Transcription
Proceedings of the 53rd Hawaii International Conference on System Sciences 2020Can Enterprise Architecture Be Based on the Business Strategy?Svyatoslav KotusevHSE Universitykotusev@kotusev.comSherah KurniaUniversity of Melbournesherahk@unimelb.edu.auAbstractEnterprise architecture (EA) is a set of documentsdescribing various aspects of an organization from anintegrated business and IT perspective. EA facilitatesinformation systems planning and helps improvebusiness and IT alignment. Traditionally, the conceptof EA was closely coupled with the business strategyand mainstream EA methodologies recommendstarting the EA effort from documenting the businessstrategy and then using it as the basis for defining therequired structure of information systems. Thisconceptual paper discusses in detail four practicalproblems with the business strategy that question itsvalue as the basis for EA initiatives. The presence ofthese problems challenges one of the most cherishedbeliefs or even axioms of the EA discipline: that EAshould be based on the business strategy. This paperraises a number of questions regarding the informationinputs necessary for the EA effort and calls for furtherresearch in respective directions.1. IntroductionNowadays the use of IT can be considered asessential for the business of many organizations.Companies invest a considerable amount of money innew IT systems and underlying infrastructure requiredto support their business operations. However,realizing the business value from these IT investmentsrequires aligning, if not intertwining, business and ITstrategies [1, 2, 3]. Enterprise architecture (EA) is acollection of organizational documents, typically calledas artifacts, describing various aspects of anorganization from an integrated business and ITperspective [4, 5]. EA intends to bridge thecommunication gap between business and ITstakeholders, facilitate information systems planningand thereby improve business and IT alignment [6, 7].In the existing literature the concept of EA is veryclosely related to the business strategy. For instance,mainstream EA methodologies [8, 9, 10, 11]recommend starting EA initiatives from documentingURI: 3(CC BY-NC-ND 4.0)Rod DilnuttPaul TaylorUniversity of MelbourneDHHS .authe organizational business strategy, strategic vision,goals and objectives and then using this informationfurther as the basis for defining the required structureof information systems. Some authors argue that EAshould be derived directly from the top-level businessstrategy [12] or even propose the definitions of EA thatexplicitly reflect an inextricable link existing betweenEA and the business strategy [13]. Other authors gofurther and claim that there may be no EA without thebusiness strategy [11, 14].However, most claims on the fundamentalimportance of the business strategy for EA are purelyprescriptive or highly speculative in nature and do notoriginate from evidence-based literature, while thebroad analysis of empirically substantiated literatureshows that the business strategy actually has a numberof undesirable properties rendering it incongruous asthe basis for EA initiatives. For example, in manyorganizations an articulate business strategy may besimply missing [15, 16], while in other organizationsthe business strategy may be extremely unstable andvolatile [17, 18].In this conceptual paper we answer the followingresearch question: “What problems with the businessstrategy may prevent its use as the basis for EA andhow can these problems be potentially addressed?”Specifically, we identify and discuss in detail fourdifferent practical problems with the business strategyfound in literature that question its value as thepotential basis for EA initiatives. This paper challengesthe status quo in the EA discipline, disputes the centralplace of the business strategy in the EA discourse,raises a number of questions regarding the necessaryinputs of the EA effort and calls for further research inrespective directions.Importantly, this paper does not intend to offerdefinite answers to the questions it raises, but rather tostimulate future research in order to clarify the actualrole of the business strategy for EA and understandwhat other information might be required as an inputfor EA initiatives.This paper continues as follows: (1) we discuss theperceived role of the business strategy for EA in themainstream EA literature, (2) we discuss four problemsPage 5613
with the business strategy identified in empiricalliterature that may prevent its use as the basis for EA,(3) we address the question whether EA can be actuallybased on the business strategy as prescribed inliterature, (4) we discuss possible solutions to theidentified problems, (5) we outline critical questionsand directions for future EA research related to thebusiness strategy, its value as the basis for EA andother possible inputs of the EA effort and (6) weconclude the paper and discuss the implications of ourfindings for research and practice.2. The role of the business strategy forenterprise architectureThe term “business strategy” has numerous slightlydifferent meanings and interpretations in literature[19]. However, it can be generally understood as “acombination of the ends (goals) for which the firm isstriving and the means (policies) by which it is seekingto get there” [20, p. xvi]1.Traditionally, the notion of business strategy playsa significant role in the EA discourse and the businessstrategy is widely considered as a starting point, orbasis, for developing EA artifacts defining the futurestructure of information systems required by theorganization. In fact, all mainstream EA methodologiespropose to start the development process of EAartifacts in some or the other form directly from theorganizational business strategy, e.g. mission, vision,drivers, goals, objectives and key performanceindicators (KPIs) [8, 9, 10, 11, 12, 21, 22, 23, 24, 25,26, 27, 28].For example, Holcman [25] recommends startingthe EA effort from explicitly documenting the businessgoals and their hierarchy. van't Wout et al. [11, p. 35]list the business vision, mission, strategy, drivers andobjectives as first EA artifacts of the contextual layer,which “sets the stake in the ground for the rest of thearchitecture by providing context”. Similarly, TOGAF[10] lists the business strategy, goals and driversamong the primary inputs of the preliminary phase ofits architecture development method (ADM). Bittlerand Kreizman [12, p. 4] claim that “future-state EA isdirectly derived from business strategy” and argue that“the goal [of the EA effort] is to translate businessstrategy into a set of prescriptive guidance to be usedby the organization (business and IT) in projects thatimplement change” [12, p. 7]. IBM’s EA consulting1The business strategy can also exist at different organizationallevels, e.g. corporate, divisional and departmental. In this paper wediscuss specifically the top-level corporate business strategy definedby C-level executives. Our analysis and conclusions may not beequally applicable to more detailed lower-level strategies definedwithin separate business unitsmethod states that EA is “driven by strategy” [27, p. 4].Likewise, Oracle’s EA framework declares that“driven by business strategy” is the first of its corevalues [26, p. 4]. Essentially, all these methodologiesconsider the business strategy as the core input for EAinitiatives.Analogous ideas regarding the primacy of thebusiness strategy are also expressed by other authors[29, 30, 31, 32], who argue that EA and IT planningefforts in organizations should stem directly from thebusiness strategy. Bernard [33, p. 12] states that “theidea of Enterprise Architecture is that of integratingstrategy, business, and technology”. Parker and Brooks[34, p. 46] argue that the business strategy and EA areinterrelated so closely that they actually represent “thechicken or the egg” dilemma. These views aresupported by Gartner as well, whose analysts evenexplicitly define EA as “the process of translatingbusiness vision and strategy into effective enterprisechange” [13, p. 2]. Moreover, Gartner analysts arguethat “the strategy analysis is the foundation of the EAeffort” and propose six best practices to align EA withthe business strategy [35, p. 1]. Unsurprisingly, similarviews are also shared by academic researchers, whoanalyze the integration between the business strategyand EA [36], modeling of the business strategy in theEA context [37, 38, 39] as well as other aspects of theirrelationship [40, 41].To summarize, in the existing EA literature thebusiness strategy is widely considered as the necessarybasis for EA and for many authors the very concepts ofbusiness strategy and EA are inextricably coupled, i.e.EA essentially cannot exist without the businessstrategy. Current views on the role of the businessstrategy for EA prevalent in literature can arguably bebest summarized in the words of Schekkerman [14, p.6], who formulates this idea in the most striking way:“No strategy, no enterprise architecture”. van't Wout etal. [11, p. 11] echo the same view almost verbatim:“No strategy, no architecture. No vision, noarchitecture”.3. Problems with the business strategy asthe basis for enterprise architectureDespite the widespread acknowledgement of theguiding role of the business strategy for EA initiativesin the current EA literature, as discussed above, anumber of important facts on the business strategyallow questioning its actual place in relation to EA.Interestingly, all the discussions of the relationshipbetween the business strategy and EA are highlyspeculative, while all the claims on the importance ofthe business strategy for the EA effort are purelyprescriptive. For instance, none of the publicationsPage 5614
cited in the previous section to highlight the centralrole of the business strategy in the EA discourse isbased on empirical research in real organizations. Allthese publications are either purely conceptual, or atbest based only on anecdotal evidence. At the sametime, a broad analysis of the empirically substantiatedliterature on business and IT alignment, informationsystems planning and EA reveals the existence of atleast four long-recognized major problems with thebusiness strategy, which suggest that it actually cannotbe considered as a sound basis for EA initiatives.3.1. Business strategy is often vague, unknownor merely absentFirstly, despite the prevalence of “no strategy, noarchitecture” thinking advocated, among others, bySchekkerman [14] and van't Wout et al. [11], manyorganizations actually face exactly this situation: theyhave no strategy, or at least no clear strategy 2. Over thelast decades numerous authors have consistentlynoticed that in many organizations the businessstrategy is very inarticulate, unknown to decisionmakers or simply absent altogether [15, 16, 42, 43, 44,45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59,60].The first observations regarding the absence of aformal business strategy and clear strategic plans inorganizations had been made in the mid-1980s byGalliers [54], Lederer and Mendelow [43] and Vitale etal. [47]. For example, Lederer and Mendelow [46, p.11] reported that “nine IS executives stated thatsometimes top business executives have no clearlydefined mission, objectives, and priorities, and do notknow their plans for the coming year”. Moreover,“some interviewees maintained that top [business]executives preferred flexibility which is lost when aplan is written” [56, p. 16]. Analogous conclusions hadbeen reported by Vitale et al. [47, p. 268]: “Many ISmanagers would feel very fortunate to have a clearpicture of where their organization is headed so thatthey could match IS and organization efforts. But manyorganizations have no well defined strategy”During the 1990s similar findings had beenreported by Baets [48], Bhide [59], Flynn and Hepburn[55], Sillince and Frost [60] and Segars and Grover[49]. For example, Baets [48, p. 206] reported that“preliminary research undertaken by the author in a2Since the business strategy is not defined within EA initiatives,its presence or absence is considered as a given immutable conditionfor architects to which they need to adapt. Moreover, the absence ofthe business strategy does not necessarily represents a problem fororganizations and in some cases may be even beneficial for business[19]. Generally, any discussions of the business strategy itself, itsvirtues and desirable qualities are out of the scope of this paperwell ranked European bank showed quite clearly thatmany of its middle managers, charged with carryingout the corporate strategy on behalf of the bank, wereunable to define the corporate strategy”. Likewise, thesurvey of 100 founders from the list of 500 fastestgrowing U.S. companies demonstrated that only 28%of them had formal business plans or strategies [59].Sillince and Frost [60, p. 111] found that in publicsector organizations clear strategies and goals might beabsent for political reasons: “[In police] the businessstrategies of change have been less clear-cut inpolitical terms; government has been ambivalent aboutthem. [.] So police goals are not at all clear”.During the 2000s similar observations had beenrepeated by Hackney et al. [58], Rosser [53], Slater[57], Campbell [42] and Chan and Reich [50]. Forexample, Slater [57, p. 85] reported that the survey byCutter Consortium found that “almost a third of therespondents had no formally articulated business planat all”. Campbell [42, p. 657] reported that “the results[of my study] indicate that the major concern ofpractitioners when considering alignment is copingwith the ambiguity surrounding the business strategiesthat are actually in use”. Chan and Reich [50, p. 299]noted that “a recurring issue seen in previousalignment research is that often corporate strategy isunknown [.] or, if known, is unclear and/or difficult toadapt”.More recently analogous observations have beenrepeated once again by Brown [15], Banaeianjahromiand Smolander [51] and Cantara et al. [52]. Forexample, Brown [15, p. 6] reports that “the espousedideal was that there should be a clear business strategyon which to base [information systems planning,while] the practical reality was that very often businessstrategy was either intangible, not clear, or deliberatelyambiguous for political reasons”. The Gartner surveyfound that “two-thirds of business leaders are unclearabout what their business strategy is and whatunderlying assumptions it is based on” [16, p. 2].Therefore, the lack of a clear business strategy inmany organizations that can be taken as the basis forfurther IT planning has been consistently reported byresearchers since the 1980s. Unsurprisingly, the surveyby Hauder et al. [61] shows that 84.8% of Europeanand U.S. organizations consider unclear business goalsas a significant challenge to their EA practices.3.2. Business strategy rarely provides a cleardirection for ITSecondly, even when organizations have a ratherarticulate formal business strategy, this strategy oftenis still unable to provide a clear direction for IT. Thisproblem with the business strategy has been alsoPage 5615
consistently noticed by many authors over the lastdecades [43, 44, 45, 46, 49, 56, 57, 62, 63, 64].Initially, Lederer and Mendelow [44, p. 393], whostudied difficulties in identifying business executives’objectives by IT planners, found that often “topmanagement fails to communicate corporate objectivesin a way to which IS personnel can relate”.Specifically, Lederer and Mendelow [56, p. 16]clarified that “in other cases [when formal plansexisted], the corporate plans were glittering generalitiesor mere financial targets which could not be translatedinto IT plans”. As Lederer and Mendelow [44, p. 393]reported, the business strategy often defines somepurely financial indicators useless for IT planningpurposes: “For example, top management told oneinterviewee that the organization’s major objective wasto increase sales by a given percentage and that ISshould provide systems to help do so. This providedlittle substantive direction as to what specific systemsto develop”. The same reasoning applies to marketshare and other similar goals as well: “For example, anobjective to “increase market share by a specifiedpercentage” does not define a computer application,leaving systems managers to draw their own,sometimes erroneous, conclusions” [45, p. 74]. Ledererand Mendelow [46, p. 11] also identified some morecomplex situations: “Finance Vice President stated thathis objective was to “maximize the financial flexibilityof the organization” but could not articulate how thisshould be done. This objective was too general topermit the [IT] director to formulate a supporting planfor [information systems]”.Later, Segars and Grover [49, p. 387] reported ananalogous story: “Many IS planners noted that thestrategic direction of the organization was notcommunicated in a manner which was understandable.In some instances strategic direction wascommunicated in terminology or documentation whichwas difficult to interpret”. Likewise, Slater [57, p. 86]noted that “business strategies are typically written at avery high level. They frequently talk about markets,sales and distribution channels, and growth targets, butrarely address how the company gets its work done”.As Ross et al. [64, p. 6] put it, “general statementsabout the importance of “leveraging synergies” or“getting close to the customer” are difficult [for IT] toimplement”.Therefore, the problem with formulating businessstrategies and plans in a way that does not provide anyclear actionable suggestions for IT has been recognizedby researchers for a long time. This problem alsoquestions the value of the business strategy as the basisfor EA initiatives.3.3. Business strategy is often unstable andfrequently changesThirdly, even when organizations have a ratherclear and actionable business strategy, this strategy isoften unstable, frequently changing and unable toprovide a steady basis for planning IT. This problemwith the business strategy is also consistently noticedby many authors over the last decades [17, 18, 44, 45,46, 47, 49, 62, 64, 65, 66].For example, Lederer and Mendelow [46, p. 11]noticed long ago that “even if top executives knowtheir plans in sufficient detail, an unstable environmentmight render them inapplicable”. Later, Lederer andMendelow [66] studied the problem of shiftingpriorities in more detail and identified the inherentinstability of the business strategy due to the ficklenessof the marketplace, changing customer needs andcorporate acquisitions as a major factor contributing tothis problem. As noted by one of the interviewed ITexecutives, “the winds change with each quarterlydirector’s meeting and we come back with a new set ofsignals” [66, p. 323].Segars and Grover [49] also identified theinstability of the business strategy as one of the riskfactors of architectural planning. For instance, anarchitect of a large U.S. financial organization vividlyillustrated this problem: “We did a thorough job ofaligning ourselves with organizational strategy. We feltconfident in our analysis and proceeded to operatewithin the enterprise models developed. However, wedid not do a good enough job of ensuring that thesemodels were maintained. It only took a period ofmonths before critical aspects of strategy and thebusiness had changed” [49, p. 388].Sauer and Willcocks [65, p. 41] reported that “most[surveyed CEOs and CIOs from 97 e-businesscompanies] were responding to an increasingly volatilebusiness environment by shrinking their developmentand planning cycles. Half don’t extend plans beyond ayear, and half of those with infrastructure plans updatethem quarterly”. Likewise, significant environmentaluncertainly and constant changes in the businessstrategy are also typical for companies in the retailindustry sector. For example, an architect from a majorretail company vividly illustrates this situation: “Theproblem with an organization like this is that in twelvemonths the organization has changed direction three orfour times. So, you’re not going to get that kind ofstability that fits those timeframes” [17, p. 34].The inability of the business strategy to offer astable foundation for EA initiatives is reported byreflective EA practitioners as well: “It is therefore afundamental flaw in many enterprise architectureapproaches that one starts from the (current) businessPage 5616
strategy and/or a set of principles that may be derivedfrom that strategy. Such a waterfall almost neverworks. [Although strategy should be taken intoaccount,] simply taking the current strategy and handthat to the architects to turn it into the starting point ofenterprise architecture will almost certainly fail,because the strategy is going to change long before theresults of enterprise architecture are visible” [18, pp.141-142].Therefore, the instability of the business strategy inmany organizations for the purposes of architecturalplanning has been also consistently reported byresearchers and practitioners since the 1980s.Unsurprisingly, the survey by Hauder et al. [61] showsthat 71.4% of European and U.S. organizationsconsider quickly changing organizational environmentas a significant challenge for their EA practices.3.4. Business strategy often requires strategyspecific, non-reusable IT systemsFinally, even when organizations have a ratherclear, actionable and stable business strategy, thisstrategy often requires highly specific, non-reusable ITsystems that cannot deliver lasting business valuebeyond the current strategy. This problem with thebusiness strategy is recognized less widely than thethree other problems discussed earlier, but is stillacknowledged by a number of authors [18, 62, 63, 64,67, 68, 69, 70].After being developed and deployed, informationsystems typically exist in organizations much longerthan the business strategies or strategic initiatives theywere intended to support [64, 69, 70]. Specifically,“average” business strategies may be active for theperiod of no longer than 3-5 years, while the ITsystems created to execute these strategies may stay inorganizations for 10-15 years or even longer [69, 70].For instance, Mocker [70] explains this mismatchmetaphorically by saying that IT exists in a different“time zone” from the business.For this reason, even a stable business strategy isunable to provide a long-lasting, sustainable view ofthe business commensurable with the lifespan of itsinformation systems and enable the proactive use of ITin the organization in the long run. As a result, “IT isleft to align with individual strategic initiatives – afterthey are announced. Thus, IT becomes a persistentbottleneck” [62, p. 1]. These attempts to chase everchanging business strategies (ever-changing in a sensethat even rather stable strategies active for the period of3-5 years can change faster and more radically thaninformation systems, which often stay active for thelonger period of 10-15 years) usually lead to theproliferation of legacy IT systems in organizations thatonce were strategic, but then lost their relevance to thebusiness [64, 69]. Thereby, today’s IT assets oftenbecome tomorrow’s IT burden.For example, Shpilberg et al. [67, p. 52] call suchsituations, when strategically aligned informationsystems eventually turn into an inefficient, entangledand fragile IT landscape, as “alignment traps” anddescribe one of these situations in the following way:“The company’s various divisions were drivingindependent initiatives, each one designed to addressits own competitive needs. IT’s effort to satisfy itsvarious (and sometimes conflicting) businessconstituencies created a set of Byzantine, overlappingsystems that might satisfy individual units for a whilebut did not advance the company’s business as awhole”. Similarly, Weill and Ross [63, p. 1] describethis situation in the following way: “IT organizationsattempt to build capabilities while addressing a laundrylist of immediate business needs. The result, in mostcases, is IT spaghetti – with ever increasingmaintenance costs and slow time to market”.The inability of the business strategy to offer along-term guidance (comparable to the typical 10-15years lifespan of information systems) regarding therequired structure of the organizational IT landscape isrecognized by reflective EA practitioners as well. Forinstance, Wierda [18, p. 134] argues that “what peopleseldom realize that if you build a landscape of elementsthat have an average life span of fifteen years with astrategy that changes direction every few years,chances are that you end up with a mess”. “Ironically,one of the most pregnant uncertainties is the strategy ofthe company itself. Systems have an average life timeof fifteen years. The strategy of a company [on averageremains constant for only] maybe four. In other words:in the time that the architecture of a system and a largepart of its surrounding systems exists, theorganization’s strategy will have changed four times,and often such changes are pretty radical” [18, pp. 140141].Therefore, the unsuitability of the business strategyas the foundation for a long-range architecturalplanning exceeding the horizon of 3-5 years is alsoacknowledged by both researchers and practitioners.The ensuing susceptibility to “alignment traps”questions the value of the business strategy as the basisfor EA initiatives.4. Can enterprise architecture be based onthe business strategy?The four problems with the business strategydiscussed above suggest that the business strategyeither cannot be taken as the basis for EA initiatives atall due to its absence, vagueness or volatility, or at bestPage 5617
can provide only some relatively short-term directionfor IT incommensurable with the typical lifespan ofinformation systems in organizations. A conceptualdecision-making framework for assessing thepossibility of using the business strategy as the basisfor EA reflecting the four common problemshighlighted in this study is shown schematically inFigure 1.The existence and widespread acknowledgement ofthe four problems with the business strategy shown inFigure 1 suggests that contrary to the popular claimsfound in popular prescriptive literature, the businessstrategy actually can hardly provide any real practicalbasis for EA, or at least the business strategy takenalone on its own. In light of these findings, the tenetthat EA should be derived directly from the businessstrategy or rooted in strategic drivers, goals andobjectives advocated by most EA methodologies [8, 9,10, 11, 12, 21, 22, 23, 24, 25, 26, 27, 28] can beconsidered more as an attractive cliché or seductivemotto, than as a realistic actionable prescription thatcan be successfully implemented in practice.Figure 1. Assessment of the business strategy as the basis for enterprise architecture5. Possible solutions to the identifiedproblemsAlthough the problems with the business strategydiscussed above (see Figure 1) have no definiteanswers in the available EA literature, someapproaches still seem promising as potential solutionsto these problems. These approaches can be groupedinto conceptual, organizational and technical ones.Conceptually, some other aspects of organizationsmight be taken as an input for EA initiatives. Forexample, Ross et al. [64, p. 25] recommend to use anoperation model (i.e. “the necessary level of businessprocess integration and standardization for deliveringgoods and services to customers”) as the basis forplanning IT. Unlike the business strategy, an operatingmodel should always exist in some or the other form,should be more clear, actionable for IT and stable inthe long run [62, 64]. However, this suggestion ishighly prescriptive, received only a limitedindependent validation [71, 72] and it is still largelyunclear whether, to what extent and how manyorganizations actually find the concept of operatingmodel helpful for their IT planning efforts.Organizationally, some problems with the businessstrategy might be resolved though establishingeffective IT governance arrangements and a closerdialog between business and IT helping intertwinebusiness and IT strategies together. For example, theIT governance literature stresses the importance ofcollaborative decision-making involving both businessand IT leaders with clearly defined responsibilities andPage 5618
decision rights [68, 73]. Similarly, the importance ofdirect communication and finding a common languagebetween business and IT stakeholders has been longrecognized in the business and IT alignment literature[46, 50, 74, 75]. However, the details of respectiveprocesses (e.g. what planning decisions get made, whomakes them and when as well as how exactly thebusiness strategy is converted into IT actions) stillremain insufficiently understood [76].Technically, the problem of the instability of thebusiness strategy in the short and long terms might bealleviated via adopting agile delivery approaches andflexible architectural paradigms (e.g. service-orientedarchitecture) respectively. In particular, agiletechniques may promote better adaptability to rapidlychanging business needs, while service-orientedarchitecture may facilitate higher reuse of theaccumulated IT assets in future business strategies.However, these approaches address only someproblems with the business strategy and may not offera “complete” solution.6. Directions for future researchThe four problems with the business strategyidentified and presented in this paper (see Figure 1)challenge the status quo in EA research and questionone of the most cherished beliefs, assumptions or evenaxioms of the EA discipline: that EA should be basedon the business strategy. As demonstrated in this paper,these beliefs are based essentia
strategy for EA prevalent in literature can arguably be best summarized in the words of Schekkerman [14, p. 6], who formulates this idea in the most striking way: "No strategy, no enterprise architecture". van't Wout et al. [11, p. 11] echo the same view almost verbatim: "No strategy, no architecture. No vision, no architecture".
