Data Governance Special Report

November 2015 waterstechnology.com/ird Data Governance Special Report

Sponsored by Harness the Power of Your Data

Managed effectively, investment data is your ﬁrm’s most valuable asset.

Today, many of the world’s most influential investment organizations are relying on Eagle’s Data Management solutions to deliver a “single version of the truth”: high-quality, governed data that maximizes the transparency and accuracy of their investment decisions. Your business demands a stable and flexible platform that enables you to consolidate accurate information from multiple global locations and business lines. From a single platform, Eagle can support data management, investment accounting, and performance measurement requirements to help you more efficiently manage your investment portfolios.

Enterprise Data Management | Investment Accounting | Performance Measurement and Attribution

Contact us today to learn how Eagle can help your business: Americas +1 781 943 2200 Asia Pacific +65 6432 0581 Europe +44 20 7163 3827 Mainland China +86 10 58771605 www.eagleinvsys.com Eagle Investment Systems LLC, a leading provider of financial services technology, is a division within BNY Mellon Asset Servicing and is a wholly owned subsidiary of The Bank of New York Mellon Corporation. ©2015 Eagle Investment Systems LLC. All rights reserved. Editor’s Letter The EDM/MDM Challenge Asked whether enterprise data management (EDM) as a method of data governance has weaknesses, AIM Software’s Olivier Kenji Mathurin says EDM’s problem is a lack of transparency (in the Virtual Roundtable begin- ning on page 8 of this report). A possible reason for this, Mathurin says, is that retracing the technical implementation of data policies can be necessary to uncover implementation information and make the data governance transparent. While EDM can ensure consistency, it does not address whether processes should be governed centrally or in more distributed or federated fashion, says Mathurin’s counterpart, Paul McInnis of Eagle Investment Systems. Master data management (MDM) uses aggregation of sources to ensure a consistent view, McInnis adds. Yet, MDM and EDM can possibly complement each other, according to Acadian Asset Management’s Brian Buzzelli. These views raise the question of whether both MDM and EDM should be used, and used together, to get data governance benefits that would otherwise be impossible to achieve—especially transparency. MDM should be viewed as a component within EDM, according to Scotiabank’s Paul Childerhose, interviewed in a Q&A on page 14. Setting up MDM and EDM this way ensures improvement of data governance and data quality, he says. These data governance professionals all say progress has been made in the field, but it’s apparent that with more attention paid to data governance and more data governance work being done, the issue of figuring out MDM and EDM plans is taking center stage.

Yours sincerely,

Michael Shashoua Editor, Inside Reference Data Email: [email protected] Tel: +1 646 490 3969

waterstechnology.com/ird November 2015 3 Complete access for your entire organisation to WatersTechnology

Our information solutions provide everyone in your organisation with access to the best information resource on the ﬁ nancial market technology industry.

WatersTechnology delivers:

• Breaking news and in-depth analysis on the ﬁ nancial market technology industry • Detailed features on your market and the deals and people who shape it • Video features and technical papers • A fully-searchable archive of content from WatersTechnology and from all of the other market- leading titles incorporated under the site (Buy-Side Technology, Sell-Side Technology, Inside Market Data, Inside Reference Data and Waters) • Full compatibility with any mobile device

To ﬁ nd out more about the beneﬁ ts an information solutions package would bring, email [email protected] or call +44 (0) 20 7484 9933 / +1 646 736 1850

ZDWHUVWHFKQRORJ\FRP Contents

FEATURES NEWS 8 Virtual Roundtable 6 Data Governance Plans Get Cross- Inside Reference Data asks Functional Emphasis leading data management professionals how techniques 6 Data Governance and Quality such as EDM and MDM can help Concerns Rise, Survey Finds make sense of our increasingly complex information 7 Datactics Integrates Data Quality Manager With Governance 14 Q&A Solutions Front ofﬁ ce involvement in data governance begins with 7 News Download value for customers, says Paul Childerhose, Scotiabank’s director of data governance – exposure and capital analysis

November 2015 waterstechnology.com/ird Data Governance Special Report

Sponsored by

Incisive Media © 2015 Incisive Risk Information (IP) Ltd. Unauthorized 28–29 Haymarket, photocopying or facsimile distribution of this copyrighted newsletter is prohibited. All rights reserved. ISSN: 1750-8517 London SW1Y 4RX

waterstechnology.com/ird November 2015 5 News Review Data Governance Plans Get Cross-Functional Emphasis As firms look to put together data Harbottle, a consultant who previously governance plans—a task many are was a data governance implementa- just starting to carry out—handling tion lead at Canadian Western Bank. cross-functional pieces is emerging as “This ends up moving out into informa- an important part of constructing any tion—it is a cross-functional effort to governance plan. gather data and turn it into information,” “You need to understand all the other she said. components being put into position, in Data governance planning should be a strategy for how you will manage data a business function, not a technology across the organization,” said Cal Rosen, or vendor function, according to Paul vice president of enterprise data gover- Childerhose, director, data governance and data quality at TD Bank in nance—exposure and capital analysis, Toronto, during the Toronto Financial at Scotiabank. Rather than “ticking Information Summit. boxes” of technology and solutions The lifecycle of data starts at origi- providers, or vendors, start by “planning nation, continues with gathering and the journey,” he said. goes on to storage, according to April Michael Shashoua Data Governance and Quality Concerns Rise, Survey Finds Data quality is a top priority for the “This confirmed what we have been buy side and is rapidly gaining in hearing in forums we held last year, importance, a new survey has found. and in industry research we’ve been In May, Rimes Technologies released reading: that firms realize data quality its third annual survey of market isn’t necessarily the same thing for the participants such as major asset manag- end-users as it is for the operations ers, pension funds, banks, hedge funds teams working under more centralized and insurance companies. models of data governance,” says Giles Sixty-seven percent of respon- Arbuthnott, Benchmark Data Service dents—twenty-four percent more than manager at Rimes. “End-users are in 2014—cited data quality improve- demanding better quality data that is fit ment as the primary data management for purpose.” priority for this year. Joanna Wright

6 November 2015 waterstechnology.com/ird News Download

Asset Control Launches Risk Datactics Integrates Data Data Management Service Data management software Quality Manager With provider Asset Control has launched AC Risk Data Manager, Governance Solutions which provides data governance for risk managers on both the buy Software provider Datactics has launched and sell sides, and handles current a browser-based application called Data and historical risk data. Quality Manager (DQM) for matching “To date, the complexity of risk and integrating entity, instrument and data has meant each specific piece regulatory data. of data has sat separately in spread- DQM automates the transfer of critical sheets or in individual trading or data in the business life cycle and allows risk platforms,” says Richard Petti, easier access to the data-processing engines CEO of Asset Control. The new of Datactics’ FlowDesigner application. software synchronizes risk data “The design goal for DQM was to allow generation using common under- us to integrate easily with clients’ existing lying master data. master data management and data governance solutions,” explains Steve Cowler, Morgan Stanley Plans Data Datactics’ London-based CTO. Governance Improvements Clients can choose to set data-quality Morgan Stanley is evaluating thresholds, so records falling below defined providers of data governance levels of accuracy are presented for manual systems and close to a deal to use validation via the DQM Master Record a data quality tool, according to an Manager. They can then review and update executive at the firm individual records, as well as viewing and Morgan Stanley is working confirming the business rules that have been on improvements in data gover- applied to data. nance and data quality, says Mark According to Datactics CEO Stuart Harvey, Reitman, executive director of DQM allows capital markets firms to profile enterprise data analysis and infor- and scrub data as it enters the organization, mation lifecycle management at rather than undertaking this data manage- the firm. “We have a lot of systems ment task retrospectively. and a lot of data to break down,” Joanna Wright he says.

waterstechnology.com/ird November 2015 7 Virtual Roundtable Data Governance: towards a single version of the truth Inside Reference Data asks leading data management professionals how techniques such as EDM and MDM can help make sense of our increasingly complex information

How is data governance catching on the benefits of operational and cost as a data management method, or not efficiency as a corporate ‘data utility.’ —and why? As we continue to evolve the data Brian Buzzelli, head of data governance, management discipline, we recognize Acadian Asset Management: The finan- the need for a data governance frame- cial industry has been managing ever- work that works in partnership with increasingly complex classes of data from data management. Data governance has the time of the initial rise of markets and recently evolved to more formally trading. We initially employed technology provide the policies and standards, data to establish the architecture and infra- quality measurements and metrics, structure, and our data management clarity on roles and responsibilities, and practices evolved from database manage- changes to how we think about data, so ment for applications within a function to we understand and maximize the value enterprise data management (EDM) and of our information assets. comprehensive data management across We evolved to be a ‘pure information’ global information architectures. industry and few touch the physical assets The data management discipline these days. We look to operational and further evolved with master data data governance excellence in other management (MDM) techniques, where industries such as pharmaceuticals, tech- we focus on ‘mastering’ classes of data— nology, and aerospace; our industry’s data at times with more centrally controlled governance discipline is far less mature. organizational structures—to promote Data governance is not just catching on,

8 November 2015 waterstechnology.com/ird but rather it’s part of our industry’s natural evolution to a more structured, engineered, standardized and quantiﬁed approach to governing data management.

Olivier Kenji Mathurin, product market- Olivier Kenji Mathurin, ing manager at AIM Software: Much Head of Product Marketing progress has happened in the past couple & Research, AIM Software years with the emergence of the chief data officer (CDO) role. Where data +43 1 512 4652 governance was often perceived as [email protected] optional, regulators and data vendors’ requirements are putting the spotlight and is almost universally recognized by on it. It’s not just reporting the data investment management firms, although anymore, it’s also about the compliance formal adoption of data governance strat- of the process to get the data. egies is only now starting to gain momen- But to work, data governance requires tum. There are a number of drivers two elements: management commit- behind this growing interest, with orga- ment and recognizing data management nizations looking to gain material compet- as a business priority. This is where itive advantages and relying on these most firms still struggle. Firms that efforts to achieve operational efficiencies. have gone through an EDM program However, there is a gap between inten- often have an advantage because these tion and execution when it comes to data programs are cross-functional and often governance, with many firms struggling lay the foundation for data governance: to get these initiatives off the ground. definition and implementation of data A recent survey we conducted with policies, data models and glossaries, WatersTechnology highlighted three ownership and stewardship functions, main obstacles: lack of senior manage- standards, processes and tools. ment involvement; lack of resources; and not knowing where to start on drafting a Paul McInnis, data management product policy. Drivers for broader adoption of manager, Eagle Investment Systems: data governance policies by investment Data governance is a crucial part of a management firms include the need to successful data management framework. efficiently and consistently handle the In fact, it is valued as a strategic asset increasing volume and pace of informa-

waterstechnology.com/ird November 2015 9 Virtual Roundtable

it promotes deep data class expertise by organizing data practitioners around the full shape, life cycle and workﬂow of a class of data (e.g., entity, security master, Paul McInnis, data pricing, books, risk, transactions, etc.) management product Both MDM and EDM can beneﬁt from, manager, Eagle Investment and work in conjunction with, data Systems governance. Practitioners need to evolve +1 781 943 2526 and mature their data management discipline and develop data governance eagleinvsys.com in our industry similar to that of others.

tion. Firms that have ignored the need Mathurin: Both EDM and MDM have to establish such policies have experi- been strong in supporting data steward- enced operational inefficiencies, client ship functions by automating the recog- losses and data security breaches. nition of irregularities and exceptions and by providing a central issue resolu- What are the strengths/weaknesses tion workflow. They offer workflow of MDM and EDM for supporting support, control, and traceability to data governance? processes previously reliant on spread- Buzzelli: EDM focused on managing data sheets and manual-intensive activities. across the enterprise with greater func- Where EDM had weaknesses was in tional standardization in operational data the transparency of the data policies processing. MDM complements EDM as a executed. It was often hard to understand technology architecture strategy, an orga- which version of the data policy was appli- nizational efficiency enabler and as a cable on a given date or hour, because ‘center of excellence’ data management they were often translated from docu- technique. MDM informs our strategy and ments into technical rules programmed drives toward ‘single version of the truth’ into the database and other technical data structures to link the enterprise to layers of the EDM system. Understanding critical data stores (e.g., clients, accounts if a tolerance was changed in a quality and holdings); it shapes organizational control often involved gathering multiple structures, supporting greater centraliza- technical log files. tion of common data processing activities Asset-servicing firms using a model of for efficiency gains and reduced cost; and working with external data owners, are

10 November 2015 waterstechnology.com/ird pioneers in EDM. AIM has worked with a resulting “single version of the truth” is a fund administrator that calculates net logical component to the broader objec- asset values for 2,400 funds by setting a tive of aligning the business through a pricing policy for each asset class of the data governance strategy. fund. This policy defines how prices are That said, both MDM and EDM have sourced, selected and controlled for NAV their limitations as far as their support for calculation. The administrator embeds data governance is concerned. For these data policies in service-level agree- example, potential challenges with MDM ments with clients. Changes clients are the assignment of ownership of the request must be applied and tracked. To mastered data, who provides guidance on respond to external auditors asking about how decisions are made on data usage, prices selected on certain days, the firm and how conflicts are remediated. EDM needed a more flexible, controlled and ensures consistency and reduces the risk transparent way to manage changes in of conflict, but it does not necessarily data policy. resolve the question over whether Modern EDM must treat data policies as processes should be governed centrally first-class citizens. Business data owners or a more federal structure is optimal for must maintain a policy in their system an organization. without re-programming. The data policies can be modified in a single place, with How should IBOR (investment book approval workflows, version controlled, of record) serve as a data governance and with audit capabilities. The audit trail tool or support mechanism? must point to the exact version of the Buzzelli: Using IBOR to give immediate policy in place at any given time. and ‘real-time’ accurate positions has been historically challenging for reasons includ- McInnis: MDM and EDM help support ing data, technology and timing. The veloc- data governance initiatives by ensuring ity of global capital movement demands consistency of data across the organiza- we understand position, exposure, risk tion. MDM aggregates multiple sources of and obligation. This velocity, along with data in a central location to ensure there increasing regulatory demands, will drive is a consistent view across the enterprise, more evolution of data management, data while EDM enables the firm to have one governance, global data infrastructure and version of the truth across the enterprise. application-processing technologies. Having access across an enterprise to a single view of a dataset as well as the Mathurin: IBOR helps consolidate a firm’s

waterstechnology.com/ird November 2015 11 Virtual Roundtable

positions in one central investment data—is consumed by the place, acting as a organization. “Positions Master.” As such, an IBOR also Can data governance be effectively ensures the execution centralized or does it require a more of data policies—data dispersed, flexible approach? quality rules per asset Buzzelli: The management and culture of Brian Buzzelli, type, issue resolution every firm is unique; thus, the central- Acadian Asset workflows, roles, and ized or distributed structure should align Management ownership. A number of accordingly. However, data governance IBOR initiatives started is best positioned for success with a as a data management initiative, as these balanced approach where governance systems specialized in acquiring, control- authority, firm level policy, governance ling, and publishing investment data of all programs, standards and measurement, kinds—including cash and asset positions leadership and guidance are centralized. and/or corporate actions. The balance of data governance is within each business function and includes McInnis: IBOR is not a new concept—it’s definition and adoption of function-level something we’ve been helping clients data usage policies, specific business with for well over a decade. That said, it function service-level expectations and has grown in popularity as the regulatory authority over data stewardship and data environment and management practices ownership. Regardless of organizational increasingly demand risk and exposure structure, data governance should be a reporting at an enterprise level. shared development, adoption and An IBOR gives organizations one view promotion responsibility across the firm. of their investment data and enables them to benefit from a single source of Mathurin: The issue relates more often to truth regarding their investment activity. roles and processes rather than tools. This can be used by compliance, Centralized systems provide a single risk, and performance and attribution place to ensure the compliant execution systems across the organization to of the data policies. They also provide a improve controls and decision-making. single place for transparency, access, and As a result, an IBOR can be regarded connectivity to multiple sources. as a by-product of data governance Regarding the organization itself, struc- as it controls how data—in this case, ture must follow strategy: the process

12 November 2015 waterstechnology.com/ird and the organization can be decentral- and agile to react to, and reflect, both ized to allow ownership and steward- organizational and industry changes. We ship functions to run at the level where it firmly believe that data governance is makes most sense. We have seen various organic and ever-evolving, rather than a configurations from a central approach to set of rules carved in stone. Accordingly, a federated approach—where local subsid- success in data governance requires iaries own and manage local fields such as ongoing oversight and systematic reviews tax, while common fields are managed by to measure the completeness, accuracy, headquarter. These approaches are fully consistency and timeliness of the data. compatible with a central system. How would you compare MDM, EDM, McInnis: Data governance policies need IBOR and centralized data gover- to be flexible, period. However, whether nance as operational approaches? policies are centralized or federated Mathurin: All of these approaches are depends on the structure and size of the fully complementary, with EDM, IBOR organization, and they can vary, depend- and MDM focusing on the data steward- ing on the data in question. ship functions on different data focus A large organization may choose to angles, and the centralized data gover- centrally manage certain datasets that nance supplementing and orchestrating are commoditized and utilized across the these initiatives. enterprise, such as security reference data or pricing data. Yet, that same orga- McInnis: These approaches all follow the nization may choose to let other more concept of manufacturing centrally and esoteric datasets be governed locally. distributing globally. Firms need a unani- For example, if a quant group calcu- mous centralized view of their commod- lates data solely for its own usage and itized datasets, as well as how they are there are no dependencies on that data consumed and utilized across the enter- elsewhere, it makes sense for that group prise. Specifically for investment manage- to set its own controls and guidelines for ment organizations, there should be no that particular dataset. If at some stage, disagreement when it comes to the likes of that dataset does need to be consumed security reference, trade, position or pric- outside of the group, instituting a global ing data across the organization to ensure governance policy may then make sense. consistency and clarity of data, and the As is the case in this example, data decision making that results from its use. governance policies have to be flexible Each approach supports that core aim.

waterstechnology.com/ird November 2015 13 Q&A

Measurement As Blueprint Front ofﬁce involvement in data governance begins with value for customers, says Paul Childerhose, Scotiabank’s director of data governance – exposure and capital analysis Paul Childerhose

How do you go about making data ment strategy, then the performance of governance a front-office function? other key components of enterprise data The primary function of the front office management, like data governance and is to build relationships with prospective data quality, will be greatly improved. new customers, and to nurture and grow relationships with existing customers. To Are data governance plans produc- effectively get the business engaged in ing progress in terms of achieving data governance, begin by framing every more effective data management? conversation with the front office in the A partial answer will be found in the self- context of creating value for the custom- assessments against ‘compliance’ with er. If your mandate is data governance, BCBS239, which the global systemically talk to the front office to help them see important banks (G-Sibs) will complete that accurate and reliable customer as of December 31 and the Canadian analytics for sales and marketing depend domestic systemically important banks on timely, accurate and complete data. (D-Sibs) one year later. Firms that are employing industry developed tools, such What are the differences between as the DMMM and/or the DCAM, to assess master and enterprise data manage- their compliance against BCBS239 will be ment? Are these differences seman- well-positioned to objectively answer the tics or substantive? self-assessment stocking-taking question- The differences are substantive. Master naire for the regulators. Secondly, to have data management is the most chal- more effective data management, your lenging component of enterprise data data governance plan needs to always management. It is the blueprint for how focus on value creation for the customer. data should be organized and classified. If your customer base grows in size and If the data is organized and classified profitability, credit your data governance according to the master data manage- and data management.

14 November 2015 waterstechnology.com/ird 1

THE INSIDE REFERENCE DATA APP GREAT NEW FEATURES

DOWNLOAD OUR LATEST UPDATE TODAY TO: ŘGet the full picture on the move – text and images now automatically adjust to ﬁ t all screen sizes perfectly ŘAccess the app on your desktop – our new Web app allows you to view it on your computer as well as your smartphone and tablet

Find out more at waterstechnology.com/static/apps#ird Incisive Media 28–29 Haymarket, London SW1Y 4RX