing itises the portfolio of ideas and develops the innovation; and a synthesis or integrative ap- selected idea into a (low cost low risk) experi- proach, which suggests that service innovation ment to test its feasibility; go/no go decision is brings to the forefront hitherto neglected ele- made quickly to speed up the chance of identi- ments of innovation that are of relevance for fying a feasible idea (or conversely the rate of manufacturing as well as services. The synthesis failures of infeasible ideas) or integrative approach is widely adopted and • Plan, design, develop and implement ideas – it is congruent to the service-dominant (S-D) lo- this phase takes the feasible idea through a gic. The best known model of this approach is rigorous service development lifecycle the Gallouj-Weinstein competence-based model • Commercialise the ideas – this phase launches (Gallouj and Weinstein 1997) that represents a the service product or a service as a system of (provider) • Review the impacts – this phase reviews the competences (PCi), technical characteristics (PTi), results of the innovation to improve current and final characteristics (Oi), where the service performance and as a feedback for future pro- outcome (Oi) is resulted from the interactions cess improvement between the customer competences (CCi) and the provider’s competences (PCi) and technical However, as alluded to in the design practices characteristics (PTi). Service innovations thus framework (Sect. 3.1), in the digital world this in- consist of changes in one or more of these ele- novation process would not necessarily occur ments. Provider competences PCi are then the in a purely linear (predictive) manner, rather direct mobilisation of service personnel compet- it would tend to be circularly iterative, akin to ences (i.e., without any technological mediation). "agile (emergent) development". PTi are knowledge, competences embodied in Research on service innovations has highlighted tangible (such as front- and back-office character- the critical importance of the front-end stages of istics) or intangible (i.e., codified and formalised new service development: idea generation, idea competences such as job analysis methods. A screening and concept development – collect- fundamental characteristic of service activities ively known as the fuzzy front-end (Alam 2006). is client participation (in various forms) in the Customer involvements in the front-end stages production of the service (Gallouj 2002). of a service innovation process are important Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 82 Eng K. Chew

so as to reduce the fuzziness (Alam 2006). Ser- service experience requirements would demand vice innovation may be incremental for steady service variability tailored to individual prefer- business growth – through exploitation of exist- ences. In general, customer participation is in- ing competences (O’Reilly and Tushman 2008); herently a source of variability since each cus- or radical for new growth idea (Anthony et al. tomer has different capabilities and must learn 2008), which could become a new growth plat- how to interact with the service process (Metters form (Laurie et al. 2006) – through exploration of and Marucheck 2007). The concept of customer new competences/capabilities (O’Reilly and Tush- efficiency is therefore a critical requirement of man 2008). But the exploratory activities must be service process design to denote the customer’s buffered from exploitative activities to ensure co- ability to participate in self service or coproduce existence (Benner and Tushman 2003), creating service (Metters and Marucheck 2007; Xue and a so-called ambidextrous organisation capable Harker 2002) – for instance the user innovation of both exploitative and exploratory innovations toolkit (Hippel 2001). Similarly, customer vari- simultaneously. ability is, thus, a design variable which can be Companies are also increasingly leveraging in- managed to improve both service quality and novative ideas from outside the firms using an efficiency (Metters and Marucheck 2007). open innovation process (Chesbrough 2003). This Firms compete through service by collaborat- means the firm needs to engage customers, part- ing (i.e., co-produce offering and co-create value) ners, suppliers, regulators, and even competitors with customers and network partners to enhance to co-generate creative ideas, co-produce service knowledge (Lusch et al. 2007). This requires the offerings and co-create value in a continual non- firm to possess absorptive capacity (Zahra and linear process of service innovation, which sup- George 2002) in order to absorb new informa- ports direct interactions with the customers to tion and knowledge from customers and partners match innovations with customers needs (Ches- to comprehend from the external environments brough 2011). The aim of customer participation, the important trends and know-how which, in as described in the next section, is to co-create a turn, give them the ability to adapt/adjust to the "unique personalised customer experience" (Pra- complex, dynamic, and turbulent external envir- halad and Krishnan 2008). onments. Firms that draw on the knowledge of 6 Customer Participation their customer base can capitalise on customer competencies for use during the course of their Central to discovering service innovation op- innovation activities (Blazevic and Lievens 2008). portunities is "knowing how customers define value" (Bettencourt 2010). As service value is al- Customer participation or involvement in service ways determined by the customer, new creative innovation can take place at various phases of ideas must be developed from the customer’s the new service development process (Alam 2006; outside-in view (Edvardsson et al. 2007; Payne Chesbrough 2011). Customer participation or in- et al. 2008). Indeed, successful firms are co-opting tegration can be conceptualised as the incorpora- customer involvement in service and value co- tion of resources from customers into the service creation (Prahalad and Ramaswamy 2000). Cus- development processes of a company (Moeller tomer participation is equally essential to both 2008). This would include participating in pro- the ’old’ physical and ’new’ digital service worlds. ducing and delivering the service (Dong et al. However, involving customer in co-production 2008). Business has to develop an adaptive or- of a service process is often confronted with con- ganisational model where customer involvement flicting design requirements. For example, scale- and innovation is persistent and inherent in the economy or efficiency requirements would de- entire service lifecycle – such that the distinc- mand service standardisation, while personalised tion between customers and employees becomes Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Service Innovation for the Digital World 83

blurred (Oxton 2008). This organisational model experience, perceived value in future co-creation, operates as a network of relationships based on and intention to co-create in the future (Dong the principles of alignment, transparency, iden- et al. 2008). tity (reputation) (Oxton 2008). Customer participation towards creating person- 7 Community-based Innovation alised experience (Prahalad and Krishnan 2008) The advent of social media and clouds-based ser- typically follows a five-stage iterative approach: vices has led many firms globally, as part of im- 1) establishment of antecedent conditions for plementing their social strategies, to directly en- customer to participate; 2) development of mo- gage with their customers online across a broad tivations or customer benefits; 3) cost-benefit range of activities (such as marketing, customer evaluation; 4) activation of co-creation process care, etc.) to co-create value for mutual bene- by choosing the stages of the "production-con- fits. This has evolved from a relatively straight- sumption" activity chain; and 5) evaluation of forward traditional online customer service plat- the effectiveness of the co-creation strategies form to a more sophisticated community based against the cost-benefit analysis (Etgar 2008). It innovation (CBI) which requires a new set of is prudent for the provider to institute a continu- organisational capabilities that interact and in- ous learning process with the customer from the tegrate with those of the customers themselves co-creation experience to improve their service- (Fuller et al. 2006). usage competence. Learning enhances the customer’s competence in seamlessly integrating CBI is defined as a new online service innovation the value proposition with their lives, objectives process that fully engages the firm’s customer and aspiration (Payne et al. 2008). Organisational community from ideation phase right through learning about customer’s value creation pro- to the test and launch phase of New Service De- cesses deepens customer insights. Organisational velopment. The community members become learning is a crucial process for nurturing the pro- the sources of new service ideas as well as the vider’s collaborative competence to improve the co-creators and evaluators of the service designs. provider’s innovation capability and competitive The most common CBI user/customer archetype advantage (Edmondson 2008). is called the "lead users" – who are highly know- ledgeable of the firm’s products/service and have The increased digitalisation of services in the ’job’ (problem) needs that are ahead of all other internet era is creating new opportunities for user groups in a given market. Lead users are knowledge coproduction between customers and allowed to design (using interactive toolkits pro- the provider (Blazevic and Lievens 2008). In a vided by the service provider) their own products/ digital world, customers may take on three dif- service by trial-and-error according to their wants ferent roles for knowledge coproduction-passive and needs. Their creativity and problem-solving user, active informer, and bidirectional creator- skills (competencies) using the toolkits (provider each with distinctive declarative and procedural competencies) will produce the ’ideal’ solutions characteristics, and distinct impacts on the three to match their problems (the ’jobs’ to be done) – innovation tasks of detection, development, and for instance, Peugeot’s "Retrofuturism" car deployment (Blazevic and Lievens 2008). The di- designs were produced using CBI1 (Fuller et al. gital world also facilitates customer participation 2006). Two other user archetypes are also com- in recovery from service failure. This may vary mon: the "insiders" who are strongly associated in degrees from firm recovery, joint recovery, to in the community and highly involved in the customer recovery (Dong et al. 2008). This would topic; the "devotee" who are highly involved with require higher levels of role clarity, but it also tends to enhance satisfaction with the service 1www.peugeot-avenue.com. Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 84 Eng K. Chew

the topic but not very much related with the com- strengthening their relationships – that is focus munity. CBI communities could be selected on on many-to-many social activities between com- the basis of the exchanged content, professional- munity members as exemplified by eBay’s Group ism, traffic volume, and number of participants Gift (Piskorski 2011). interacting with each other (Fuller et al. 2006). Users could be accessed directly or more often 8 Strategic Management for Innovation they recommend access via a trustworthy mem- Success ber of the community or via the webmaster to Innovative service firms have strong commit- increase acceptance. Feedback to users on their ment to innovation from top management backed input is regarded critical as is getting users’ feed- by well structured innovation processes and gov- back on their participation experience and their ernance together with the aligned culture and willingness and expectations to participate again systems, and the attendant prioritised resources in future virtual product/service development allocated to innovation efforts. In service in- projects (Fuller et al. 2006). novation "it is not the service itself that is pro- Community members engagement in CBI can duced but the pre-requisites for the service" (Ed- be fostered and sustained in a three-step pro- vardsson and Olsson 1996). Due to services’ real- cess: (1) understand consumer needs and mo- time production, new service development would tivations; (2) promote community participation, require modifications of the service delivery pro- including encourage content creation, cultivate cess and changes in frontline employees’ skills. connections, and create enjoyable experiences; This would require strong fit between the new and (3) motivate cooperation, including mobil- service and existing systems; and close alignment ising member-leaders, inspiring idea creation and between the customer-service-focused front-end selection via a panel/polling (Porter et al. 2011). and the operational-excellence-focused back-end Community engagement is motivated intrinsic- systems. ally by the value created when community spon- But despite its strategic importance, service in- sors help user-members meet their needs with novation is notoriously difficult to accomplish their virtual community. So the community spon- (Dorner et al. 2011). This could be attributed to sor’s judicious and targeted efforts to encourage such managerial deficiencies as: lack of ability to members to act in ways that create greater value protect services hinders investment; lack of clear for themselves and for the firm are crucial to "organisational anchoring" of service innovation success (Porter et al. 2011). Members’ "embed- activities; lack of systematic innovation process; dedness" (willingness to act in value-creating lack of customer participation; and "bad ideas not ways toward a community sponsor) and "em- consistently eliminated" (Chandy and Tellis 1998). powerment" are seen to be fundamental to driv- So managers need to be vigilant in all innovation ing cooperative, engaging behaviour from the stages to assess ideas against the company’s stra- community members (Porter et al. 2011). This, in tegic goals and market needs in order to determ- turn, would require the community sponsor to ine their commercial viability. Further, managers understand the needs of its community members, need to focus on people (evolving competences build trust with and create value for its members in line with changing customer value expecta- (Porter et al. 2011). CBI tends to focus on firm- tions) and structural support (systematic new community (one-to-many and many-to-one) col- service development process supported by spe- laboration. More recently, new social strategies cific innovation tools, multi-disciplinary teams, are being proposed that seek to reduce company the availability of resources, market testing and costs and/or increase customer willingness to pay market research) to ensure successful service in- by helping the community to meet online and novation (Dorner et al. 2011). Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Service Innovation for the Digital World 85

Service innovation is technology-enabled but is accentuated by the design practices frame- more human-centred and process-oriented. work for service innovation which serves as a Therefore, the "envisioning, energising and en- foundation for systematic service conceptualisa- abling" capabilities, sound communication/co- tion, design, architecture and innovation. Service ordination, and reducing intra-organisational con- innovation commercialisation is contingent on flicts and power struggle have been identified mindful alignment of the firm’s service strategy, as fundamental and very critical for new ser- service concept and business model. Firm needs vice development to minimise organisational in- collaborative, absorptive capacity and dynamic ertia/resistance (Nijssen et al. 2006). Innovative capabilities (including organisational learning firms commonly possess "willingness to canni- processes) to continuously adapt its service in- balise" mindset and capability – i.e., willingness novations with the changing external environ- to make obsolete its existing products/services, ments including the value networks to which it is prior investments, and/or existing organisational connected. From strategic management perspect- capabilities (Chandy and Tellis 1998; Nijssen et al. ive, the firm needs to be ambidextrous capable 2006). These innovative organisations are said to of pursuing exploitative and exploratory service possess ambidexterity capable of pursuing sim- innovations simultaneously to create sustained ultaneous exploitative and exploratory innova- value for itself and its customers. tions. An ambidextrous organisation "requires a coherent alignment of competencies, structures Acknowledgements and cultures to engage in exploration, a contrast- This paper is based on and provides an extension ing congruent alignment focused on exploitation, to the author’s earlier work described in parts of and a senior leadership team with the cognit- Chew and Gottschalk (2013). ive and behavioural flexibility to establish and nurture both" (O’Reilly and Tushman 2008). References 9 Conclusion Alam I. (2006) Removing the fuzziness from the fuzzy-end of service innovations through cus- Service innovation is focused on creating customer interactions. In: Industrial Marketing tomer value, and service is about relationship Management 35(4) with the customer. Customer co-creates value Anthony S. D., Johnson M. W., Sinfield J. V. with the provider by integrating his/her com- (2008) Institutionalizing innovations. In: MIT petences/capabilities with those of the provider. Sloan Management Review 49(2) Thus customer productivity is as important as Arnould E. J. (2008) Service-dominant logic and that of the provider in service provision as it resource theory. In: Journal of the Academy impacts directly the service experience. Increas- of Marketing Science 36, pp. 21–24 ingly, in a digital world, customer and member- Benner M. J., Tushman M. L. (2003) Exploitation, community participation across the firm’s en- exploration and process management: the tire service innovation lifecycle is becoming a productivity dilemma revisited. In: Academy critical innovation strategy for sustained value of Management Review 28 co-creation. It has become a core and distinct- Bettencourt L. A. (2010) Service innovation: how ive organisational capability for service organ- to go from customer needs to breakthrough isations to develop and adapt in line with the services. McGraw-Hill, New York evolving external environments and the custom- Bitner M. J., Ostrom A. J., Morgan F. N. (2008) ers’ increasingly mature service competences. Service blueprinting: a practical technique Service innovation is technology-enabled but for service innovation. In: California Man- more human-centred and process-oriented. This agement Review 50(3), pp. 66–94 Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 86 Eng K. Chew

Blazevic V., Lievens A. (2008) Managing innov- Edvardsson B., Olsson J. (1996) Key concepts for ation through customer co-produced know- new service development. In: Service Indus- ledge in electronic services: an exploratory tries Journal 16(2), pp. 140–164 study. In: Journal of the Academy of Market- Edvardsson B., Gustafsson A., Roos I. (2005) Ser- ing Science 36, pp. 138–151 vice portraits in service research: a critical Bolton R. N., Grewal D., Levy M. (2007) Six review. In: International Journal of Service strategies for competing through service: an Industry Management 16(1), pp. 107–121 agenda for future research. In: Journal of Re- Edvardsson B., Gustafsson A., Enquist B. (2007) tailing 83(1), pp. 1–4 Success factors in new service development Casadesus-Masanell R., Ricart J. E. (2010) From and value creation through services. In: Spath strategy to business models and onto tactics. D., Fahnrich K.-P. (eds.) Advances in Ser- In: Long Range Planning 43, pp. 195–215 vices Innovations. Springer, Berlin Heidel- Chandy R. K., Tellis G. J. (1998) Organizing for berg, pp. 166–183 radical product innovation: the overlooked Eisingerich A. B., Rubera G., Seifert M. (2009) role of willingness to cannibalize. In: Journal Managing service innovation and interorgan- of Marketing Research 35(4), pp. 474–487 izational relationships for firm performance: Chesbrough H. (2003) A better way to innovate. to commit or diversify? In: Journal of Service In: Harvard Business Review 81(7), pp. 12–13 Research 11(4), pp. 344–356 Chesbrough H. (2011) Bringing open innovation Engel J. F., Thompson A. M., Nunes P. F., Linder to services. In: MIT Sloan Management Re- J. C. (2006) Innovation unbound. In: Accen- view 52(2), pp. 85–90 ture Publication 1, pp. 28–37 Chesbrough H., Davies A. (2010) Advancing Etgar M. (2008) A descriptive model of the con- service innovations. In: Maglio P. P., Kiel- sumer co-production process. In: Journal of iszewski C. A., Spohrer J. C. (eds.) Hand- the Academy of Marketing Science 36 book of Service Science. Springer, New York, Fitzsimmons J. A., Fitzsimmons M. J. (2010) Ser- pp. 579–601 vice Management: Operations, Strategy, In- Chew E., Gottschalk P. (2013) Knowledge driven formation Technology., 7th ed. McGraw-Hill, service innovation and management: IT Irwin, New York, NY strategies for business alignment and value Fixson S. K. (2005) Product architecture assess- creation. IGI Global, Hershey, PA, USA. ment: a tool to link product, process, and sup- Clark G., Johnston R., Shulver M. (2000) Exploit- ply chain design decisions. In: Journal of Op- ing the service concept for service design and erations Management 23(3/4), pp. 345–369 development. In: Fitzsimmons J. A., Fitzsim- Froehle C. M., Roth A. (2004) New measure- mons M. J. (eds.) New Service Design. Sage, ment scales for evaluating perceptions of the Thousand Oaks, CA, pp. 71–91 technology-mediated customer service exper- Dong B., Evans K. R., Zou S. (2008) The effects ience. In: Journal of Operations Management of customer participation in co-created ser- 22(1), pp. 1–22 vice recovery. In: Journal of the Academy of Fuller J., Bartl M., H. E., Muhlbacher H. (2006) Marketing Science 36 Community based innovation: how to integ- Dorner N., O. G., Gebauer H. (2011) Service in- rate members of virtual communities into novation: why is it so difficult to accomplish? new product development. In: Electronic In: Journal of Business Strategy 32(3), pp. 37– Commerce Research 6, pp. 57–73 46 Fynes B., Lally A. M. (2008) Innovation in ser- Edmondson A. C. (2008) The competitive im- vices: from service concepts to service experi- perative of learning. In: Harvard Business ences. In: Hefley B., Murphy W. (eds.) Service Review 86(7/8) Science, Management and Engineering Edu- Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Service Innovation for the Digital World 87

cation for the 21st Century. Springer, New ation Management 18, pp. 247–257 York, pp. 229–333 Holmlid S., Evenson S. (2008) Bringing service Gadrey J., Gallouj F. (2002) Productivity, Innov- design to service sciences, management and ation and Knowledge in Services: New Eco- engineering. In: Hefley B., Murphy W. (eds.) nomic and Socio-economic Approaches. Ed- Service Science, Management and Engineer- ward Elgar, Cheltenham, UK ing Education for the 21st Century. Springer, Gallouj F. (2002) Innovation in the service eco- New York, pp. 341–345 nomy: the new wealth of nations. Edward Johannessen J. A., Olsen B. (2010) The future Elgar, Cheltenham, UK of value creation and innovations: aspects Gallouj F., Weinstein O. (1997) Innovation in of a theory of value creation and innovation services. In: Research Policy 26, pp. 537–556 in a global knowledge economy. In: Interna- Glushko R. J., Tabas L. (2009) Designing ser- tional Journal of Information Management vice systems by bridging the ‘front stage’ 30, pp. 502–511 and ‘back stage’. In: Information Systems, E- Laurie D. L., Doz Y. L., Sheer C. P. (2006) Creat- business Management 7, pp. 407–427 ing new growth platforms. In: Harvard Busi- Goldstein S. M., Johnston R., Duffy J., Rao J. ness Review 84(5) (2002) The service concept: the missing link Lovelock C., Gummesson E. (2004) Whither in service design research? In: Journal of Op- services marketing?: in search of a new erations Management 20(2), pp. 121–134 paradigm and fresh perspectives. In: Journal Grawe S. J., Chen H., Daugherty P. J. (2009) The of Services Research 7, pp. 20–41 relationship between strategic orientation, Lusch R. F., Vargo S. L., O’Brien M. (2007) Com- service innovation, and performance. In: In- peting through service: insights from service ternational Journal of Physical Distribution dominant logic. In: Journal of Retailing 83(1), & Logistics Management 39(4), pp. 282–300 pp. 5–18 Hastings H., Saperstein J. (2013) A practice- Lusch R. F., Vargo S. L., Tanniru M. (2009) Ser- driven service framework for value creation. vice, value networks and learning. In: Journal In: 2013 IEEE International Conference on of the Academy of Marketing Science 38(1), Business Informatics. Piscataway, NJ, USA pp. 19–31 Helkkula A., Kelleher C., Pihlstrom M. (2012) Madhavaram S., Hunt S. D. (2008) The service- Characterizing value as an experience: im- dominant logic and a hierarchy of oper- plications for service researchers and man- ant resources: developing masterful oper- agers. In: Journal of Service Research 15(1), ant resources and implications for market- pp. 59–75 ing strategy. In: Journal of the Academy of Hertog P. d. (2002) Co-producers of innovation: Marketing Science 36, pp. 67–82 on the role of knowledge-intensive business Maglio P. P., Spohrer J. (2008) Fundamental of services in innovation. In: Gadrey J., Gallouj service science. In: Journal of the Academy F. (eds.) Productivity, Innovation and Know- of Marketing Science 36, pp. 18–20 ledge in Services: New Economics and Socio- Maglio P. P., Srinivasan S., Kreulen J. T., Spohrer Economic Approaches. Edward Elgar, Chel- J. (2006) Service systems, service scientists, tenham, UK, pp. 223–255 SSME and innovation. In: Communications Heskett J. L., Jones T. O., Loveman G. W., Sas- of the ACM 49(7), pp. 81–85 ser W. E., Schlesinger L. A. (2008) Putting Maglio P. P., Vargo S. L., Caswell N., Spohrer J. the service-profit chain to work. In: Harvard (2009) The service system is the basic abstrac- Business Review 86(7/8) tion of service science. In: Information Sys- Hippel E. v. (2001) Perspective: user toolkits for tems E-Business Management 7(4), pp. 395– innovation. In: The Journal of Product Innov- 406 Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 88 Eng K. Chew

Metters R., Marucheck A. (2007) Service manage- 122 ment – academic issues and scholarly reflec- Porter C. E., Donthu N., MacElroy W. H., Wydra tions from operations management research- D. (2011) How to foster and sustain engage- ers. In: Decision Sciences 38(2), pp. 195–214 ment in virtual communities. In: California Moeller S. (2008) Customer integration – a key Management Review 53(4), pp. 80–110 to an implementation perspective of service Prahalad C. K., Krishnan M. S. (2008) New age of provision. In: Journal of Service Research innovation: driving cocreated value through 11(2), pp. 197–210 global networks. McGraw-Hill, New York Nijssen E. J., Hillebrand B., Vermeulen P. A. Prahalad C. K., Ramaswamy V. (2000) Co-opting M., Kemp R. G. M. (2006) Exploring product customer competence. In: Harvard Business and service innovation similarities and differ- Review 78(1), pp. 79–87 ences. In: International Journal of Research Roth A. V., Menor L. J. (2003) Insights into in Marketing 23 service operations management: a research Normann R., Ramirez R. (1993) From value chain agenda. In: Production and Operations Man- to value constellation: designing interactive agement 12(2), pp. 145–164 strategy. In: Harvard Business Review 71, Schneider B., Bowen D. E. (2010) Winning the pp. 65–77 service game. In: Maglio P. P., Kieliszewski, Ordanini A., Parasuraman A. (2011) Service in- C. S. J. (eds.) Handbook of Service Science. novation viewed through a service domin- Springer, New York, pp. 31–59 ant logic lens: a conceptual framework and Spohrer J., Maglio P. P., Bailey J., Gruhl D. (2007) empirical analysis. In: Journal of Service Re- Steps toward a science of service systems. In: search 14(1), pp. 3–23 IEEE Computer 1, pp. 70–77 O’Reilly C. A., Tushman M. L. (2008) Ambidex- Teece D. J. (2007) Explicating dynamic capab- terity as a dynamic capability: resolving the ilities: the nature and microfoundations of innovator’s dilemma. In: Research in Organ- (sustainable) enterprise performance. In: Stra- izational Behavior 28 tegic Management Journal 28, pp. 1319–1350 Osterwalder A., Pigneur Y. (2005) Clarifying Teece D. J. (2010) Business models, business business models: origins, present, and future strategy and innovation. In: Long Range Plan- of the concept. In: Communications of the ning 43, pp. 172–194 Association for Information System 16, pp. 1– Thomke S. (2003) R&D comes to services – Bank 25 of America’s pathbreaking experiments. In: Oxton G. (2008) An integrated approach to ser- Harvard Business Review 81(4), pp. 71–79 vice innovation. In: Hefley B., Murphy W. University of Cambridge, IBM (2007) Succeed- (eds.) Service Science, Management and En- ing through Service Innovation: A Discussion gineering Education for the 21st Century. Paper. University of Cambridge Institute for Springer, New York, pp. 97–105 Manufacturing, Cambridge, United Kingdom Patricio L., Fisk R. P., Cunba J. F. (2008) Design- Vargo S. L., Lusch R. F. (2004) Evolving to a new ing multi-interface service experiences: the dominant logic for marketing. In: Journal of service experience blueprint. In: Journal of Marketing 69, pp. 1–17 Service Research 10(5) Vargo S. L., Lusch R. F. (2008) Service-dominant Payne A. F., Storbacka K., Frow P. (2008) Man- logic: continuing the evolution. In: Journal of aging the co-creation of value. In: Journal of the Academy of Marketing Science 36, pp. 1– the Academy of Marketing Science 36, pp. 83– 10 96 Vargo S. L., Maglio P. P., Akaka M. A. (2008) Piskorski M. J. (2011) Social strategies that work. On value and value co-creation: a service In: Harvard Business Review 89(11), pp. 117– systems and service logic perspective. In: Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Service Innovation for the Digital World 89

European Management Journal 26, pp. 145– 152 Voss C. A., Hsuan J. (2009) Service architecture and modularity. In: Decision Sciences 40(3), pp. 541–569 Womack J. P., Jones D. T. (2005) Lean consumption. In: Harvard Business Review 83(3), pp. 58–68 Xue M., Harker P. T. (2002) Customer efficiency: Concept and its impact on e-business management. In: Journal of Service Research 4(2) Zahra S. A., George G. (2002) Absorptive capacity: a review, reconceptualization, and extension. In: Academy of Management Review 27(2), pp. 185–203 Zomerdijk L. G., Voss C. A. (2010) Service design for experience-centric services. In: Journal of Service Research 13(1), pp. 67–82 Zott C., Amit R. (2010) Business model design: an activity system perspective. In: Long Range Planning 43, pp. 216–226

Eng K. Chew School of Systems, Management and Leadership University of Technology Sydney Australia [email protected] Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 90 Stéphane Marchand-Maillet and Birgit Hofreiter

Stéphane Marchand-Maillet and Birgit Hofreiter

Big Data Management and Analysis for Business Informatics A Survey

Modern communication networks have fueled the creation of massive volumes of data that may be valued as relevant information for business activities. In this paper, we review technologies for enabling and empowering business activities, leveraging the content of this big data. We distinguish between data- and user-related technologies, and study the parallel brought by the overlap of these categories. We show how the trend of Big Data is related to data security and user privacy. We then investigate automated ways of performing data analysis for Business Intelligence. We finally review how groups of users may be seen as a workforce in business through the notion of human computation or crowdsourcing, associated with the notions of trust and reputation. We conclude by discussing emerging trends in the domain.

1 Introduction characteristics, potential risks and benefits. User- generated data is considered a potentially rich Progress in Business Informatics aim to develop source of information for business and user be- business administration using computational and haviors are modeled using this data. Users and information technologies. As such, business in- data are therefore two inter-related main actors formatics may use any method providing a tech- within this landscape that we explore via these nology useful for its end purpose. two perspectives, as illustrated in Fig.1.

Modern business activities essentially rely on Data Users

an accurate management of knowledge (often Storage and access Trust referred to as Business Intelligence). The devel- Visual Analytics opment of communication technologies and the Security Privacy wide-spread and ubiquity of communication net- Management Knowedge Big Data Communities works have created an opportunity for gather- Analysis Management ing and analysing data in view of deriving use- Ontologies Social Knowledge ful knowledge. Hence, business informatics is Social Networks (Web 2.0) primarily supported by data management and Machine Intelligence Human Computation data analysis technologies. In addition, users and user groups remain at the center of any busi- Figure 1: A classification of domains for enabling ness. They may assist performing data analysis and empowering technologies in business informatics. as much as benefiting from it. Square boxes indicate technical domains, whereas cir- cular items relate to multi-disciplinary binding fields In this paper, we review and analyse the main of study enabling technologies in business informatics. We explore as thoroughly as possible the information landscape in which business informatics Investigating data-related aspects allows to un- operates, to understand the aspects and their derstand the technical infrastructure that should Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Big Data Management and Analysis for Business Informatics 91

be set up and sustained, from base data collec- • Process data: sensing data (GPS, weather, traffic tion and housing to sophisticated data analysis. monitoring, ...), computation (sociological, sci- In parallel, studying user-related issues allows to entific, financial trends, ...); model the user and his community, as originator • Logs: traces of human interaction with sys- of this data. We therefore distinguish between tems (information, e-commerce, entertainment, data- and user-related technologies, although the ...), logs of machine-machine communications split is somewhat artificial since these techno- (web services, distributed computing, ...). logies generally overlap. We further look into passive management technologies (whose main Before considering data analysis, a choice is to aim is not to create any knowledge) against act- be made on the form of data housing, if any. ive analysis technologies (that transform data The emergent paradigm of big data (Sect. 2.1) is into knowledge). The extend of our review is addressing some choices there. In turn, data pre- symbolised in Fig.1. Section2 reviews the hous- servation and access immediately open security ing and preservation of data, in relation to the issues, reviewed in Sect. 2.2. current challenge of Big Data. In Sect.3, we review automated technologies for data analysis. 2.1 Big data These are crucial for their aspect of scalability, since any necessity for user intervention would Every study on the topic shows clearly that the create prohibitive costs at large scale. As a mir- volume of data created by individuals and com- ror to the data-related sections, Sect.4 reviews panies is growing exponentially (see, e.g., Ma- user-related strategies, from preserving user pri- nyika et al. 2011). In parallel, analysts predict vacy to exploiting the force and intelligence of success to anyone who will exploit this data ac- the crowd. We discuss the future potential of re- curately, thus implicitly supporting the mean- viewed technologies and foreseeable extensions ingfulness of this data. However, this data is in Sect.5. everything unlike what companies are used to deal with. It is unstructured and redundant and 2 Data management potentially noisy or corrupted. Every piece of the data may be seen as noise that would pollute While information and communication networks a database of clean and structured data. There is have triggered the creation of an overwhelming nevertheless a clear intuition that the mass com- mass of data, they have also created opportunit- pensates for the defects of the pieces. A global ies to monitor and mine this data, thus augment- picture of the data should contain information ing drastically the volume of contextual informa- that could be exploited to many ends. This is tion potentially available. Massive collections of the challenge of the recent trend identified as big digital documents are made available, either pub- data (Big Data: Science in the Petabyte Era 2008; licly on the Web or in private networks related ERCIM News Special Theme: Big Data 2012). to companies, workgroups or social associations. Data may be very diverse and arise from any Big data has initially been characterised by its form of information exchange. Examples of this “three Vs” (Fayyad 2012), mostly addressing its include: technical specificity: • Textual: Web pages (personal, professional, • Volume: In itself, the volume of this data is an from individuals, groups or companies), emails, issue. It surpasses many of the simple storage blogs posts, news feeds, exchanges over social strategies classically used. At this scale and networks; evolution rate, it is hardly possible to structure • Multimedia: photos, videos, music, audiovisual and clean the data, for both technical and cost blogs, ...; reasons; Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 92 Stéphane Marchand-Maillet and Birgit Hofreiter

• Variety: In order to transform data into know- Many “Vs” have been added to big data (e.g., Vi- ledge, its multiple facets should be taken into ability, Veracity (Vossen 2013), Volatility, ...) but account (see Sect.3). The data in question the main “V” business is concerned with is therefore encompasses a high diversity in its • Value: The question is “how to make value out content, format, structure and interpretation. of this large, complex and unstable stream of Again, this goes against most principles of clas- data?” sical data management and storage paradigms where the structure of the data should be un- There are many answers to that question, includ- derstood and stable; ing: elocity • V : One of the main characteristics of big • By better understanding actions and behavior data is the pace at which it is generated and of its customers, traced by the log of their ac- at which it evolves and gets obsolete. In other tions, a company will be able to offer better words, this data inherently bears a strong tem- and more relevant services; poral dimension. Usage logs, trends, news, • By better understanding the context within are all content that have a strong interest in which it operates, characterised by the min- their immediate history and quickly decline ing of environmental factors, a company will into useless or even polluting data. However, lower its risks. this data may also have a behavioral interest at long-term on a more global temporal scale The first common step is always to transform (e.g., Morrison et al. 2012). data into knowledge, partly thanks to the technologies reviewed in the next sections. One finds These technical factors prevent the use of the typ- reports on success stories of big data analytics3 ical database models (e.g., a relational structure relating how such insurance company could fine- made usable via SQL) and impose to move onto tune its risk model using deep data analysis, in- more flexible, agile and scalable structures (including exploiting inferred customer social re- NoSQL 1 cluding the trend advocating for schema- lationships (which in turn poses questions on free storage or the MapReduce model (Dean and privacy - see Sect. 4.1. Looking at business as Ghemawat 2004; Mohamed and Marchand-Maillet a permanent complex constrained optimisation 2 2012) to support indexing, e.g., via the agnostic problem, where the right balance should be found name-value pair model). Decentralised storage (“price vs volume, cost of inventory vs the chance and processing systems (a.k.a Peer-to-Peer sys- of a stock-outrisk”3, etc), the success of big data is the Cloud tems or ) rely on structures having in providing insights into how to rationalise de- these characteristics to make the data safe, ubi- cisions on these tradeoffs. In that case, the noise quitous, and accessible (“Anything, Anywhere, in the data refers to any potential inconsistency Anytime”). in data patterns, unintentional (e.g., failures in In practice, the current appeal for big data has process or communication), or intentional (e.g., mapped into new functions coined as data scient- spam (Mukherjee et al. 2012)). ists, i.e, data analysts able not only to perform It should be emphasised there that as much as technical operations on the data such as prepara- there may be some benefit for a company to ex- tion, cleaning, compaction, etc, but also to contex- tract knowledge from the data, there is an inher- tualise the data and read it with all surrounding ent risk that this data is used by an adversarial parameters (e.g., social context, as detailed in party in many ways. These includes industrial Sect. 4.3). 3e.g., http://www.forbes.com/sites/mckinsey/ 1Web-scale databases: http://nosql-database.org 2012/12/03/big-data-advanced-analytics-success 2The Apache Hadoop library: http://hadoop.apache.org/ -stories-from-the-front-lines Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Big Data Management and Analysis for Business Informatics 93

espionage for adversarial reasons and unfair com- sector and the associated reliance on government petition, signals intelligence collection and ana- financial resources place public agencies under lysis4 for (state) security reasons and customer strong political influence. As a result, informa- privacy breach or exposure5, either intended or tion management systems in public organisations accidental. This then forces to ensuring the secur- emphasise more the environmental factors rather ity of the data and the privacy of the user, which than internal characteristics from the organisa- we study next. tion. As demonstrated in Conklin (2007); Wang (2009), these differences play an important role in 2.2 Data security the diffusion of technology in e-government settings. In particular, decisions made on informa- Securing data first aims at preserving its integ- tion security management in public organisations rity and confidentiality, while not imposing con- do not always follow technological rationales straints on its availability to authorised parties. (Ruighaver et al. 2007). Issues linked to its authenticity and accountability Technologically, the challenge is to define secur- are related to its integrity, while data access is ity strategies in the Information Management Sys- characterised by non-repudiation and reliability. tem that will support business processes (see, e.g., Data security is related to data usage and there, Diesburg and Wang 2010 for technical surveys related services include user authentication, user on digital data security). As given in Place and authorisation, access accountability and user reli- Hyslop (1982), Information management focuses ability. This finally mirrors to user privacy and on “plans and activities that need to be performed trust, studied in detail in Sect. 4.1 and Sect. 4.2, to control an organisation’s records”. Here, secur- respectively. The relationship between data se- ity should “ensure the continuity and minimise curity and user privacy is established via “guar- business damage by preventing and minimising anteeing privacy by securing access to private the impact of security incidents” (Solms 2006, data”. On a sociological level, the Security Cul- 2010). Authors of the latter references structure ture (Alnatheer et al. 2012) is defined as the level the evolution of security policies into several of belief and expectations members of a group waves (illustrated in Fig.2) showing the focus of (e.g., an organisation or company) have regard- every development phase, from purely technolo- ing security. This is valid at all levels of the gical to management-related concerns. organisation and, in Ruighaver et al. (2007), it is demonstrated that the effectiveness of operational security policies in an organisation is posit- Cyber Security Wave ively correlated with the belief that the top management (decision makers) have in these policies. Governance Wave Greene and D’Arcy (2010) verify empirically the Institutional Wave hypothesis that security culture (i.e., beliefs) and Management Wave job satisfaction lead to a increased security beha- Technical Wave vior in the organisation. - 1983 1985 - 1990 1995 - 2000 2000 - 2005 2006 -

A further distinction regarding the integration Figure 2: The 5 waves of Information Security (created of data security is to be made between the pub- after Solms 2010). The drift from technical to societal lic and private sector. The absence of economic issues is clearly visible markets for final product outputs in the public

4e.g., the Echelon Network (2001) over communication networks, or the Prism Initiative (2013) over the Web From its origins (first wave), information security 5e.g., the case of “AOL user 927” (2006) has been seen from a technical perspective. With Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 94 Stéphane Marchand-Maillet and Birgit Hofreiter

the stability of the technical solutions, the ques- 3 Data Intelligence tion has moved onto the integration of security Business intelligence is related to an accurate practices at a management (second wave), institu- use of the data collected at large scale. The tional (third wave) and governance (fourth wave) main aspects of this adequate management is the levels. The interweaving of private and public accessibility and legibility of knowledge where communication networks (e.g., private compan- available, and the creation of knowledge by auto- ies exposing themselves on the Web) has gen- mated or supervised processes. Figure3 schem- erated the related security issues of cybersecur- atises the stages for the creation of knowledge ity (fifth wave). Here, the construction of a se- from data, leading to accurate decision-making cure context over highly complex interconnected support. communication networks (cybersecurity ERCIM News Special Theme: Cybercrime and Privacy Is- sues 2012) should go with the help of reference organisations such as the ISO/IEC (COBIT (Control Data Organising Objectives for Information and related Techno- "Cleaning" logy) 4.1: Framework for IT Governance and Con- Mining Summarising trol Last retrieved: June 2013; ISO/IEC 27002:2005. Filtering Information technology – Security techniques – Code of practice for information security man- Information agement 2005). These institutions supervise the Analysis creation of standards whose role is to protect Recognition Synthesis an organisation’s information asset in the context of confidentiality integrity and availability. Knowledge Standards are generally largely biased, depending on economical, political or simply technical Towards Decision-making interests. In every domain, the debate of which Figure 3: Creating knowledge from data standard is better always exists. Data security is no exception (see, e.g., Solms 2005). In the particular case of big data, valuable information is Data is the raw material that can be collected, potentially hidden within massive amounts of either as characteristics (of users, products, etc) or data. Hence, in this context, defining the cost- as traces of activities (logs, sensing, results, etc). benefit tradeoff of securing data is a hard task. Information is obtained by compacting data into As much as there are open technical challenges its coherent structures (e.g., patterns, summary). for securing data at very large-scale, there are This step consists in aligning the data onto a also strategic open questions on the overall gain model. Knowledge is obtained via processing of such efforts. In the context of business inform- Information with a high-level of understanding atics, the data often originates from customer (e.g., semantic understanding). This step consists behavior or input. Security questions therefore in matching the Information with known high- go beyond the technical benefits (e.g., the qual- level concepts. ity of data modelling), they encompass ethical issues, related to user rights issues related to data 3.1 Knowledge Management preservation (e.g., including privacy, as detailed The domain of knowledge management, via the in Sect. 4.1) definition of extensive data description schemes Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Big Data Management and Analysis for Business Informatics 95

offers solutions for accurate (semantic) query- characterised by the emergent surges of inform- based information or service access. The grail of ation patterns such as recurring keywords or Knowledge Management Systems (KMS) is reas- phrases within text, or repeating events in us- oning and inference. From a non-redundant, but age logs. A particular case of data mining, suited complete, knowledge representation, data con- to business informatics, is therefore Emerging tent or actor behaviour may be predicted and Trend Detection (ETD) (Kontostathis et al. 2003). It linked together. However, the complexity and studies flows of information along a timeline and induced costs of design, creation, maintenance extracts specific topic areas whose focus becomes and compatibility of such descriptions generally more important at a point of time. It should be impede their usage and development at large. viewed as an automated mining process, since These strategies nevertheless find applications the manual inspection of flows of information in well-understood, closed domains. Hence, be- at large-scale is simply not feasible. ETD is of sides their utility in representing knowledge via crucial importance for data analysis, event pre- ontologies and inference models aligned with diction and decision-making in many areas, in- the semantic web, KMS (Abramowicz et al. 2010; cluding business, finance, or politics. As such it Hu et al. 2010; Jiang et al. 2009) have been apis fully relevant for the analysis of big data. By plied to specialised domains, including domains identifying growing interests, actors of these do- related to business. This is the case for enterprise mains will be able to react accordingly and even modelling (Frank 2013) and cartography (Tribolet predict future evolution. For example, based on and Sousa 2013) and the modelling of business mining discussions over a social network, a com- processes (Abramowicz et al. 2009; Sanz 2013). pany may decide to create a new product associ- The reader is referred to the latter references ated with a trendy product (e.g., sensitive pens for further insights in future developments of for tablets) or, on the contrary, retract a product knowledge management in all aspects of busi- whose philosophy goes against current trends ness modelling. (e.g., large cars go against emerging “green” feeling). Investors may also anticipate fruitful niches 3.2 Data mining and filtering if they can detect emerging trends at an early stage. They may also learn from the past by min- Data mining is the unsupervised (or weakly su- ing historical data to understand what caused the pervised, where weak assumption may be made success or failure of such investment or product. over the inner structure of the data at hand) discovery of recurrent or coherent patterns in the ETD generally considers documents as being data (Fayyad et al. 1996; Rajaraman and Ullman aligned along a timeline and emerging trends 2012). It develops in parallel to the field of Ma- also appearing and growing along that timeline chine Learning (see, e.g., Domingos 2012) where (Ganesh et al. 2011). An early survey in (Kon- the aim of the supervised process is to teach the tostathis et al. 2003) lists and analyses systems machine specific decisions via the processing of proposed in the late 90’s that operate on textual examples. As such, data mining maybe coined as technical data such as the INSPEC database and Knowledge Discovery (finding recurrent patterns the IBM DB2 US Patent database. In Le et al. in the data), complementing knowledge manage- (2005), a technique also applying on the scientific ment, whereas machine learning is about know- literature is proposed to track trendy topics using ledge propagation (extending known decisions to counts and bibliographic measures along time. unknown situations), related to the field of Pre- Several temporal models have been proposed for dictive Analysis. The flow of information evolves the analysis of topics over time such as the Dy- with time and trends develop. Such trends are namic Topic Model (Blei and Lafferty 2006), Topics Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 96 Stéphane Marchand-Maillet and Birgit Hofreiter

over Time (Wang and McCallum 2006) and the opinion mining explores blog texts, customer re- Trend Analysis Model (Kawamae 2011). views or comments to track the acceptance or re- jection of a product, an idea or a decision within The medical literature, notably with the availabil- a population (customers, voters, etc). Several ap- ity of the PubMed database is a domain of interest proaches exist, including using sentiment diction- for ETD (Mörchen et al. 2008). Goorha and Ungar aries to map text words to opinions or sentiments (2010) apply it to news wire articles, blogs posts, with polarity (is/is not) (Liu 2012). review sites and tweets, in search for interest rises in products or companies. A huge flow of 3.3 Information access: retrieval, information is processed daily based on word and filtering and browsing phrases counts. Leskovec et al. (2009) correlate the appearance of given phrases in news with its Search and retrieval operations have installed occurrence in blogs. Similar studies have more themselves as a base paradigm for accessing items recently be applied on Twitter data (e.g., Weng from within a repository. They are mostly based et al. 2010). on the notion of a query formulation (Baeza- Yates and Ribeiro-Neto 2011). (Seidel et al. 2008), Collaborative and hybrid recommender systems for example, demonstrates how such tools may (Park et al. 2012) leverage the wisdom of the support creativity in a business context. crowd and propagate user interests across a community. They can result in the emergence or fall Since precise data description is often a costly of an item, an idea by aggregating and propagat- operation (or simply incompatible with the pace ing adequately consistent user judgements. Col- at which data is produced), in the case of systems laborative filtering operations may be seen as a operating over poorly described or non-textual local form of mining and trend detection within data, the idea of query-by-example has emerged user interests. As such, they are also very close (Rui et al. 1998) as a help to construct accurate to the notion of crowdsourcing (see below). The queries. Positive and negative examples are ag- main idea is to create a bipartite graph between gregated over intermediate search operations, in products and customers where user ratings (judg- order to form a descriptive set for the sought ments) are used as edge weights. Information is items. Examples then become the base for online then propagated along this graph to group cus- learning operations, so as to generalise classes of tomers and/or products and thus, predict new provided relevant and non-relevant items (Bruno edge weights (i.e., the judgement of a costumer et al. 2007; Wyl et al. 2011). over a product). Browsing systems have been proposed and are This framework for recommendation is used in also mostly based on the definition of a search Selke and Balke (2011) to cater for the lack of objective (Heesch 2008). Such systems are typic- relevant or accurate information available to cus- ally oriented towards the localisation of a known tomers over “experience products”. Authors demon- information, be it media copy detection or user’s strate the effectiveness of their technique in the mental model localisation (Ferecatu and Geman context of movie recommendations. This relates 2009) (see also Fig.4). They iterate user judge- the idea of creating online and automatically item ments over appropriately-chosen sample sets of descriptions and therefore also relates to inform- information to estimate the target item the user ation retrieval. An early study on how such sys- has in mind. This framework has been extended tems may be formally evaluated is proposed in by (Lofi et al. 2010) from photo search to product Herlocker et al. (2004). browsing for mobile e-commerce. Whereas collaborative filtering uses implicit or Adaptive Hypermedia (AH) also relies on the nav- explicit user judgements, sentiment analysis (a.k.a igation paradigm for information exploration Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Big Data Management and Analysis for Business Informatics 97

to resolve the issue of complex query formula- Information filtering comes as a helper solution tion. As accurately given in De Bra et al. (2004): for the interactive formulation of search queries. Rules are defined over product characteristics, in “The core problem in finding the inform- order to define the class of the sought items as the ation you want, in all the above cases, is intersection of solution sets for the rules. Rules describing what you want. Results from are generally based on information facets. Facets search engines are often disappointing be- are orthogonal, mutually exclusive dimensions cause most search requests are too short and of the data whose range is quantised in relevant unspecific to yield good results. Once a Web intervals (Hearst 2008). Facets may be determ- site with interesting information is found, it ined from the data model itself by highlighting is often difficult to navigate to interesting important characteristics of the data. In explor- pages only, because the site can only be nav- atory conditions however, i.e., when the data is igated using its predefined link structure, not fully understood, it may be interesting to independently of the search request that let facets emerge automatically or interactively brought you to that site. The community for providing interpretation of its organisation of user modelling and adaptive hypermedia and to facilitate its exploration (Zwol and Sigur- offers solutions for this problem: using in- björnsson 2010). Several routes may be taken to formation gathered about the user during automatically determine data facets. They all con- the browsing process to change the inform- sist in using the data or a representative sample ation content and link structure on-the-fly. in a mining process to identify a reduced set of User modelling captures the mental state of orthogonal projection operators whereby every the user, and thus allows that knowledge to data item is identified by its set of projections. be combined with the explicit queries (or links) in order to determine precisely what Faceted search is extensively used over e-com- the user is looking for. To support the design merce sites when products bear inherent ortho- of this user model-based adaptation, refer- gonal characteristics. For example, this is the ence models like AHAM (De Bra et al. 1999; case for real estate commerce with facets such as Wu 2002) and Munich (Koch and Wirsing product type, surface, region, price range, ... 2002), both based on the Dexter Model by Halasz and Schwartz (1994), have been in- 3.4 Visual analytics troduced in an attempt to standardise and unify the design of adaptive hypermedia The above tools are used to make sense of the applications, used mostly in isolated inform- data itself, using the intrinsic data content or the ation spaces such as an online course, an usage context of the data. Visual analytics refers electronic shopping site, an online museum, to “the science of analytical reasoning facilitated etc”. by interactive visual interfaces” (Heer and Schnei- dermann 2012; Keim et al. 2010; Thomas and In Brusilovsky (2001), a taxonomy of AH tech- Cook 2006) and creates a link between content- nologies is further presented. The taxonomy is based data mining and interactive data explora- analysed in detail in Stash (2007), along with an tion techniques, as described above. extensive review of AH systems. Visual Analytics supports the user in exploring The above involves the notion of user modelling the data and to interactively guide the system to and a comprehensive review on personalisation find a formal solution that matches an intuitive research in e-commerce is presented in Adolphs solution (the mental model) to the problem (see and Winkelmann (2010). Fig.4). Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 98 Stéphane Marchand-Maillet and Birgit Hofreiter

As much as data should be secured, the privacy of a user should be guaranteed. This will allow the user to act freely in the environment s-he is confronted with. Consequently, a consistent and Data reliable user behaviour may lead to trust and high reputation that may be used in many contexts for information access and recommendation. Use case Data Analysis Tool Thanks to ever-developing communication me- (problem) (support) dia, users may also group into communities and form social groups. The emergence and identification of such social networks allows the ana- Interactive lysis to move from the individual to the proto- Mental Model Visualisation typical user-community s-he belongs to. This electronic crowd represents a task-force and a User mass of semantic knowledge that crowdsourcing (decision) efforts aim at capturing.

Figure 4: The process of Visual Analytics where the user 4.1 User privacy is matching a mental model of the solution with the User privacy (Danezis and Gürses 2010; ERCIM knowledge inferred from the data (see also Fig.3) News Special Theme: Cybercrime and Privacy Issues 2012; Hansen et al. 2008) is directly related to data security. The relationship between data In Zhang et al. (2012), a review of commercial sys- security and user privacy is established via “guar- tems for Visual Analytics, to support facing this anteeing privacy by securing access to private big data era is proposed in the context of Busi- data”. ness Intelligence. Various use cases are explored User privacy can be understood as a two-fold (inc. medical, microblogging) and performance concept, ethical and technological. Ethics should over factors such as scalability and effectiveness prevent the usage of user data to infer specific for supporting decision-making are given. Au- user needs and thus make that user fragile over thors then issue a number of future challenges communication networks. User data should be related to effective large-scale data analysis. studied statistically and anonymously so that it One of the key parameters in interactive data returns to the user as a member of one user class, analysis is to offer proper user interfaces and to not as an individual. Many more ethical aspects adequately leverage the potential of user inter- should be defined in parallel of the advent of action, seen as a source of semantic knowledge big data (Davis 2012). This is the role of govern- into the system (Morrison et al. 2012). The role mental or not-for-profit independent organisa- of users and user groups is studied in the next tions (e.g., the UN World Trade Organisation) to sections. counter the temptation of inadequate usage of this data from large Internet companies, even though it is known that individuals value their 4 From the user to the community privacy but tend to give it up easily as customers While an accurate use of the data is fundamental (Pogue 2011). to the decision process, the ultimate actor in the Technological solutions should ensure that the process remains the user. There are many user- user data and behavior (e.g., mirrored into us- related issues technologies should take care of. age logs) remain private and are not accessible Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Big Data Management and Analysis for Business Informatics 99

in their raw form to anyone. Anonymity may 4.3 Social Network analysis be a solution to privacy (Edman and Yener 2009), The constitution of social communities and but again, this approach may not be feasible any- groups of interest have allowed to move from more, as soon as the individual becomes a cus- the individual perspective to mass-address for tomer or a user of social networks (De Cristofaro business (essentially for push-based advertise- et al. 2012; Fung et al. 2010). ment, the main revenue model for the Web). The Also related to privacy is the possibility for se- study of social networks is therefore essential cure authentication (Poller et al. 2012), prevent- to structure the potential of such communities, ing identity spoofing. These fields, associated to including via the detection of key network fea- digital forensic and secured biometrics, directly tures such as connectivity and influential nodes relate to the notion of trust over communication (Gomez-Rodriguez et al. 2012; Sun et al. 2013). networks. In relation to adaptive hypermedia and recom- 4.2 Modelling trust mendation systems, where it is the study of user interaction that leads to recommendation, the Trust is a social notion that an individual or a study of social media (media hyperlinked in so- group (persons or organisation) develops over cial networks) may allow the inference of re- time and along experience. It measures the belief commendation (friends over Facebook or connec- that the actions of an individual or a group may tions over LinkedIn) (Backstrom and Leskovec be predicted (e.g., from social knowledge of the 2011). One of the difficulties here is the scale at individual or group) and stay within the limits of which algorithms should operate. A compensat- a predefined frame. Trust is closely related to the ing advantage of human-structured networks is notion of reputation (Castelfranchi and Falcone their reputed low diameter (originally valued to 6 1998; Pinyol and Sabater-Mir 2013). It is opposed (Schnettler 2009), but said to be reduced to 4 over to the adverse behaviour of cheating via fraud social networks) enabling local computations. and attacks (Hoffman et al. 2009). As such, the estimation of trust and reputation represents the 4.4 Social labor: Human Computation estimation of a risk for the environment where and crowdsourcing the individual or group in question is active. There is a large labor potential to leverage over Models for trust and reputation over communic- the Internet. This is known as Human Com- ation networks such as the Internet have been putation (Ahn 2005; Quinn and Bederson 2011) proposed with essentially two approaches. The and also relates to crowdsourcing (Jones 2013). game theory approach formalises a competition This strategy is, for example, used to help di- context where the objective is to maximise payoff gitising characters via the ReCAPTCHA (Com- with minimising risks. Trust estimation therefore pletely Automated Public Turing Test To Tell relies on associated risk-minimisation tools. The Computers and Humans Apart) system. Here, cognitive approach accounts for elements such the trust in the user is evaluated by presenting as beliefs, goals, desires and intentions. As such a problem with a known answer. The answer to the resulting trust models bear as much value an unknown problem proposed simultaneously in their result than in their capability to explain is then used as a statistical clue towards the the result. A thorough review of these models right solution of this latter problem. In general, and their classification is proposed in Pinyol and these tools, along with the Games With a Purpose Sabater-Mir (2013). These models are important (GWAP6), use the fact that human capabilities to estimate the value of user interaction in sys- to perform (visual) pattern recognition surpass tems such as recommender systems (Maida et al. 2012) (see also Sect. 3.3 above). 6Games With a Purpose: http://www.gwap.com Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 100 Stéphane Marchand-Maillet and Birgit Hofreiter

by far that of an automated process, with the 5 Discussion and conclusion incentive of fun or commercial advantage. Re- Modern communication networks have fueled commender systems may also be seen as a form the creation of massive volumes of data. In this of crowdsourcing in that they seamlessly federate paper, we have discussed how this data may be- user judgements to create semantic information come an asset for business activities. This thor- about items, products or services. ough overview of the information landscape, aug- mented with a large number of key references The impressive performance of such collaborat- aims at providing a faithful picture and guideline ive systems demonstrate the potential of labor to for the practitioner who wants to attack the prob- be federated over the internet. Another way of lem of data management and analysis in a busi- federating the crowd as a workforce is the use ness context. We highlighted and exemplified the of digital labor (Larson et al. 2012). For example, potential benefits of data analysis but also the the Amazon Mechanical Turk mediates between complexity and issues related to this task. We job requesters and workers. A requester creates advocated for considering in parallel the data a HIT (Human Intelligence Task) and proposes and the user viewpoints. Both perspectives share a reward for it. This HIT generally consists of commonalities in their structure and analysis. a short but repetitive task such as asserting the The first being that most of the data originates presence of an object in an image. The trust into from the users and that the users will then be workers’ competences may be evaluated by ini- modeled (in their behavior) via the analysis of tial trials and a reputation system is active for data. Further, as much as data may be seen at both workers and requesters. different scales, user and user communities may be modeled at different scales. There is therefore Eventually, if enough workers act on a simple much to gain in keeping this relationship alive task, this workforce constitutes a parallel pro- when exploring and exploiting the data. click-farms cessing machine (e.g., the to cheat In this era of big data, large-scale data analysis Internet ads) and software APIs have even been becomes a strategic field of development. The developed to make that process fully transparent. promise of a reasoning machine by the field of artificial intelligence in the 1960’s has been replaced by the statistical crunching of massive 4.5 Social knowledge : Folksonomies data with the side effect of smoothing out interesting details. The original three Vs of big data While human labor may be organised over the impose shallow processing for scalability. It is Internet, there are also several initiatives to feder- still an open challenge to design scalable process ate human knowledge, following the “Wisdom of to filter (project or denoise, however a volume the Crowd” paradigm (Surowiecki 2004). Beyond reduction strategy may be based on) the data to the ever-growing Wikipedia and its collaborat- lower volumes and enable more effective ana- ive edition model, including trust and reputation lysis. In parallel, distributed infrastructures ac- mechanisms, the combination of the semantic commodating hierarchical processing of the data web and Web 2.0 for social behaviour enables may help finding the essence of information and the gathering of a social knowledge, known as focus on these sparse interesting needles in the folksonomy (Lohmann and Díaz 2012). This know- data haystack. It is also a commonplace that the ledge is made accessible accessible to machines potential of big data for business profitability is 7 via semantic web technologies and also offers a more an intuition than a frequent reality . Hence, great potential for the development of adaptive 7https://www.facebook.com/dan.ariely/posts/ or human-tailored business services. 904383595868 Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Big Data Management and Analysis for Business Informatics 101

not only effective processing and efficient infra- References structures are needed but also the right analysis models are still lacking. Abramowicz W., Haniewicz K., Kaczmarek M., Zyskowski D. (2009) Semantic Modelling of The ubiquity of communication platforms and Collaborative Business Processes. In: Kusiak networks have shaped the culture and raised new A., goo Lee S. (eds.) eKNOW. IEEE Comp. concerns and approaches towards privacy and Soc., pp. 116–122 security issues. Further deployments are likely Abramowicz W., Fensel D., Frank U. (2010) Se- to be guided by the further integration of con- mantics and Web 2.0 Technologies to Sup- nected hardware into our everyday lives. The port Business Process Management. In: Busi- news ways of interacting with the data that these ness & Information Systems Engineering 2(1), devices create may be exploited for further aug- pp. 1–2 menting the ubiquity and usefulness of the data Adolphs C., Winkelmann A. (2010) A rigorous for the customer. In turn, these powerful sensing literature review on personalization research devices will enable companies to tailor recom- in e-commerce (2000-2008). In: Journal of Elec- mendation and targeting systems even further. tronic Commerce Research 11(4), pp. 326–341 Where the original Web was about simple data, von Ahn L (2005) Human Computation. PhD the Web 2.0 about people and their relationships, thesis, Carnegie Mellon University Last Ac- the emergent web of things (suggested as “Web cess: (UMI Order Number: AAI3205378) Alnatheer M., Chan T., Nelson K. (2012) Under- 3.0”) proposes to connect “Anything, Anywhere, standing And Measuring Information Secur- Anytime”. Objects will be part of the communica- ity Culture. In: Pacific Asia Conference on tion process and their usage, location, proximity, Information Systems (PACIS2012) Proceed- etc will be tracked for better user behavior under- ings. Ho Chi Minh City, Vietnam standing. Extending the concept of Tangible User Backstrom L., Leskovec J. (2011) Supervised ran- Interface, devices may symbol any data or virtual dom walks: predicting and recommending entity and be moved, exchanged, or combined as links in social networks. In: Poceedings of any object can be, with an associated impact in the WSDM’2011 Conference, pp. 635–644 the virtual world. Baeza-Yates R., Ribeiro-Neto B. (2011) Modern Information Retrieval: the concepts and tech- Affective computing (Picard 2000), assessing user nology behind search, 2nd. Add. Wesley emotions via physiological sensors will also allow Big Data: Science in the Petabyte Era. Nature, providers to penetrate even further into users’ vol. 455, num. 7209. Last Access: (special is- wishes. The trade-off between privacy and utility sue) is thus very likely to continue evolving. Blei D. M., Lafferty J. D. (2006) Dynamic topic models. In: Proceedings of the 23rd interna- 6 Acknowledgments tional conference on Machine learning. ICML ’06. Pittsburgh, Pennsylvania, pp. 113–120 Bruno E., Kludas J., Marchand-Maillet S. (2007) This work has been partly funded by Swiss Na- Combining Multimodal Preferences for Multi- tional Science Foundation (SNSF) via the Swiss media Information Retrieval. In: Proceedings NCCR(IM)2 and the Swiss State Secretariat for of the international workshop on Workshop Education and Research (SER) supporting ICT on multimedia information retrieval COST Action IC1002 “Multilingual and Multifa- Brusilovsky P. (2001) Adaptive Hypermedia. In: ceted Interactive Information Access” (MUMIA) User Modelling and User-Adapted Interac- . tion 11, pp. 87–110 Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 102 Stéphane Marchand-Maillet and Birgit Hofreiter

Castelfranchi C., Falcone R. (1998) Social Trust. ERCIM News Special Theme: Big Data. 89 In: Proceedings of the first workshop on de- ERCIM News Special Theme: Cybercrime and ception, fraud and trust in agent societies. Privacy Issues. 90 Minneapolis, USA Fayyad U., Piatetsky-Shapiro P., Smyth P. (1996) COBIT (Control Objectives for Information and Knowledge Discovery and Data Mining: To- related Technology) 4.1: Framework for IT wards a unified framework. In: Proceedings Governance and Control Institute, Informa- of the ACM SIG KDD Conference tion Systems Audit and Control Association Fayyad U. (2012) Big Data Analytics: Applica- (ISACA) http://www.isaca.org tions and Opportunities in On-line Predict- Conklin W. (2007) Barriers to Adoption of ive Modeling. In: Proceedings of the ACM e-Government. In: System Sciences, 2007. SIGKDD Conference. (material available on- HICSS 2007, pp. 98–98 line – June 2013) Danezis G., Gürses S. (2010) A critical review Ferecatu M., Geman D. (2009) A statistical frame- of 10 years of privacy technology. In: Surveil- work for image category search from a men- lance Cultures: A Global Surveillance Soci- tal picture. In: IEEE Transactions on Pat- ety? UK tern Analysis and Machine Intelligence 31(6), Davis K. (2012) Ethics of Big Data. O’Reilly Me- pp. 1087–1101 dia Frank U. (2013) Enterprise Modelling: the next De Bra P., Houben G.-J., Wu H. (1999) AHAM: steps. In: IEEE Conference on Business In- A Dexter-based Reference Model for Adapt- formatics (IEEE–CBI 2013). Vienna, Austria ive Hypermedia. In: Proceedings of the 10th Fung B. C. M., Wang K., Chen R., Yu P. S. ACM conference on Hypertext and Hyperme- (June 2010) Privacy-preserving data publish- dia. Darmstadt ing: A survey of recent developments. In: De Bra P., Aroyo L., Chepegin V. (2004) The Next ACM Comput. Surv. 42(4), 14:1–14:53 Big Thing: Adaptive Web-Based Systems. In: Ganesh M. S., Reddy C. P., N.Manikandan, Ven- Journal of Digital Information 5(1) kata D. P. (2011) TDPA: Trend Detection and De Cristofaro E., Soriente C., Tsudik G., Willi- Predictive Analytics. In: International Journal ams A. (2012) Hummingbird: Privacy at the on Computer Science and Engineering 3(3) Time of Twitter. In: Security and Privacy (SP), Gomez-Rodriguez M., Leskovec J., Krause A. 2012 IEEE Symposium on, pp. 285–299 (Feb. 2012) Inferring Networks of Diffusion Dean J., Ghemawat S. (2004) MapReduce: Sim- and Influence. In: ACM Trans. Knowl. Discov. plified Data Processing on Large Clusters. In: Data 5(4), 21:1–21:37 OSDI’04: Sixth Symposium on Operating Sys- Goorha S., Ungar L. (2010) Discovery of signific- tem Design and Implementation. San Fran- ant emerging trends. In: Proceedings of the cisco, CA 16th ACM SIGKDD international conference Diesburg S. M., Wang A.-I. A. (Dec. 2010) A sur- on Knowledge discovery and data mining. vey of confidential data storage and deletion Washington, DC, USA, pp. 57–64 methods. In: ACM Comput. Surv. 43(1), 2:1– Greene G., D’Arcy J. (2010) Security Culture and 2:37 the Employee-Organization Relationship in Domingos P. (Oct. 2012) A few useful things to IS Security Compliance. In: Proc. of the 5th know about machine learning. In: Commun. Annual Symposium on Information Assur- ACM 55(10), pp. 78–87 ance. New York, USA Edman M., Yener B. (Dec. 2009) On anonym- Halasz F., Schwartz M. (1994) The Dexter Hyper- ity in an electronic society: A survey of an- text Reference Model. In: Communications of onymous communication systems. In: ACM the ACM 37(2), pp. 30–39 Comput. Surv. 42(1), 5:1–5:35 Hansen M., Schwartz A., Cooper A. (2008) Pri- Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Big Data Management and Analysis for Business Informatics 103

vacy and Identity Management. In: Security timestamps. In: Proceedings of the Forth In- Privacy, IEEE 6(2), pp. 38–45 ternational Conference on Web Search and Hearst M. A. (2008) UIs for Faceted Navigation: Web Data Mining, WSDM 2011, Hong Kong, Recent Advances and Remaining Open Prob- China, pp. 317–326 lems. In: Workshop on Computer Interaction Keim D., Kohlhammer J., Ellis G., Mansmann F. and Information Retrieval, HCIR. Redmond, (eds.) Mastering the Information Age – Solv- WA ing Problems with Visual Analytics. Euro- Heer J., Schneidermann B. (2012) Interactive Dy- graphic Digital Library namics for Visual Analytics. In: Communica- Koch N., Wirsing M. (2002) The Munich Ref- tion of the ACM 55(4) erence Model for Adaptive Hypermedia Ap- Heesch D (2008) A survey of browsing models plications. In: 2nd International Conference for content based image retrieval. In: Multi- on Adaptive Hypermedia and Adaptive Web- media Tools and Applications 40 (2) Based Systems, pp. 213–222 Herlocker J. L., Konstan J. A., Terveen L. G., Kontostathis A., Galitsky L., Roy S., Pottenger W. Riedl J. T. (Jan. 2004) Evaluating collaborat- M., Phelps D. (2003) A survey of ETD in Tex- ive filtering recommender systems. In: ACM tual Data Mining. In: Berry M. (ed.) A Com- Trans. Inf. Syst. 22(1), pp. 5–53 prehensive Survey of Text Mining. Springer Hoffman K., Zage D., Nita-Rotaru C. (Dec. 2009) Larson M., Soleymani M., Eskevich M., Serdy- A survey of attack and defense techniques ukov P., Jones G. J. (2012) The Community for reputation systems. In: ACM Computer and the Crowd: Developing large-scale data Surveys 42(1), 1:1–1:31 collections for multimedia benchmarking. In: Hu S., Wan L., Zeng R. (2010) Web2.0-based En- IEEE Multimedia, Special Issue on Large- terprise Knowledge Management Model. In: Scale Multimedia Data Collections Information Management, Innovation Man- Le M.-H., Ho T.-B., Nakamori Y (2005) Detect- agement and Industrial Engineering (ICIII), ing Emerging Trends from Scientific Corpora. 2010 International Conference on Vol. 4, In: International Journal of Knowledge and pp. 476–480 Systems Sciences 2(2) ISO/IEC 27002:2005. Information technology – Leskovec J., Backstrom L., Kleinberg J. (2009) Security techniques – Code of practice for Meme-tracking and the dynamics of the information security management (update of news cycle. In: Proceedings of the 15th ACM ISO/IEC 17799) International Standard Organ- SIGKDD international conference on Know- isation ledge discovery and data mining. KDD 2009. Jiang H., Liu C., Cui Z. (2009) Research on Know- Paris, France, pp. 497–506 ledge Management System in Enterprise. In: Liu B. (2012) Sentiment Analysis and Opinion Computational Intelligence and Software En- Mining. Synthesis Lectures on Human Lan- gineering, CiSE 2009. International Confer- guage Technologies 1 Vol. 5. Morgan & Clay- ence on, pp. 1–4 pool Jones G. J. F. (2013) An introduction to crowd- Lofi C., Nieke C., Balke W.-T. (2010) Mobile sourcing for language and multimedia tech- Product Browsing Using Bayesian Retrieval. nology research. In: Proceedings of the In: 12th Conference on Commerce and En- 2012 international conference on Informa- terprise Computing (CEC 2010). Shanghai, tion Retrieval Meets Information Visualiz- China, pp. 96–103 ation. PROMISE’12. Springer-Verlag, Zinal, Lohmann S., Díaz P. (2012) Representing and Switzerland, pp. 132–154 visualizing folksonomies as graphs: a ref- Kawamae N. (2011) Trend analysis model: trend erence model. In: Proceedings of the Inter- consists of temporal words, topics, and national Working Conference on Advanced Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 104 Stéphane Marchand-Maillet and Birgit Hofreiter

Visual Interfaces (AVI’12). ACM, Capri Island, Electronic Identity Cards for User Authen- Italy, pp. 729–732 tication – Promise and Practice. In: Security Maida M., Maier K., Obwegeser N., Stix V. (2012) Privacy, IEEE 10(1), pp. 46–54 A Multidimensional Model of Trust in Recom- Quinn A. J., Bederson B. B. (2011) Human commender Systems. In: EC-Web, pp. 212–219 putation: a survey and taxonomy of a grow- Manyika J., Chui M., Brown B., Bughin J., Dobbs ing field. In: Proceedings of the SIGCHI Con- R., Roxburgh C., Byers A. H. (2011) Big data: ference on Human Factors in Computing Sys- The next frontier for innovation, competition, tems. CHI ’11. ACM, Vancouver, BC, Canada, and productivity. Mc Kinsey Global Institute pp. 1403–1412 Mohamed H., Marchand-Maillet S. (2012) Dis- Rajaraman A., Ullman J. D. (2012) Mining tributed media indexing based on MPI and Massive Datasets. Cambridge University MapReduce. In: Multimedia Tools and Applic- Press ations, pp. 1–25 Rui Y., Huang T. S., Ortega M., Mehrotra S. Mörchen F., Dejori M., Fradkin D., Etienne J., (1998) Relevance Feedback: A Power Tool in Wachmann B., Bundschus M. (2008) Anticip- Interactive Content-Based Image Retrieval. ating annotations and emerging trends in bio- In: IEEE Trans. on Circuits and Systems for medical literature. In: Proceedings of the 14th Video Technology 8(5) ACM SIGKDD international conference on Ruighaver A., Maynard S., Chang S. (2007) Or- Knowledge discovery and data mining. Las ganisational security culture: Extending the Vegas, Nevada, USA, pp. 954–962 end-user perspective. In: Computers and Se- Morrison D., Tsikrika T., Hollink V., de Vriesand curity 26(1) E. Bruno A. P., Marchand-Maillet S. (2012) Sanz J. L. C. (2013) Enabling Customer Ex- Topic modelling of clickthrough data in im- perience and Front-office Transformation age search. In: Multimedia Tools and Applic- through Business Process Engineering. In: ations, pp. 1–23 IEEE Conference on Business Informatics Mukherjee A., Liu B., Glance N. (2012) Spotting (IEEE-CBI 2013). Vienna, Austria Fake Reviewer Groups in Consumer Reviews. Schnettler S. (2009) A structured overview of In: Proceedings of the 21st International Con- 50 years of small-world research. In: Social ference on World Wide Web. WWW’12. Lyon, Networks 31(3), pp. 165 –178 France, pp. 191–200 Seidel S., Muller-Wienbergen F. M., Rosemann Park D. H., Kim H. K., Choi I. Y., Kim J. K. (2012) M., Becker J. (2008) A Conceptual Frame- A Literature Review and Classification of Re- work for Information Retrieval to Support commender Systems Research. In: Expert Sys- Creativity in Business Processes. In: Proceed- tems with Applications (in press) ings 16th EuropeanConference on Informa- Picard R. W. (2000) Affective Computing. MIT tion Systems. Galway, Ireland Press Selke J., Balke W.-T. (2011) Turning Experience Pinyol I., Sabater-Mir J. (2013) Computational Products into Search Products: Making User trust and reputation models for open multi- Feedback Count. In: 13th IEEE Conf. on Com- agent systems: a review. In: Artificial Intelli- merce and Enterprise Computing (CEC 2011). gence Review 40(1), pp. 1–25 Luxembourg Place I., Hyslop D. (1982) Records management: von Solms B. (2005) Information Security gov- controlling business information. Reston Pub. ernance: COBIT or ISO 17799 or both? In: Co. Computers and Security 24(2) Pogue D. (2011) Don’t Worry about Who’s von Solms B. (2006) Information security – the watching. In: scientific American 304(1), p. 32 Fourth Wave. In: Computer and Security Poller A., Waldmann U., Vowe S., Turpe S. (2012) 25(3), pp. 165–168 Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Big Data Management and Analysis for Business Informatics 105

von Solms B. (2010) The 5 Waves of Informa- data mining. Proceedings of the third ACM tion Security – From Kristian Beckman to the international conference on Web search and Present. In: Rannenberg K., Varadharajan V., data mining WSDM ’10. New York, New York, Weber C. (eds.) Security and Privacy – Silver USA, pp. 261–270 Linings in the Cloud. IFIP Advances in In- Wu H (2002) A Reference Architecture for Ad- formation and Communication Technology aptive Hypermedia Applications". PhD thesis, Vol. 330. Springer Berlin Heidelberg, pp. 1–8 Eindhoven University of Technology Stash N. (2007) Incorporating Cognit- von Wyl M., Mohamed H., Bruno E., Marchand- ive/Learning Styles in a General-Purpose Maillet S. (2011) A parallel cross-modal Adaptive Hypermedia System. PhD thesis, search engine over large-scale multimedia Eindhoven University of Technology, The collections with interactive relevance feed- Netherlands back. In: Proceedings of the 1st ACM Interna- Sun K., Morrison D., Bruno E., Marchand- tional Conference on Multimedia Retrieval. Maillet S. (2013) Learning Representative ICMR ’11. ACM, Trento, Italy, 73:1–73:2 Nodes in Social Networks. In: 17th Pacific- Zhang L., Stoffel A., Behrisch M., Mittelstadt S., Asia Conference on Knowledge Discovery Schreck T., Pompl R., Weber S., Last H., Keim and Data Mining. Gold Coast, AU D. (2012) Visual analytics for the Big Data Surowiecki J. (2004) The wisdom of crowds: era – A comparative review of state-of-the- Why the many are smarter than the few art commercial systems. In: Visual Analytics and how collective wisdom shapes business, Science and Technology (VAST), 2012 IEEE economies, societies and nations. New York: Conference on, pp. 173–182 Doubleday van Zwol R., Sigurbjörnsson B. (2010) Faceted Thomas J., Cook K. (2006) A Visual Analytics exploration of image search results. In: Pro- Agenda. In: IEEE Computer Graphics and Ap- ceedings of the 19th international conference plications 26(1), pp. 10–13 on World Wide Web (WWW’2010) Tribolet J., Sousa P. (2013) Enterprise Gov- ernance and Cartography. In: IEEE Conf. on Stéphane Marchand-Maillet Business Informatics (IEEE-CBI 2013). Vienna Vossen G. (2013) Big data as the new enabler in Department of Computer Science business and other intelligence. In: Vietnam Centre Universitaire Informatique (CUI) Journal of Computer Science 1(1), pp. 1–12 University of Geneva Wang J.-F. (2009) E-government Security Man- Carouge agement: Key Factors and Countermeasure. Switzerland In: Information Assurance and Security, 2009. [email protected] IAS09. Fifth International Conference on Vol. 2, pp. 483–486 Birgit Hofreiter Wang X., McCallum A. (2006) Topics over time: Electronic Commerce Group a non-Markov continuous-time model of top- Institute of Software Technology and Interactive ical trends. In: Proceedings of the 12th ACM Systems SIGKDD international conference on Know- Vienna University of Technology ledge discovery and data mining. KDD ’06. Vienna Philadelphia, PA, USA, pp. 424–433 Austria Weng J., Lim E.-P., Jiang J., He Q. (2010) Twitter- [email protected] Rank: finding topic-sensitive influential twit- terers. In: Proceedings of the third ACM international conference on Web search and Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 106 Thomas Setzer

Thomas Setzer

Data-Driven Decisions in Service Engineering and Management

Today, the frontier for using data to make business decisions has shifted, and high-performing service companies are building their competitive strategies around data-driven insights that produce impressive business results. In principle, the ever-growing amount of available data would allow for deriving increasingly precise forecasts and optimised input for planning and decision models. However, the complexity resulting from considering large volumes of high-dimensional, fine-grained, and noisy data in mathematical models leads to the fact that dependencies and developments are not found, algorithms do not scale, and traditional statistics as well as data-mining techniques collapse because of the well-known curse of dimensionality. Hence, in order to make big data actionable, the intelligent reduction of vast amounts of data to problem- relevant features is necessary and advances are required at the intersection of economic theories, service management, dimensionality reduction, advanced analytics, robust prediction, and computational methods to solve managerial decisions and planning problems.

1 Introduction gain efficiency through data-driven decisions, an- ticipatory action and accelerated service support Increasingly automated data capturing, the ubi- and delivery processes. As an example, those quity of sensors, the spread of smart phones, and companies can utilise knowledge extracted from the penetration of life by social media leads to past customer behaviours to better understand enormous and ever growing amounts of data. customers in order to better convince them with Novel technological advances in analytics and smart, individualised offers and services. scalable data management promise to facilitate the capturing, storage, searching, sharing, analys- 1.1 Service Management ing, and visualisation of relationships and trends hidden in large, high-dimensional data sets. Traditionally, the aim of service management is to optimise service-intensive supply chains, While, traditionally, scientists in areas such as which are typically much more complex than the meteorology, genomics, physic simulations, or supply chains of typical goods. Those require environmental research were primarily faced with tighter integration with field service and third the challenges of exploring large, very high-di- parties and must also accommodate inconsistent mensional data sets, today such challenges also and uncertain demand by establishing more in- affect areas like business informatics. In par- tegrated and more robust information flows. In ticular service design and management need to addition, most processes must be coordinated process data in order to spot business trends, de- across numerous service locations. Interestingly, termine and anticipate bottlenecks and quality of among typical manufacturers, after-sale services service, or prevent customer churn by identify- (support, repair, maintenance, etc.) comprise less ing churn risk and triggering appropriate actions, than 20% of revenue. Among the most successful to name only a few tasks. In general, enterprises companies, those same activities on average gen- that can use their data quickly and correctly can erate more than 50 percent of their total profits Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Data-Driven Decisions in Service Engineering and Management 107

(Accenture 2006). This is one of many observa- allows for identifying novelty patterns in cus- tions indicating that a profound understanding tomer behaviour and improving short and long- of customers and business partners and establish- term performance of enterprise business systems, ing high-quality service and information man- which is vital for running a competitive service agement is of crucial importance. company.

However, today enterprises provide an increas- In ‘Competing on Analytics: The New Science of ing number of services in an automated or semi- Winning’, Davenport and Harris (2006) argue that automated fashion by means of information tech- the frontier for using data to make business de- nology (IT services), where customer behaviour cisions has shifted. Many high-performing com- and experience can only be ‘observed’ by track- panies are building their competitive strategies ing what a customer is doing, in particular how around data-driven insights that generate im- he uses one or more services over time. Providers pressive business results. Those companies use even of IT-only services can no longer afford to advanced analytical procedures, sophisticated focus on technology and their internal organisa- quantitative and statistical analysis and predict- tion, but need to consider the quality of the ser- ive modelling. Examples of analytics are the us- vices they provide and focus on the relationship age of novel tools to determine the most profitable customers and offer them the right price, to with customers. IT service management (ITSM) accelerate product innovation, to optimise and refers to the implementation and management integrate supply chains, and to identify the major of high quality IT services that meet the needs drivers of financial performance. Many examples of customers. ITSM is performed by IT service from organisations such as Amazon, Barclay’s, providers through an appropriate mix of people, Capital One, Harrah’s, and Procter & Gamble are process and information technology (Office of presented, showing how to leverage analytics to Government Commerce (OGC) 2009). drive business. However, various potential defin- Unfortunately, in particular with IT services, pro- itions for advanced analytics exist. Typically, the viders typically do not receive regular direct cus- ‘advanced’ indicates quantitative, predictive or tomer feedback that is required for marketing, prescriptive models as described later in this pa- further service improvements, and service innov- per. ation. However, there is an ever-growing amount of information how a customer uses a services 1.3 Big Data Analytics (e.g., sensors of a rental car, log files of a Web- Over the last two years, the term Big data is shop, browsing behaviour in on-line manuals, propagated by major companies offering inform- etc.), and these datasets can be analysed to get ation management software such as Intel1, SAP2, ‘implicit’ feedback as described for example in or IBM3, and has become more and more a syn- Choi and Ahn (2009). onym for data analysis and advanced analytics. For many SMEs and also for larger com- 1.2 Advanced Analytics panies, this is in some sense counter-productive as nowadays enterprises collect massive amounts In fact, today’s service enterprises have more 1http://www.intel.de/content/www/de/de/big-data/ data at hand about their markets, customers, and big-data-analytics-turning-big-data-into-intelligence. rivals than ever before. Analysing those vast html amounts of historical and current data in an auto- 2http://www54.sap.com/pc/tech/ nomic or semi-autonomic fashion allows for pre- in-memory-computing/hana/software/analytics/big-data. html dicting service demand and usage, customer be- 3http://www-01.ibm.com/software/data/infosphere/ haviour, and market dynamics. In addition, it hadoop/what-is-big-data-analytics.html Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 108 Thomas Setzer

of various metrics, such as historical sensor, mon- even if a company has Big Data, making use of itoring, and customer usage data, hoping that such data typically not only requires appropriate the data will turn out to be useful one day for tools but also data scientists with expertise and prediction and optimisation. know-how, hacking-skills, domain knowledge, and deep mathematical and data management Accordingly, as Big data analytics is now a pop- skills; unfortunately, as of yet data scientists of ular topic for management, many information that sort are still a very scarce human resource management companies offer tools and solutions (Davenport and Patil 2012). to extract and project relationships between a vast amount of high-dimensional data vectors The result is that – in practice – data are often (structured, semi-structured, or unstructured collected and then ignored or aggregated in a ones), and to process, reduce, correlate and inter- problem-agnostic fashion, and finally for most pret data in a much more flexible fashion com- problems rather simple and conservative solution pared to traditional database management and heuristics are applied by rules of thumb or using business intelligence systems. coarsened data. The authors of this article are not aware of many companies besides the financial Over the last years, enterprises such as Software institutions and telecommunications companies AG, Oracle, IBM, Microsoft, SAP, EMC, and HP that make excessive use of their collected data; have spent more than $15 billion on software however, most enterprises spend an increasing firms only specialising in data management and amount of money and effort in monitoring sys- analytics. Since the last three years, this industry tems and data collection. That is also the out- was worth more than 100 billion US-dollars and come of numerous studies and expert interviews was growing at around 10 percent a year: about conducted and summarised by Ross et al. (2013). twice as fast as the software business in general (The Economist 2010). Interestingly, already today leading data scientists are telling us that Big Data can and must be 1.4 The Curse of Dimensionality reduced intelligently to small data, so that finally for most decision problems one does not need While in principle the vast and ever-growing Big Data at all.4,5 sets of available data would allow for deriving increasingly precise predictions and optimised 1.5 Collecting the Right (Amount) of planning and decision models, the complexity Data resulting from the consideration of large volumes Large, global companies already recognise that of multivariate, fine-grained, often noisy and in- there is a need to stop collecting more data and complete data leads to the fact that relationships start a focused collection of the right data re- within the data are not found, algorithms do not quired to make decisions and to run a business scale, and traditional statistics as well as data- successfully (Nokia Siemens Networks 2013). mining techniques collapse because of the well- Suppose a company is gathering the right data: known curse of dimensionality (nowadays also attributes and dimensions really relevant for plan- called the curse of big data) (Bellman 1961; Lee ning and decision-making. There is still the ques- and Verleysen 2007). tion whether the return on adding more data

Despite these dimensionality-intrinsic problems, 4Big Data: Maybe You Don’t Need It : http://www. biases in how data are collected, a lack of context, datacenterjournal.com/it/big-data-dont/ gaps in what’s gathered, artefacts of how data are 5Most data isn’t big, and businesses are wasting money processed and the overall cognitive biases that pretending it is: http://qz.com/81661/most-data-isnt lead even experienced researchers to determine -big-and-businesses-are-wasting- non-existing patterns (and vice versa) shows that money-pretending-it-is/ Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Data-Driven Decisions in Service Engineering and Management 109

points diminishes after passing a certain volume reliable, more conservative allocation mechan- of data collection, or certain data granularities isms might be appropriate. (such as monitoring intervals), and if – in a par- If the forecasting horizon approaches zero time ticular situation – gathering additional data will intervals, conservative online mechanism should cost more than it will actually yield. be applied that allow for handling unexpected de- Cleary, an answer to that question depends on mand phases immediately, as sophisticated offline- the concrete enterprise planning and decision planning would not beneficial in such situations: problem, the importance of the problem, the plans would be invalid shortly after their compu- scalability of engines/algorithms processing the tation. data, the tolerance of the algorithms regarding This paper reviews theory and practice of data artifacts and noise, the skills of the managers reduction in service management with regard to processing and interpreting the data, and many the various targets addressed with the different more factors. data reduction techniques. First, we argue that a really efficient and intelligent data reduction However, independent of particular problems requires the prior definition of business problems and individual factors as aforementioned, the an- and algorithms how to address these problems swer also depends on purely statistical or math- with reduced data. Second, we argue that mathematical criteria regarding redundancy and noise ematical programs and algorithms for planning within the datasets. That is because such criteria and decision-making should not be applied in a can determine if another piece of data can bring data-agnostic fashion. In contrast, programs and novel information at all, or whether it can be algorithms should be sensitive and adjustable to fully or approximately derived from data already available data and the amount of dependencies, available (for example by means of collaborative reliability, and stochastics within data, which mechanisms such as regression or causal reason- typically vary over time, use-case, domain, and ing). planning horizon. Furthermore, for reasons of robustness and scalability it is disadvantageous to parametrise predic- 2 Data Understanding and Reduction tion models and mathematical decision programs The first and most important step in analytics is with correlated or even collinear data vectors. a proper understanding of the available data, the In fact, efficient decision mechanisms should be involved variables and how these are measured. rather elastic and adaptable to the anatomy and Data quality, appropriate data cleaning and hand- the information contained in the input data, while ling missing values as well as detecting outliers today typically the signatures and internal al- and errors must be performed prior to any data gorithms of enterprise decision modules are of analysis. Knowing that data preprocessing is ar- rather static nature. guable the most complex and time-consuming Consider a resource allocation mechanism for step in analytics, for now we assume these tasks enterprise services in a data centre. If demand have been already performed. forecasts were expected to be highly precise for We will now characterise various techniques to certain indicators over a defined period of time, reduce data to relevant features, structures, and a rather aggressive allocation mechanism oper- developments. In order to separate approaches ating with deterministic demand curves would aimed at descriptive, predictive, and prescript- be appropriate. Once the demand prediction tool ive analytics, we will group the techniques ac- downgrades its confidence levels and shrinks the cordingly. Descriptive analytics will be further horizon of the look-ahead period considered as differentiated in simple aggregations (Sect. 2.1), Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 110 Thomas Setzer

and approaches that exploit statistical dependen- derives the well established and broadly used re- cies in and between data objects and variables porting functions based on information stored (Sect. 2.2). In Sect. 2.3, we focus on data mining in data warehouses. Management dashboards approaches aimed at gaining knowledge from the usually provide the means of presenting such data to reduce uncertainty regarding the realisa- aggregated data to managers to support their tion of a particular variable (or label). A typical business decisions. task would be the determination of the probability of a positive response of a customer, and 2.2 Data Compression and the determination of data (features) necessary Approximation to learn this probability. In Sect. 2.4 we then summarise approaches to predict whole vectors The most generic way to reduce (and not just or time series. Finally, in Sect. 2.5, we focus on aggregate) data is to exploit dependencies in and prescriptive data reduction techniques that dif- between data vectors – in a problem-agnostic fer from prescriptive techniques as data selection way – by multivariate statistics and matrix ap- and reduction needs to be aligned with a particu- proximation techniques, mostly based on linear lar, potentially combinatorial and computational algebra. Examples are variance-preserving ap- very complex mathematical optimisation prob- Empirical Ortho- lem. In the latter case, the goal is not only to proximation techniques such as gonal Defactorisation Eigen-approaches gain insights and reduce uncertainty of future derived by Truncated Singular Value Decomposition values of data, but to select and transform data such as compact Principal Component Analysis PCA in a way that is beneficial for solving a particular or ( ). Independ- planning and decision problem More and more, techniques such as ent Component Analysis (ICA) are applied to derive more meaningful features (in contrast to 2.1 Data Aggregation for Descriptive solely reducing data). By exploiting communal- Service Analytics ities, such techniques are very useful to reduce data to the maximum amount of variation (as a The purpose of aggregating data for descriptive proxy for information) in the data sets and are service analytics is to summarise what happened often shown to derive the best low-dimensional in the past. For example, in Web analytics met- approximation of data in very useful mathemat- rics are considered such as number of page views, ical senses such as the L norm. conversion rates, check-ins, churns, etc. There 2 are literally thousands of such metrics, on their Other examples are topology-preserving tech- own typically simple event counters. Other ag- niques such as Local-Linear-Embedding (LLE) (Ro- gregations for descriptive service analytics might weis and Saul 2000) or isoMap (Tenenbaum et al. be the results of simple arithmetic operations, 2000), where the objective of data reduction is such as share of voice, average throughput, aver- not to capture maximum variance of the data age number of positive responds to a campaign, sets with fewer dimensions, but to preserve the etc. Most of what the industry called analytics topology of the data objects, i.e., their distance is nothing but applying filters on the data before relationships. computing the descriptive statistics, sometimes combined with a linear statistical forecast. For Likewise, multivariate techniques such as vector example, by applying a geo-filter first, a company quantisation and linear and non-linear regres- can get metrics such as average revenue per week sion techniques fall into this category of data from USA vs. average revenue per week from reduction according to pre-defined mathematical Europe. Structuring aggregated data to reports objectives. Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Data-Driven Decisions in Service Engineering and Management 111

2.3 Data Reduction by Information (see Tsymbal et al. 2002 as an example) propose Gain and other Criteria approaches that do not work (cluster) on original Unlike the approaches described in Sect. 2.1 and data but on reduced data as a result of compres- Sect. 2.2, the analysis step of discovering know- sion steps as described in Sect. 2.2. ledge in databases is aimed at discovering patterns in sets of data involving methods at the in- 2.4 Data Reduction for Predictive tersection of artificial intelligence, machine learn- Service Analytics ing, statistics, and database systems. The over- Predictive analytics is based on information ex- all goal is to extract pattern in a data set and tracted by the three previous data understanding transform it into structural dependencies for fur- and reduction steps; it uses all of the gained in- ther use. Aside from the raw analysis step, it sights to make robust prediction of developments involves database and data management aspects, of important indicators, metrics, and variables inference considerations, interestingness metrics, (Stewart et al. 2012). complexity considerations, post-processing of identified structures, visualisation, and on-line An intuitive way to understand predictive ana- updating mechanisms. lytics is to apply it to the time domain. The most familiar predictive analytic tool is a time series Typical goals are the automatic or semi-automatic model (or any temporal model) that summarises analyses of large quantities of data to extract pre- past trajectories found in the data, and use either viously unknown patterns such as groups of data auto- or (lagged) cross-correlations and regres- records (segmentation analysis), unusual records sion to extrapolate time series to a future time (anomaly detection) and dependencies via associ- where data is not yet existing. This extrapolation ation rules, decision trees, or other methods. For in the time domain is what scientists refer to as instance, data mining techniques might identify forecasting or prediction. multiple groups in the data, which can then be used to obtain more accurate prediction results Although predicting the future is a common use and more focused marketing campaigns by a de- case of predictive analytics, predictive models cision support system. Here, data is reduced to are not limited to predictions in temporal dimen- gain information about the general structure of sions. Such models can theoretically predict any- the data (clustering), or the class prediction of thing and, hence, predictive analytics are some- records with an unknown label due to similarit- what overlapping with data mining and know- ies with other records where labels are already ledge extraction as described in Sect. 2.3. The known. predictive power of a model needs to be prop- erly validated by criteria addressing the robust- As discussed in Sect. 1.4, clustering and classifica- ness of the prediction such as using pre-whitened tion do not perform well with high-dimensional predictors, perpendicularity of predictors, by us- data because of the curse of dimensionality. Beyer ing information criteria such as BIC or AIC, and et al. (1999) and Aggarwal et al. (2001), amongst finally out-of-sample testing using consecutive others, have shown that standard measures for samples. The essence of predictive analytics, in proximity or distance that are used for stand- general, is that we use existing data to build a ard k-means clustering, are becoming more and model. Then we use the model to predict data more meaningless with growing dimensionality. that doesn’t (yet) exist. To circumvent this problem, approaches as proposed in Aggarwal et al. (2001) introduce novel However, only with concrete use cases in terms distance calculations that are still meaningfull of business problems in mind, one can decide even in high-dimensional data space of 15 dimen- which pieces of information in the data set are sions and more. Alternative streams of research ultimately relevant for a company, and which Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 112 Thomas Setzer

pieces are not. This brings one directly to data 3 Information Gain versus reduction for prescriptive analytics that will be Optimisation Gain described in the next subsection. Each department of a service provider has a set of typical tasks to perform on an operational, 2.5 Data Reduction for Prescriptive tactical, or strategic level. Taking for instance Service Analytics the Customer Relationship Management (CRM) Prescriptive analytics not only predicts a possible department. CRM is aimed at the optimisation of future, it predicts multiple futures based on the a company’s interactions with current and future decision maker’s actions. Therefore a prescript- customers. Objectives of CRM are the reduction ive model is, by definition, also predictive and of overall churn by adequate customer service significant effort must be undertaken to guaran- and support, or by identifying and rewarding tee internal and external model validity. As it is customers that have been loyal over a period of seen today, a prescriptive model is actually a com- time but now show certain behaviours that in- bination of multiple predictive models running in crease churn probability (reduced call frequency, parallel, one for each possible input. Since a pre- churns of neighbor nodes in the telecommunic- scriptive model is able to predict the possible con- ation network, etc.) Another objective might sequences based on different choices of action, be the identification of customer segments for it can also recommend the best course of action particular campaigns such as cross-selling offers for any pre-specified outcome, given the data based on score-values of customers. Scores are set used to predict the future (together with its derived by data analytics and reflect the probab- confidence or uncertainty). The goal of most pre- ility of a certain customer to respond positively scriptive analytics is to guide the decision maker depending on a customer’s profile and past be- towards decisions that will ultimately lead to an haviour. Such procedures are aimed at gaining (near) optimal and robust business outcome. information from datasets regarding the probability of an unknown label in data records (for In prescriptive analytics, one also builds a pre- instance, class predictions such as churn: yes/no, dictive data model. However, the model must upselling: yes/no, etc.) and are in the primary have two more added components in order to focus of business intelligence solutions. be prescriptive. A company not only needs a However, usually strict business rules exist that rigorously validated predictive model, the model complicate the selection of target customers. As must be actionable, i.e., managers must be able a simple example, consider the case where one to take actions supported by the model. In addi- single customer is not allowed to be contacted tion, the prescriptive model must have a feedback more than twice a year (a common rule-type in system that collects feedback data for each type telecommunications companies’ campaign man- of action, which will additionally increase data agement). This in fact leads to predictive and volume by some orders of magnitude. There- finally to prescriptive analytics, as combinatorial fore, prescriptive analytics is very challenging decision problems based on expected behavioural even with scalable data infrastructures and the developments of customers are required (bey- talent/expertise to make sense of the feedback ond the calculation of current scores). Besides a data (e.g., sensitivity analysis, causal inference, customer’s score-value for a planned campaign, or risk models). knowledge of future campaigns are of import- That makes prior data reduction even more im- ance as well as on future developments of cus- portant and requires a focus on the pieces of tomers in order to predict their responses. In input data really relevant for decision-making addition, it has been shown by Goel and Gold- and optimisation. stein (2013), amongst others, that the structure Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Data-Driven Decisions in Service Engineering and Management 113

of the communication or social network and the mathematical model or algorithm to additional prediction of future behaviour of a customer’s data, which might be more fine-grained data, neighbors play important roles, which brings a more data in terms of a longer reliable planning decision maker to network models, multivariate horizon, or simply an additional attribute or di- forecasting models and collaborative prediction. mension under consideration. While there is a huge body of knowledge of Optimisation gain also differs from concepts such broadly used methods and sophisticated tools as sensitivity, robustness, or stability of a solu- exist to perform individual tasks such as classific- tion. With optimisation gain we address the dif- ation, time series prediction, or mathematical op- ferent and more general problem of quantifying, timisation, the integration of these tasks to derive if (and how much) the optimality or robustness efficient and robust overall solutions is still left of a solution would benefit for example from the to the expertise and preferences of individual de- consideration of a novel data feature in a partic- cision makers, typically based on trial-and-error ular planning or decision problem. Addressing procedures or rules of thumb. such questions is challenging as this typically requires the re-formulation of the mathematical For each task, different data reduction techniques program formulation for numerous input-data and feature-combination might be adequate, combinations and transformations. The intuition while the interplay of these tasks might lead to of optimisation gain is the quantification of the the fact that certain data considered as highly solution quality expected with different input relevant in one task might not or only slightly data for a particular type of optimisation prob- impact the overall solution (and vice versa). For lem analytically, without expensive and time- instance, it might turn out that the prediction of consuming (and potentially infeasible) trial-and- features relevant to compute current scores are error-procedures. The vision is a new generation too difficult to predict for future campaigns and of criteria by integrating data and model selec- the forecast cannot be considered as reliable. For- tion and configuration. mulating a stochastic optimisation model might reveal that the solution is highly sensitive to even Please notice that optimisation gain can become small planning errors or rather insensible to lar- negative as too many parameters can lead to an ger ones, which makes the predictability of a explosion of the search spaces and increased com- feature either less or more important. Hence, plexity, where optimal solutions are much harder each type of problem requires individual data to find. For instance, node-sets of branch & cut and model selection procedures if the goal is to solvers might increase dramatically, and the qual- make optimal decisions. ity of solutions that can be found in pre-defined periods of time might decline sharply with the This leads to a novel concept in prescriptive ana- number of features and constraints under consid- lytics that we will refer to as optimisation gain eration. Furthermore, models operating with too of data. Optimisation gain differs from inform- many data dimensions are more likely subject ation gain (or derivatives such as information to over-fitting as artefacts and collinear config- gain rations, GINI, etc.) or matrix approxima- urations of (stochastic) variables used as model- tion quality norms of a residual matrix. Those input worsen the quality of decision-making. metrics are aimed at quantifying the quality of From a business perspective, the marginal gain of a data prediction or approximation without con- considering more data might further decline as textual knowledge on how information is used in collecting and managing data comes at additional subsequent optimisation steps. costs for data scientists that need to analyse the By optimisation gain we mean the dependency data, as well as costs for monitoring, IT infra- of a solution (the solution quality) derived by a structures, storage, and licenses. Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 114 Thomas Setzer

We argue that the role of optimisation gain of the expected service demand and the correspond- data is a highly relevant concept in prescript- ing resource demand that needs to be supplied ive analytics, and key to reducing Big Data ef- in future points of time. Considering the case ficiently to a manageable and actionable set of of private clouds, with the potential of hosting features Also, INFORMS, the leading scientific services in virtual machines (VM) in a flexible and professional organisation for OR profession- manner, e.g., by co-hosting VMs temporarily on als, decided to stake its claim on the analytics the same physical server, sharing and multiplex- movement. The organisation recognised that the ing a servers capacity for resources such as CPU, trend toward data-driven and analytical decision- memory, or I/O. In such an environment, IT ser- making presents tremendous opportunities and vice managers try to minimise the number of challenges for OR professionals (Libertore and servers by assigning enterprise services in vir- Luo 2011). Since 2009, INFORMS organises an tual machines efficiently to physical servers, but own conference at the intersection of analytics at the same time provide sufficient computing re- and OR named Business Analytics and Operations sources at each point in time. It is worthwhile to Research, with a focus on how to apply data sci- notice that running servers are (independent of ence to ‘the art of’ business optimisation. It fea- their utilisation levels) the main energy drivers tures presentations on real-world applications in data centers, where energy costs already ac- of analytic solutions, presented by industry and count for 50% or even more of total operational university leaders. costs (Filani et al. 2008).

Optimisation gain can provide a means of signi- Without going into too much detail, the result- ficantly reduce the effort spent for monitoring, ing VM allocation problem can be reduced to a collecting and managing data, as ideally only stochastic multi-dimensional bin-packing prob- data is collected that is indeed supposed to im- lem, a well-known NP-hard problem. As it is the prove decisions. Unnecessary frequent measure- case with every bin-packing problem, the goal ments are also avoided as the collection of correl- is to fill-up the available spaces (resource capa- ated data that is (statistically) already captured cities) of bins (servers) as much as possible, and, by other variables. These ideas are closely re- hence, come out with fewer servers while not lated to visions such as smart measurement and exceeding the capacity of servers, as this would collaborative monitoring systems, but with an result in overload and SLA violations. additional focus on the impact on the business relevance of gathered data. We will further detail Theoretically, historical workload data would al- on this in Sect.4. low for accurate workload demand forecasting (for more than 80% of typical operational busi- 4 Feature-based Optimisation and ness services) and optimal allocation of enter- Model-Data-Integration prise applications to servers. In various experiments and studies with smaller VM sets it has As aforementioned, certain units in enterprises been shown that such approaches lead to a reduc- have specific tasks to perform, usually composed tion of required server by around 30% (Speitkamp by structured or at least semi-structures pro- and Bichler 2010). Unfortunately the volume of cesses. For instance, in IT service management, data and the large number of resulting capacity the role of capacity management is to ensure constraints in a mathematical problem formu- sufficient capacity to provide high-quality ser- lation renders this task impossible for any but vices to customers efficiently, i.e., at reasonable small instances and is of little use for IT service (low) costs to the business. In capacity manage- providers with server parks of hundreds or thou- ment, it is important to have a clear picture of sands of VMs to be consolidated. Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Data-Driven Decisions in Service Engineering and Management 115

Looking at the core of each packing problem, in The same authors argue that the overall approach particular at bin-packing problems, the challenge can also be applied to other large packing prob- is to find complementarity in objects to be packed lems. For instance, in Setzer (2013), the authors (in our case, the demand profiles of VMs for vari- show that high-dimensional knapsack problems ous resources over time) to achieve high average can also be intelligently reduced to smaller and server utilisation levels. It makes sense to co-host computationally tractable ones, as long as there is VMs with peak loads in the morning hours and a significant amount of shared variance amongst VM having their peak loads later during a day. the dimensions to be considered. Please notice Similarly it makes sense to combine a VM with that, according to recent studies, knapsack-prob- high CPU and low memory demand with one lems are amongst the top four problems to be having lower CPU but high memory demand. solved in enterprises, although managers often do not know that their particular problems could When we consider relevant features of workload be formulated as knapsack-problems. profiles for the packing problem as aforementioned, features describing the complementarit- Overall, we believe that there is a huge poten- ies between VM profiles could be of great value, tial for solving particular decision problem with besides features describing the absolute resource Big Data made small. However, to exploit these demand curves of VMs. potentials, problems must be formalised before integrated data reduction and optimisation mod- Setzer and Bichler (2012) use techniques based on els can be developed. singular value decomposition (SVD) to extract significant features from a matrix of the expec- Reconsidering the example of capacity manage- ted (fine-grained) demand vectors of hundreds ment in private cloud infrastructures, we will of VMs and provide a new geometric interpret- now detail on the need for a decision model fab- ation of these features as principal demand pat- ric that not only aligns the model to be used terns, complementary between these patterns, to changing environments by considering novel and uncertainty. The extracted features allow parameters. In contrast, completely different for formulating a much smaller allocation model solution techniques are required depending on the (recent) structures and developments found based on integer programming and allocating in the data. Again, we will use private clouds for large sets of applications efficiently to physical illustration. servers. While SVD is typically applied for analytical purposes only such as time series decom- Nowadays, live migration allows to move VMs position, noise filtering, or clustering, here fea- to other servers reliably even during runtime tures are used to transform a high-dimensional and promises further efficiency gains (VMWare allocation problem in a low-dimensional integer ESX, amongst others) (Nelson et al. 2005). Some program with only the extracted features in a platforms such as VMware or vSphere closely much smaller constraint matrix. The approach monitor the server infrastructure in order to de- has been evaluated using workload data from a tect resource bottlenecks by tracking threshold- large IT service provider and results show that it violations. If such a bottleneck is detected they leads to high solution quality. At the same time take actions to dissolve it by migrating VMs to it allows for solving considerably larger problem different servers. For instance, if the CPU utilisa- instances than what would be possible without tion exceeds 80%, a VM is migrated away from prescriptive analytics, intelligent data reduction that server to reduce total server load. On the and model transform. This work provides a first other hand, if a controller detects phases of low example of a highly integrated data reduction overall workload, there is the possibility to con- and optimisation approach. centrate workloads on fewer servers by vacating Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 116 Thomas Setzer

servers and shutting down these source servers 6 Conclusion and Vision temporarily to further reduce energy consumption. We will refer to such techniques as dynamic Analysing historical and current data in order to resource allocation or dynamic control, as op- make better predictions is vital for running a composed to static VM allocation where allocations petitive service company. Data-driven design are computed and kept fixed for a longer period and management of services demand interdis- of time. ciplinary knowledge from the business domain, processes, data analytics, and mathematical op- 5 Towards Data-Elastic timisation. While in principle the ever-growing Decision-Making amounts of available data would allow for de- On the one hand, dynamic control strategies riving increasingly precise forecasts and optim- are more flexible and should therefore lead to ised input for planning and decision models, the lower energy costs. On the other hand, migra- complexity resulting from the consideration of tions cause significant additional overheads and large volumes of ever-growing volumes of mul- response-time peak, which are avoided with static tivariate, fine-grained data leads to the fact that allocation mechanisms. It has been shown that dependencies and relationships within the data with well-predictable workloads of business ap- are not found, algorithms do not scale, and tra- plications, dynamic resource allocation during ditional statistics as well as data-mining tech- operational business hours does not lead to higher niques collapse because of the well-known curse energy efficiency compared to static allocation of dimensionality. Hence, in order to make Big even if future demand is known only to a certain Data actionable, we are interested in the intelli- extend (Wolke et al. 2013). However, if demand gent reduction of vast amounts of data to small is completely unknown, dynamic control is the sets of problem-relevant features. We argue that only reasonable option to avoid both: massive mathematical optimisation and planning mod- overprovisioning and service degradation. De- els need to be transformed to be able to operate pending on the share of stochastic developments efficiently on highly reduced data. In addition, in workload demand curves, hybrid models might the selection of adequate planning and decision be appropriate where basic allocations are com- models must be adapted to (current) data and puted for a given planning horizon in a more the reliability of relations and predictions extrac- conservative fashion, considering the option of ted from that data, which requires time-dynamic potential migrations to cope with uncertainty. and data-driven model selection and evaluation In summary, dynamic, data-based model selec- techniques. tion is required that differs from parameter alignment, which simply would mean that for instance References the alpha parameter in an exponential smoothing model is adjusted from time to time (which then Accenture (2006) Service Management – En- leads to a different and hopefully better short abling High Performance Through Supply term prediction), but where the same mathemat- Chain Management. Accenture. ical model is used for prediction. Aggarwal C. C., Hinneburg A., Keim D. A. (2001) In the example above, depending on the predict- On the Surprising Behavior of Distance Met- ability of demand behaviour, which might be rics in High Dimensional Space. In: Database well predictable throughout certain periods but Theory – ICDT 2001. LNCS. Springer, London, rather unpredictable in other periods of time, pp. 420–434 completely different allocation mechanisms are Bellman R. E. (1961) Adaptive Control Processes: advised. A Guided Tour. Princeton University Press Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Data-Driven Decisions in Service Engineering and Management 117

Beyer K. S., Goldstein J., Ramakrishnan R., Shaft Roweis S., Saul L. (2000) Nonlinear Dimension- U. (1999) When is ’Nearest Neighbor’ Mean- ality Reduction by Locally Linear Embedding. ingful. In: Proc. Int. Conf. Database Theory In: Science 290(5500), pp. 2323–2326 (ICDT ’99). Springer, London, pp. 217–235 Setzer T. (2013) An Angle-Based Approach Choi D., Ahn B. S. (2009) Eliciting Customer to Solving High-Dimensional Knapsack- Preferences for Products From Navigation Be- Problems. KIT havior on the Web: A Multicriteria Decision Setzer T., Bichler M. (2012) Using Matrix Ap- Approach With Implicit Feedback. In: IEEE proximation for High-Dimensional Discrete Transactions on Systems, Man and Cyber- Optimization Problems. In: European Journal netics, Part A: Systems and Humans 39(4), of Operational Research 227, pp. 62–75 pp. 744–760 Speitkamp B., Bichler M. (2010) A Mathematical Davenport T. H., Harris J. G. (2006) Competing Programming Approach for Server Consolid- on Analytics: The New Science of Winning. ation Problems in Virtualized Data Centers. Harvard Business Review Press In: IEEE TSC 3(4), pp. 266–278 Davenport T. H., Patil D. J. (2012) Data Scientist: Stewart T., Thomas R., McMillan C. (2012) De- The Sexiest Job of the 21st Century. http: scriptive and Prescriptive Models for Judg- //hbr.org/2012/10/data-scientist-the-sexiest- ment and Decision Making: Implications for job-of-the-21st-century/ Knowledge Engineering. In: Expert Judgment Filani D., He J., Gao S., Rajappa M., Kumar A., and Expert Systems – NATO AS1 Senes 35, Shah P., Nagappan R. (2008) Dynamic Data pp. 314–318 Center Power Management: Trends, Issues, Tenenbaum J. B., Silva V. d., Langford J. C. (2000) and Solutions. In: Intel Technology Journal A Global Geometric Framework for Nonlin- Goel G., Goldstein D. G. (2013) Predicting In- ear Dimensionality Reduction. In: Science dividual Behavior with Social Networks. In: 290(5500), pp. 2319–2323 Marketing Science The Economist (Feb. 2010) A Special Report Lee J. A., Verleysen M. (2007) Nonlinear Dimen- on Managing Information: Data, Data Every- sionality Reduction. Springer, Heidelberg where. In: The Economist Libertore M., Luo W. (2011) INFORMS and the Tsymbal A., Pechenizkiy M., Baumgarten M., Analytics Movement: The View of the Mem- Patterson D. (2002) Eigenvector-Based Fea- bership. In: Journal Interfaces 41(6), pp. 578– ture Extraction for Classification. In: Proc. 589 Int. Artificial Intelligence Research Society Nelson M., Lim B.-H., Hutchins G. (2005) Fast Conference. Florida, USA Transparent Migration for Virtual Machines. Wolke A., Bichler M., Setzer T. (2013) Energy In: Proc. USENIX. Berkeley, USA, pp. 25–35 efficient capacity management in virtualized Nokia Siemens Networks (2013) Exit Big data centers: optimization-based planning vs. Data. Enter Right Data. http : / / www . real-time control. In: Proc. INFORMS Work- nokiasiemensnetworks.com/news- events/ shop on Information Systems Technologies. insight-newsletter/articles/exit-big-data- Milan, Italy enter-right-data Office of Government Commerce (OGC) (2009) Thomas Setzer Official ITIL Website. Last Access: [Online]. Available: http://www.best-management- Karlsruhe Institute of Technology practice.com/IT-Service-Management-ITIL Englerstraße 14 Ross J. W., Beath C. M., Quaadgras A. (2013) 76131 Karlsruhe You May Not Need Big Data After All. In: Germany Harvard Business Review 91(12) [email protected] Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 118 Imprint

Enterprise Modelling and Information Systems Architectures

The journal Enterprise Modelling and Information Systems Architectures is the official journal of the Special Interest Group on Modelling Business Information Systems within the German Informatics Society (GI-SIG MoBIS). The journal Enterprise Modelling and Information Systems Architectures is intended to provide a forum for those who prefer a design-oriented approach. As the official journal of the German Informatics Society (GI-SIG-MoBIS), it is dedicated to promote the study and application of languages and methods for enterprise modelling – bridging the gap between theoretical foundations and real world requirements. The journal is not only aimed at researchers and students in Information Systems and Computer Science, but also at information systems professionals in industry, commerce and public administration who are interested in innovative and inspiring concepts. The journal’s editorial board consists of scholars and practitioners who are renowned experts on various aspects of developing, analysing and deploying enterprise models. Besides Information Systems, they cover various fields of Computer Science.

Subscription Information The journal is distributed free of charge for members of the GI-SIG-MoBIS. Membership can be acquired through the German Informatics Society (http://www.gi-ev.de/verein/mitgliedschaft/). Single issues, priced at EUR 25 each (plus shipment), can be ordered online (http://www.fg-mobis.gi-ev.de/). Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 Editorial Board 119

Editorial Board Editors in Chief Manfred Reichert, Ulm University Klaus Turowski, Otto von Guericke University Magdeburg Associate Editors Wil van der Aalst, Eindhoven University of Technology Witold Abramowicz, Poznan University of Economics Colin Atkinson, University of Mannheim Jörg Becker, University of Münster Jörg Desel, University of Hagen Werner Esswein, Dresden University of Technology Fernand Feltz, Centre de Recherche Public Gabriel Lippmann Ulrich Frank, University of Duisburg-Essen Andreas Gadatsch, Bonn-Rhine-Sieg University of Applied Sciences Martin Glinz, University of Zurich Norbert Gronau, University of Potsdam Wilhelm Hasselbring, University of Kiel Brian Henderson-Sellers, University of Technology, Sydney Stefan Jablonski, University of Bayreuth Manfred Jeusfeld, Tilburg University Reinhard Jung, University of St. Gallen Dimitris Karagiannis, University of Vienna John Krogstie, University of Trondheim Thomas Kühne, Victoria University of Wellington Frank Leymann, University of Stuttgart Stephen W. Liddle, Brigham Young University Peter Loos, Johannes Gutenberg-University of Mainz Oscar Pastor López, Universidad Politècnica de València Heinrich C. Mayr, University of Klagenfurt Jan Mendling, Vienna University of Economics and Business Markus Nüttgens, University of Hamburg Andreas Oberweis, University of Karlsruhe Erich Ortner, Darmstadt University of Technology Erik Proper, Radboud University Nijmegen Michael Rebstock, University of Applied Sciences Darmstadt Stefanie Rinderle-Ma, University of Vienna Michael Rosemann, Queensland University of Technology Matti Rossi, Aalto University Elmar J. Sinz, University of Bamberg Friedrich Steimann, University of Hagen Stefan Strecker, University of Hagen Bernhard Thalheim, University of Kiel Oliver Thomas, University of Osnabrück Juha-Pekka Tolvanen, University of Jyväskylä Gottfried Vossen, University of Münster Mathias Weske, University of Potsdam Robert Winter, University of St. Gallen Heinz Züllighoven, University of Hamburg

ISSN 1866-3621 (Online) Enterprise Modelling and Information Systems Architectures Vol. 9, No. 1, June 2014 120 Guidelines for Authors

Guidelines for Authors

The journal serves to publish results of innovative research on all facets of creating and analysing enterprise models and information systems architectures. For research papers, it is required to satisfy academic standards in terms of originality, level of abstraction and justification of results. Experience reports serve to describe and analyse success stories as well as practical obstacles and resulting research challenges. Topics covered by the journal include, but are not restricted to the following subjects: • Languages and Methods for Enterprise Modelling • Reusable Domain Models (Reference Models) • Analysis and Design Patterns • Modelling of Business Processes and Workflows • Process-Oriented System Architectures • Component-Oriented System Architectures • Conceptual Modelling for Component-Oriented Design • Ontologies for Enterprise Modelling • Modelling for Enterprise Application Integration • Modelling for Data Warehouses • Modelling to support Knowledge Management • Model-Driven Development • Aspect-Oriented Design • Agile Methods for Enterprise Modelling Authors are asked for electronic submissions, which have to be sent to the editor in chief as e-mail attachment. In case of multiple authors, it is required to name one author who acts as contact person. The submission should include a cover page with the paper’s title and the names, affiliations and e-mail addresses of all authors. The first page of the paper starts with the title and does not carry the authors’ names. A manuscript must be in PDF format. It should not exceed 5.000 words – this includes an abstract of around 150 words. Submitted papers will be reviewed within no more than two months. The review process is double blind. Authors who submit a manuscript guarantee that it has not been published elsewhere, nor is intended to be published elsewhere. Papers that were accepted for publication must be written according to the style defined for the journal. A comprehensive description as well as a corresponding LaTeX template is provided on the web portal of the GI-SIG-MobIS (http://www.fg-mobis.gi-ev.de/).

ISSN 1866-3621 (Online)