Product Complexity: a Definition and Impacts on Operations Mark A

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by University of Dayton

University of Dayton eCommons Department of Management Information Systems, MIS/OM/DS Faculty Publications Operations Management, and Decision Sciences

10-2007 Product Complexity: A Definition and Impacts on Operations Mark A. Jacobs University of Dayton, [email protected]

Follow this and additional works at: https://ecommons.udayton.edu/mis_fac_pub Part of the Business Administration, Management, and Operations Commons, Databases and Information Systems Commons, Management Information Systems Commons, Management Sciences and Quantitative Methods Commons, Operations and Supply Chain Management Commons, and the Other Computer Sciences Commons

eCommons Citation Jacobs, Mark A., "Product Complexity: A Definition and Impacts on Operations" (2007). MIS/OM/DS Faculty Publications. 94. https://ecommons.udayton.edu/mis_fac_pub/94

This Article is brought to you for free and open access by the Department of Management Information Systems, Operations Management, and Decision Sciences at eCommons. It has been accepted for inclusion in MIS/OM/DS Faculty Publications by an authorized administrator of eCommons. For more information, please contact [email protected], [email protected]. RESEARCH ISSUES

■ SHAWNEE VICKERY, Feature Editor, Eli Broad Graduate School of Management, Michigan State University

Product Complexity: A Definition and Impacts on Operations by Mark Jacobs, The Eli Broad Graduate School of Management, Michigan State University

s evidenced by recent articles in need to be explained (Ramdas, 2003). AThe Wall Street Journal (Lawton, However, this cannot happen until 2007) and Forbes (Patton, 2007), there complexity can be explained theoreti- is a growing emphasis on product de- cally. But, to build theory there must first sign resulting in products that are in- be a common understanding about the creasingly more differentiated and construct of interest (Wacker, 2004). aimed at more and more narrowly de- Only then can researchers operationa- fined market segments. The result is lize it and search for meaningful rela- product portfolios manifesting increas- tionships. In light of this, I develop a ing levels of complexity. While adding definition of complexity below. A sam- to the portfolio may enhance revenue, it pling of the operations management lit- appears to be at a high cost. erature is then presented within the In a recent survey, 57 percent of ex- context of the definition. Then, given the ecutives reported that the cost to man- definition, an example of how theory age customer orders, procure and can be applied is offered and proposi- inventory materials, and deliver prod- tions drawn therefrom. ucts to end users threatens to undermine operational efficiencies and to consume Definition profits (Hoole, 2006). Product complex- The study of product complexity has ity in business supply chains is the pri- been hampered by the lack of consen- mary driver of these costs (Bozarth, sus around a precise definition. My goal Warsing, Flynn, & Flynn, 2007). Case is to establish a basis for consensus be- Mark Jacobs research confirms that many companies ginning with a formal and robust defi- is a PhD candidate in are indeed struggling with product nition of the construct ‘complexity.’ To Michigan State Univer- complexity decisions (Closs, Jacobs, do so I investigated several different sity’s Supply Chain Man- Swink, & Webb, 2007), and marketing disciplines to gain a comprehensive agement Department initiatives appear to be a major culprit. majoring in operations understanding of how complexity has management with an em- Marketers constantly pushing for been conceptualized to date. These find- phasis on product develop- greater differentiation of their products ings are summarized in Table 1. For ment. His research interests added 1.7 new products for each prod- brevity, the elucidation of these findings involve advancing theory in managerially rel- uct retired (Hoole, 2006). Thus it ap- will be reserved to other publications evant ways in the areas of product complexity, pears that the challenges presented by organizational performance, and operations (Jacobs & Swink, 2007). strategy. His dissertation investigates theoreti- product complexity are pervasive and Inspection of Table 1 reveals har- cal relationships between product complexity and significant to organizations mony amongst the uses of the word demand and supply chain costs. Prior to begin- (ATKearney, 2004). complexity in the academic literature. ning an academic career, Mark consulted for Cap The difficulty for organizations These similarities include multiplicity, Gemini Ernst & Young, held positions at IBM arises because neither complexity nor and Cummins, and was chief operating officer at relatedness, and difficulty of compre- JL Analytical Services. Mark holds a BSIE from its impacts on performance are well hension. Therefore, I propose the fol- Cal Poly, SLO, and an MBA from the University of understood (Fisher & Ittner, 1999b). The lowing definition of complexity. Minnesota’s Carlson School of Management. mechanisms through which it affects [email protected] cost, quality, delivery, and flexibility

Decision Line, October 2007 6 Table 1: Findings on how complexity has been conceptualized to date.

Complexity is the state of possessing among its elements than the other. We and complementarity. Similarity in- a multiplicity of elements manifest- therefore define product complexity as cludes sharing characteristics such as ing relatedness. follows: part geometries or components, offering Complexity in a product is mani- Product complexity is a design state the same functionality, fulfilling the fested by both the multiplicity of, and resulting from the multiplicity of, and same strategic role in the portfolio as a relatedness among, elements contained relatedness among, product architec- prior product, or any other such indi- within the product portfolio or the prod- tural elements. cation of a like kind relationship. uct itself. Ceteris paribus, one product Multiplicity relates to an enumera- Interconnectedness relates to a connec- is considered more complex than another tion of items. However, as can be seen tion via an interface such as that identi- if it contains a greater multiplicity of in Figure 1, relatedness has three dimen- fied by Ulrich’s (1995) slot, bus, and elements or more inter-relationships sions; similarity, interconnectedness, sectional typology. The gist is that there

Decision Line, October 2007 7 beneficial as long as the unit cost of the component being standardized was not too high relative to alternative suitable components (multiplicity complexity). Others (Krishnan, Singh, & Tirupati, 1999; Sanderson & Uzumeri, 1995) ar- ticulate how the use of a common platform can be advantageous to cost-effectively pursuing additional market segments. One interpretation of this work is that there can be increasing returns to decreasing complexity, but that the benefit is bounded by component costs. Therefore the benefit to the relational dimension of complexity may Figure 1: Three dimensions of Relatedness. be concave. A significant body of work has emerged on the topic of modularity— is a mechanical connection or the pass- tional markets to offset economic or po- modularity representing an increase in ing of signals between two elements. litical risks, as well as offer broader lines reledness complexity. Modularity en- The interconnectedness of elements in the hope of increasing the chance of ables scale economies (Pine, Victor, & also includes logical interconnected- at least one becoming a runaway suc- Boynton, 1993), inventory reductions ness. For example, a product that sup- cess. There are further forces such as (Fisher, Ramdas, & Ulrich, 1999a; plants another in the portfolio, the competitive positioning and responses Ramdas & Randall, 2004; Swink & proverbial new and improved product, that work to cause firms to offer more Closs, 2006; Tu, Vonderembse, Ragu- is connected to the old though the simi- products. Nathan, & Ragu-Nathan, 2004), engi- larity of position in the portfolio, func- The impacts of product complexity neering efficiencies (Collier, 1981), and tionality offered, market segment on firm operations are explored prima- improved coordination (Nobeoka & targeted, or other logical connection. rily in three separate research streams: Cusumano, 1997; Sanchez & Mahoney, Complementary relatedness is intended complexity management, measures, 1996; Schilling, 2000). However, the in the economic sense; an mp3 player and inventory. Inventory is the thread benefits are shown analytically to be a and digital music are complements. which ties the streams together as much function of the cost of the components of the management literature looks at being standardized (Fisher et al., 1999b; The Literature effectiveness in reducing inventory lev- Karmarkar & Kubat, 1987). Empirical As presented, product complexity rep- els or costs, and the measures are also research shows that the advantages of resents a multiplicity of related ele- focused on improving inventory posi- modularity can have a positive impact ments. Systems theory (Boulding, 1956; tions. However, even though elements on elements of competitive performance Simon, 1962) informs us that product related to portfolio complexity have (Jacobs, Droge, Vickery, & Calantone, complexity can be represented on sev- been studied since the 1970’s, there has 2006; Jacobs, Vickery, & Droge, 2007). eral levels. My review of the literature yet to emerge a unified framework. Plac- However, remaining is the need to de- finds that these levels include the port- ing the collective work of these schol- scribe the nature of the functional rela- folio of a firm’s offerings and the prod- ars into a new context, it becomes tionships between the dimensions of uct family, and extend down to the apparent that relational complexity has complexity and competitive performance. component level of the products within different outcomes than multiplicity A logical area for OM researchers the portfolio. complexity. This becomes most evident to explore in relation to product com- My view is that the genesis of prod- in the treatments of platforms and modu- plexity is inventory. Indeed this is where uct complexity resides at the portfolio larity within the literature. the primary focus of the operations level. The twin objectives of funding re- The relationship between related- management literature has been. Most quirements (generating large amounts ness and multiplicity complexity was of this research builds on that of Collier of cash currently and long term sales tacitly addressed by Krishnan and (1981) by looking at the impact of vari- growth potential) and risk mitigation Gupta (2001) who found that the ben- ety upon inventory. The first of these (Henderson, 1970, 1972a, 1972b) are efit to increasing the use of common was the seminal work of Collier (1982) powerful forces driving added levels of platforms (relational complexity) was who demonstrated that as the magni- complexity. Firms are pressured to in- a function of the component costs. They tude of the Degree of Commonality In- troduce product variants into addi- found that increasing platform use was dex (DCI) increased, the safety stock

Decision Line, October 2007 8 required decreased. Similar works, for ity Index (BCCI), Total Constant Com- (Meade, 1962). Several economists built example, Baker, Magazine and Nuttle monality Index (TCCI), and Within upon this foundation to establish that (1986), Gerchak, Magazine and Gamble Product Constant Commonality Index there is a diminishing return to invest- (1988), McClain, Maxwell, Muckstadt, (WCCI). Focusing at the component ment, and that substitution of resources Thomas, Weiss, and Collier (1984), fol- level, these measures account for the could positively impact productivity lowed shortly afterward, presenting degree of complexity across products, (Keynes, 1936; Leontif, 1941; Pareto, similar findings. Later, Fisher and Ittner the degree of standardization, and how 1906; von Bohm-Bawerk, 1889). Thus (1999b) explored this topic and found much complexity is present within a there is a limit to the performance an through simulation that the reduction product respectively. These measures organization can achieve given a cho- is attributable to risk pooling. Others did prove to be of value in modeling and sen set of assets. Schmenner and Swink furthered this stream and clarified the forecasting the inventory effects from chang- (1998) refer to this limit as the “asset relationship when they found that pro- ing the level of component complexity. frontier.” An organization may move its duction volume is a significant driver There is a second class of measures level of performance closer to the asset of the benefit to sharing components that has appeared recently in the litera- frontier by revising its policies and pro- (Fisher et al., 1999a). Gerchak and his ture that focuses on the interactions be- cedures in ways that more fully utilize colleagues have explored impacts of tween components or modules. its assets. The resulting increased effec- standardizing components on service Researchers (Browning, 2001; Eppinger, tiveness should be reflected by gains in levels (Gerchak et al., 1988), finding that 2001; Yassine & Braha, 2001) have em- productivity and financial performance standardization improves them. ployed the product structure matrix to (Clark, 1996; Hayes et al., 1996; Schmenner Swaminathan and Tayur (1998) chose visually represent interconnections. A et al., 1998; Skinner, 1996). to look at the problem at a different level. calculation of the percentage of connec- Transaction Cost Economics is gen- They indirectly address product com- tions (Mac Cormack & Rusnak, 2006) erally used to explain the structure of plexity in their modeling of the benefits yields the degree of component complex- organizations and why certain busi- of delayed differentiation. They find that ity. Another technique uses a ratio of ness transactions are chosen over oth- the use of vanilla boxes can reduce the connections within modules to those be- ers. TCE assumes that firms will act to cost of supplying variants of comput- tween modules to ascertain the degree of minimize costs, including both out of ers relative to the make to stock model. product complexity (Gershenson, pocket expenses and costs associated There is a second well-developed Prasad, & Allamneni, 1999). Most re- with risk. The three risks that TCE iden- stream of literature: measures of com- cently, Fixson (2005) suggests that com- tifies are asset specificity, environment, plexity. These measures are predomi- plexity can be operationalized by and opportunism. Putting TCE into the nately used to identify opportunities to creating a two dimensional space with context of the product architectural optimize inventory. Note that these ‘number of components’ as one axis and complexity, interconnections within the measures have been presented in the ‘number of functions provided by the product architecture represent transac- context of commonality. However, com- component’ as the other. The result is tions, and related costs include direct monality is just a reduced state of com- the number of components per function. production costs, as well as costs asso- plexity and hence should be viewed as ciated with the risks of asset specificity one end of the complexity spectrum. A Theoretical Perspective on and the environment. Opportunism Therefore, these measures assess de- Product Complexity would not be applicable, as the compo- grees of complexity. There are two theoretical perspectives nents are not independent actors pos- The first to apply a measure of com- that offer insights into the effects that sessing the capacity to rationalize their ponent complexity (multiplicity) was product complexity will have on opera- actions. The implication of TCE in this Roque (1977), who identified the aver- tions. These two theories are the Theory context is that a rational actor (the de- age number of applications per compo- of Performance Frontiers (TPF) (Clark, sign engineer) will seek to minimize the nent as a measure of standardization. 1996; Hayes & Pisano, 1996; Schmenner total number and concentration of His suggestion was that resource sav- & Swink, 1998; Skinner, 1996) and transactions, the cost of components, ings would be realized through an in- Transaction Cost Economics (TCE) and the influence of the environment. crease in standardization. However, it (Coase, 1937; Williamson, 1981, 1991, Using TPF and TCE as theoretical was Collier’s (1981) degree of common- 1996, 2002). frameworks, propositions can be con- ality index (DCI) that proved to be the The Theory of Performance Fron- structed that, when tested, will advance measure that other scholars built on. tiers has its basis in the neoclassical the theoretical understanding of the Wacker and Treleven (1986) built school of economics, which holds that impacts of product complexity on op- upon the DCI by creating indices that economic growth arises from techno- erations. One example for each dimen- captured the degree of complexity logical progress, and output can be rep- sion of complexity follows. across various dimensions, for example, resented by a production function Between Product Constant Commonal- See RESEARCH, page 21

Decision Line, October 2007 9 sity. Practically all of these waves are and implementation possibilities for Distribution & Logistics Management, driven by, enabled by, or amplified by EC+CDM+SC structures and their con- 32(10), 851-871. the advances of electronic commerce nections to competitiveness. Here we Hartono, E., & Holsapple, C. W. (2004). over the past fifteen years. Successful have endeavored to furnish some ideas Theoretical foundations for collabo- competitors are those who find ways to and structure that may offer guidance rative commerce research and prac- surf at least one of these six kinds of in taking on this task. tice. Information Systems and e-Business waves without being inundated by the Management, 2(1), 1-30. others. Moreover, they find ways to cope References Holsapple, C. W., & Singh, M. (2000). with environmental storms that arise, Arino, A., & Reuer, J. J. (2004). Design- Electronic commerce: From a defini- sometimes quite unexpectedly, so as not ing and renegotiating strategic alli- tional taxonomy toward a knowledge- to be knocked off course, sink into an ance contracts. Academy of Management management view. Journal of abyss of mediocrity, or worse. It is in Executive, 18(3), 37-48. Organizational Computing and Elec- this turbulent environment that deci- tronic Commerce, 10(3), 149–170. sions must be made. Collaboration in Caridi, M., Cigolini, R., &De Marco, D. the making of these decisions poten- (2005). Improving supply chain col- Jiao, J., You, X., & Kumar, A. (2006). An tially gives a wider base (of knowledge), laboration by linking intelligent agent-based framework for collabora- a more expansive span (of attention), agents to CPFR. International Journal tive negotiation in the global manu- and a greater flexibility (of processing) of Production Research, 43(20), 4191- facturing supply chain network. Robotics for dealing with the turbulent environ- 4218. and Computer-Integrated Manufactur- ment in PAIR directions. Chi, L., Hartono, E., Holsapple, C. W., & ing, 22(3), 239-255. The SoC ideas portrayed in Figure Li, X. (2007). Organizational decision Markus, M. L., & Christiaanse, E. (2003). 5 provide a frame of reference for future support systems: Parameters and Adoption and impact of collaboration consideration and study of the CDM benefits. In Handbook of Decision Sup- electronic marketplaces. Information nexus linking EC and SC. By their very port Systems: Basic Themes (eds. F. Systems and e-Business Management, nature, EC+CDM+SC structures are Burstein & C. Holsapple). Berlin/ 1(2), 139-155. necessarily concerned with knowledge, Heidelberg: Springer-Verlag. Massey, A. (2007). Collaborative tech- networks, and processes. In the interest Golicic, S. L., Davis, D. F., McCarthy, & nologies. In Handbook of Decision Sup- of helping organizations survive and Mentzer, J. T. (2002). The impact of e- port Systems: Basic Themes (eds. F. even excel in the competitive environ- commerce on supply chain relation- Burstein & C. Holsapple). Berlin/ ment, the decision sciences community ships. International Journal of Physical Heidelberg: Springer-Verlag. ■ needs to more fully elucidate the design

RESEARCH, from page 9

As the number of products offered units. This results in greater difficulty then become the guide by which the or components required to manufacture diagnosing, isolating, and repairing topic is explored, the ultimate result a product increases, the effort dedicated product failures (Karmarkar et al., being further development of TPF and to ensuring conformance will increase. 1987). Therefore, while the frequency of TCE, theoretical understanding of prod- The cost increases because with in- product failures may not be affected by uct complexity, and the opportunity to creasing numbers of items to sample, the interconnectedness of product ele- use the improved understanding to im- the number of samples must increase if ments, the cost to re-work failed prod- prove practice. ■ a constant detection rate is to be main- ucts will increase. Similarly, if an tained (Grant & Leavenworth, 1980; assembly is used across several prod- References available in the pdf version Kapur & Lamberson, 1977). Further, ucts in the portfolio, its failure will have of the article on the October 2007 Deci- these costs will grow at a decreasing larger ramifications than had it been sion Line Web site or upon request from rate due to better utilization of the qual- used in a single product. This leads to the author. ity function’s infrastructure. Therefore, Proposition 2: Proposition 1: P2: As interconnectedness increases, P1: As multiplicity increases, the cost warranty costs will increase. of inspection for conformance quality In conclusion, by formalizing the will increase at a decreasing rate. definition of complexity and clearly Greater interconnectedness in the specifying the underlying dimensions, product architecture creates greater in- appropriate theoretical perspectives terdependence among functional sub- can be identified. These perspectives

Decision Line, October 2007 21