Mrinal Ghosh & George John

Governance Value Analysis andMarketing Strategy

The authors extend transaction cost analysis into a governance value analysis (GVA) framework to address mar-keting strategy decisions, especially with regard to strategies grounded in cooperative relationships. The GVA is afour-part model. Heterogeneous resources, positioning, the consequent attributes of exchange, and governanceform all interact to determine success in creating and claiming value. The trade-offs among these factors are thecore insight offered by the modei. The authors illustrate these trade-offs and specify empirically refutable implica-tions. Finaiiy, they sketch directions for future work and a blueprint for managerial decision making.

Transaction cost economics has become the dominantparadigm for analyzing issues in several areas of mar-keting, including interfirm relationships, channel

structure, foreign market entry, and so on. A large number ofempirical studies corroborate the principal refutable predic-tions that connect transaction attributes such as specific in-vestments, uncertainty, and performance ambiguity togovernance forms' such as vertical integration and alliances(for a review, see Rindtleisch and Heide 1997). Despitethese successes, transaction cost analysis (TCA) has madelittle headway into the marketing strategy literature (notableexceptions include Day and Klein 1987; Dwyer and Oh1988). Indeed, some of the sharpest criticism of TCA hasemerged from the strategy-oriented work, and it convergeson common themes,

One strand contends that TCA is a mere cost-minimiza-tion calculus based on exogenous attributes of an exchange,and as such, it provides little insight into strategic marketingchoices that are grounded in firm-specific differences.Hence, Hunt and Morgan (1995) charge that TCA essential-ly predicts that all competing firms in a market will chooseidentical governance forms (such as vertical integration) be-cause they all face the same level of exogenous attributesimplicated in the theory. Plainly, this is unsatisfactory, as ca-sual observation of the sharp differences between competinglirms in the same market demonstrates.

Another strand charges that TCA's single-minded focuson co.st minimization provides little insight into strategicmarketing choices that are undertaken principally to en-hance and/or claim value- (Zajac and Olsen 1993). Accord-ing to this view, strategic alternatives that are more effective

'Governance forms are the formal and informal rules of ex-change,

'A traction of the gross customer benefits flows back to the firmthat creates and delivers the product to the customer. After sub-tracting costs from this fraction, the remainder is termed "value."Joint value is similar, except that it includes all the firms that co-operate to create and deliver the product. We use value, profits, andmargins interchangeably in this article,

Mrinal Ghosh is Assistant Professor of Marketing, University of MichiganBusiness School. George John is Research Professor of Marketing, Carl-son School ol Management. University of Minnesota.

(in the sense of increasing profits) could be attractive de-spite concomitant losses in efficiency. A classic examplewould be expenditures to raise entry barriers even if the ad-ditional costs meant that some marginal output had to besacrificed.

On occasion, these strands are joined to reject any valuewhatsoever that practicing managers can derive from TCA(e.g., Ghoshal and Moran 1996). When these critiques arejuxtaposed against the increasing number of empirical stud-ies that largely corroborate refutable implications from thetheory, it suggests that some clarification and expansion ofthe approach is in order.

We propose a revision and expansion of the basic TCAmodel to enable it to address marketing strategy decisionsmore closely. Specifically, we propose and answer questionssuch as the following:

•Why do firms in the same industry demonstrate starkly dif-ferent approaches to bringing their products to market?

•How do positioning differences between firms infiuence thedesign of their supply chain and end-customer govi.-nanceforms? and

•How do customer brand equity, technology, and channelresource differences between firms inlluence their supplychain and end-customer governance forms?

To distinguish our variant, we label it "governance val-ue analysis" (GVA), in the spirit of Zajac and Olsen's (1993)label, "transaction value analysis." We emphasize both val-ue maximization and cost minimization; we incorporate het-erogeneity from the resource-based view (RBV) of strategy(Peteraf 1993) and accommodate path-dependent influences(Ghemawat 1991). Day and Klein (1987) offer the first stepin this direction of integrating strategy and TCA. We buildon their foundation and capitalize on recent empirical andconceptual advances (e.g., Nickerson 1997) to support ele-ments of our framework.

Marketing Strategy from aGovernance Lens

Theoretical Foundations of TCA

Transaction costs associated with economic exchange in-clude ex ante and ex post costs. Ex ante costs include the

Journal of MarketingVol. 63 (Special Issue 1999), 131-145 Govemance Value Analysis /131

out-of-pocket costs of negotiating contracts ("ink" costs), aswell as the opportunity cost of forgone transactions. Ex postcosts are the out-of-pocket costs of enforcement agreementsand the opportunity cost of not shifting to more profitableactivities in the light of new information (Rindfleisch andHeide 1997). At its core, TCA is an elaboration of the im-plications of minimizing these costs. Although the relevanceof cost savings per se is self-evident, its relevance to mar-keting strategy decisions is not at all clear.

Williamson (1991, p. 76) first argued for the relevanceof TCA to strategy by noting that "economy is the best strat-egy," but the ramifications of this initial statement have beenunderdeveloped. To address this issue, we turn to his obser-vation tbat TCA can be summarized compactly as unpack-ing the ramifications of the Coase theorem (Williamson1996). According to this theorem, in the absence of transac-tion costs, parties to an exchange will devise joint valuemaximizing exchanges, regardless of their power differen-tials or resource endowments. Wemerfelt (1994) illustratesthe wide-ranging utility of this principle in his efficiency ap-proach to designing marketing systems. Of course, transac-tion costs are never really at zero in the real world, so it isthe subsequent implications for nonzero transaction costworlds that are really of interest.

With positive transaction costs, the core principle ofTCA is that parties will strive to align governance formswith exchange attributes to minimize the transaction costs ofexchange. Why? Because such an alignment gets them clos-er to the desired goal of joint value maximization. Notice thelogic connecting minimization of transaction costs with themaximization of joint value. In our model, we use this jointvalue metric as our criterion to judge the value of strategicchoices. As we elaborate subsequently, greater joint valuealigns itself with greater firm profits when the choice ofgovernance form is made along TCA lines.

Relationship to Other Strategy ModeisContemporary strategy models are coalescing around thecore processes of value creation and claiming. To discernthis, consider a brief overview of prominent models of strat-egy. In Ghemawat's (1991) model, creation and claiming ofvalue are the two fundamental processes that are necessaryand sufficient to understand strategy. Creation of value isachieved tbrougb favorable positions^ that are commitment-intensive, or difficult to reverse, choices that involve lock-inand lock-out effects. Claiming value from these prospec-tively favorable positions depends on "sticky" factors or re-sources accessible to tbe firm.'* Absent access to suchfactors, competitors and/or end customers claim the valuegenerated from a prospectively attractive position.

The RBV of strategy (for an overview, see Peteraf 1993)makes a similar argument. It takes a Ricardian view of com-petition in which greater profit fiows to the holders of supe-rior productive factors. Consider a simple stylized example

•'Positions are points in a muUidimensional space of customerbenefit attributes. Their favorabilility is assessed from the vantagepoint of the customer.

^Sticky resources are both scarce and appropriable by theirowners.

of Ricardian competition. Suppose that n firms compete in a

commodity tnarket. Each competitor possesses a differentset of assets and skills and thus is able to produce (i.e., cre-ate value) at a different cost (including cost of capital). Dcnote these costs as I, 2, ... n, from smallest to largest.Game-theoretic analysis reveals a market-clearing price of 2(the second-lowest cost); thus. Firm I enjoys an economicprofit equa! to the difference between the two lowest costs.Firm 2 will make a return that is just enough to cover its costof capital. All other firms will be unprofitable.

Observe the differences between RBV and the core roleof market power etnbedded in Porter's (1985) five-forcesstrategy model. The RBV invokes an efficiency view ofstrategy: Firm I's profits derive from its unique skills andassets. All the firms in this stylized commodity market be-have fully competitively, yet profit differences occur andpersist. If the skill differences between Firm 1 and its nearest competitor erode, prices fall and the created value is notclaimed.

This can be extended to more realistic markets with dif-ferentiated products with no loss in insight. Ricardian prof-its in these RBV models flow to firms that possessimperfectly mobile (sticky) resources that are superior t(»those of their competitors. In marketing. Hunt and Morgan(1995) draw on RBV principles to elaborate their resourceadvantage theory of competition.

Finally, Porter's (1996) recent strategy model establish-es positioning and operational effectiveness as the means bywhich value is created and claimed. He notes that creatingvalue is not a sufficient basis for strategic analysis if the firmcannot also claim its share of the value. He criticizes the recent trend of single-minded pursuit of operational effectiveness through practices such as benchmarking, outsourcing,and nurturing core competencies as dangerous half-truths.Both creation and claiming must be given equal attention.

The relevance of these contemporary strategy models toour effort is the burden it imposes on TCA as a strategyanalysis tool. Unless TCA is capable of disclosing insightsinto value creation and claiming, it is highly unlikely to be-come a useful tool for analyzing strategic marketing choices. We therefore examine TCA from the viewpoint olcreating and claiming value.

Value Creation in TCA

Webster (1992. p. 1) notes that the (re)emergence of coop-eration among firms is a "fundamental reshaping of the fielil[marketing strategy]." Writing to a strategy audience. Brandenburger and Nalebuff (1997) similarly observe that firmsrarely create value in isolation, particularly in advancedeconomies. Instead, they "align themselves with customers,suppliers and many others to develop new markets and ex-pand existing ones" (Brandenburger and Nalebuff 1997, p,4). Many observers of contemporary phenomena such assupply chains, strategic partnerships, and so forth echo thispoint. Much of the empirical work in TCA has focused onstudying such cooperative interfirm ties. To organize thesefindings, we begin with a stylized example of value creationin a cooperative setting.

Consider a firm. A, that has a set of resources and is con-templating entering into a supply arrangement for a compo-

132 / Journal of Marketing, Special Issue 1999

nent with another firm, B, that has its own set of resources.If A desires a standardized component, B can produce it atmarginal cost Co. Because the component is standardized,neither party is required lo make any partner-specific in-vestment.^ Therefore, the marketplace (through terminationof the arrangement) can discipline any inappropriate behav-ior by either party. Suppose that A's end product (which con-tains this component) offers gross benefits of VQ to the endcustomers, and the joint value (total margins) earned by thetwo firms is TUQ.

Alternatively, B might invest some money that is specif-ically targeted at this component. For example, it might re-design the component to fit better with A's final product. Italso might use different materials so that the weight of A'sproduct is reduced. Broadly speaking, the effects of a part-ner-specific investment, i, can be broken into two cate-gories. First, such investments might reduce the marginalcost of the component down to C[. Second, they might en-hance the appeal of A's end product in end customers' eyeslo a higher level, v,. Let the total margins earned in the newarrangement equal Ti;.

Neither firm desires the new exchange, with the addedinvestment cost, unless TCJ exceeds TUQ. This is fundamental toTCA, The model goes further: If the firms identify severalpossible JtjS. they will consider the one with the highest in-cremental joint value, as per the Coase theorem. This is theoften misunderstood efficiency aspect of TCA. It simplymeans that opportunities to realize further value should bepursued; not doing so is inetficient.

Absent such prospective gain, there is no need to con-(cmplate any special governance form. However, theprospect of gains is not the final story. We also must consid-er the process by which these parties claim their share ofthese prospective gains.

Value Claiming in TCARecall that the prospective increase in joint value (TIJ-TIO) isthe motivation for devising alternative governance forms.However, firms remain self-interested, so they will imple-ment the activities associated with the larger joint value ifand only if their own share of the joint value also exceedstheir previous profits. If complex, contingent contracts werefeasible, any side payment or sharing rule required to meetthis own-profit constraint could be devised and enforced.We do not mean to imply that the contract design problem istrivial. The agency theory and related mechanism designapproaches that otier insight into fashioning appropriatecontracts to split the gains are quite complex. Nevertheless,alternative governance forms lack significance.

Suppose, however, that there are limits to comprehen-sive contracting. For example, specific investments becomenoncontractible when third parties cannot verify them easi-ly because contractual enforcement involves third partiessuch as courts of law. Organizing a satisfactory split of thegains becomes nontrivial in these circumstances, and alter-

A partner-specific investment has significantly lower value inDthcr uses. Sunk costs are the defining characteristics of these in-vestments.

native governance forms become a significant matter. Thisis the arena of the standard TCA model.

The theory makes specific assumptions about the par-ties' behavior with respect to ex post bargaining behavior tosplit the gains. Although they cooperate to produce thegains, they remain self-interested to tbe point of oppor-tunism^ within this ex post bargaining process. As such, nei-ther party will accept less than the profits available fromtheir individual next-best alternative. This minimum is sim-ply their ex post^ costs. Each party's ex post bargainingstrength determines its share of the gains beyond this mini-mum and is a function of three critical attributes of the trans-action: specific investments, adaptation probletns, andperformance measurement problems. Greater exposure toeach of these attributes reduces the ex post bargainingstrength of the focal party.

Both parties realize that each will attempt to minimizeits ex post disadvantage at the value claiming stage by (I)scaling back investment, (2) adapting less, and (3) forgoingactivities that are hazardous from a measurement stand-point. All these steps reduce value creation and thus moti-vate the search for better govemance forms. Put differently,unless governance mechanisms are devised to manage thevalue claiming problem, value creation is affected negative-ly. Thus, aligning exchanges with governance forms to max-imize net value is the core prohlem of the standard TCAmodel. The integral role of value maximization in this align-ment process is found most clearly in the incomplete con-tracting models that formalize this TCA argument (e.g.,Bakos and Brynjolfsson 1993; Grossman and Hart 1986). Inthese models, the governance form offering the greatestjoint value from the set of feasible alternatives is the one se-lected in the alignment process.'

Aligning Exchanges with Governance Formsin TCAWe have noted that TCA characterizes exchanges alongthree critical attributes. Likewise, TCA clusters governanceforms into three types. Although we could draw up long listsof the elements of each governance form, it is possible tosummarize them compactly with reference to value creationand claiming.

Market governance describes the rules of arm's-lengthmarket exchanges. One variant uses spot contracts. Theseafford each party considerable autonomy and latitude toform and break ties, thereby keeping incentives aligned withchanging circumstances, which enhances the ability to take

^Opportunism is self-interest seeking with guile. There are twodistinct varieties of such behavior. First, parlies may engage in be-havior that reduces their own cost, regardless of its effects on thetotal proHls. An example is shirking behavior Second, they mayengage in behavior ihat imposes costs on their trading partner toforce a more favorable rearrangement of the original terms oftrade. Examples include "hoid-up" threats and haggling.

^Notice thai ex post costs do not include sunk costs.Hn particular, we focus on Bakos and Brynjolfsson (1993), who

provide a numerical illustration of the genera! model to demon-strate thai reducing the number of suppliers improves joint valuewhen suppliers are a more important source of specific invest-ments, and vice versa.

Governance Value Analysis /133

advantage of new opportunities (to create value). Valueclaiming depends entirely on the threat of loss of futurebusiness acting as the safeguard against opportunistic be-havior. Another variant of market governance relies on com-plex contingent contracts to define more specific burdens onthe parties. Although this sharpens further the incentives forcreating value, it suffers at the value claiming stage becausethese more precisely defined burdens cannot be verified eas-ily by third parties.

Hierarchical governance describes the rules of internal-ized or vertically integrated exchange. In contrast to marketgovernance, it suffers from lower-powered incentives forvalue creation but is not as vulnerable to opportunism at thevalue claiming stage. It relies on de facto ownership or con-trol over the relevant assets and activities to safeguard theclaim to a person's share of the value created.

Relational governance includes a host of diverse formsthat combine elements of the previous two types. Examplesinclude franchising, partnering, and alliances. The uniqueaspect of relational governance is its reliance on mutualpartner-specific investments, paired with implicit socialnorms to safeguard claims to shares of the value created. Al-though it is difficult to generalize across these diverse rela-tional governance forms, roughly speaking, their incentivesfor value creation are stronger than hierarchical governancebut poorer than the high-powered incentives of market gov-ernance. Likewise, their value claiming properties arestronger than the tenuous safeguards afforded by marketgovernance, but they are weaker than the safeguards of hi-erarchical governance.

Each governance form is best aligned with particular ex-change attribute levels. Because the empirical work on thealignment between exchange attributes and governance or-ganizes itself according to the exchange attributes, we re-view extant findings in this fashion to disclose tbe utility ofthe core TCA model, as well its shortcomings for informingmarketing strategy decisions. We rely heavily on Rind-fieisch and Heide's (1997) recent review for summarizingthe thrust of the empirical work.

Securing specific investments. When partner-specific in-vestments offer the prospect of enhancing value creation,this poses a dilemma. On the one hand, tbere are ex post out-of-pocket costs from opportunistic behavior. On the otherhand, scaling back these investments or forgoing the deal al-together involves ex ante opportunity costs.

The parties are motivated to devise governance formsthat possess sufficient safeguards to secure these valuablebut vulnerable investments in order to get as close as possi-ble to joint value maximization. Although the empiricalwork on this particular prediction is quite significant andshows consistent support, virtually all the studies rely on areduced-form version of the prediction; namely, the pres-ence of larger specific investments will be associated withstronger governance safeguards. Tests of this reduced-formprediction are not capable of corroborating the value maxi-mizing argument embedded in the prediction. Indeed, thereduced-form test suppresses the role of value maximizationthat is inherent in the preceding conceptual argument and in-vites counterarguments and confusion because the specific

investments appear to t>e exogenous. The very reason for thespecific investment is obscured.

This contrast between tbe theory and the reduced-formtest in the empirical studies confuses many readers and, in-deed, some of the critics. For example, Zajac and Olsen(1993, p. 132, emphasis added) aver that "transaction costapproach,.., in neglecting the issue of joint value, can lead ...to faulty analysis." Their position is clearly not consistentwith the theory reviewed previously.

A more complex alternative explanation is offered bythe RBV^ view of strategy. According to tbis view (e.g.,Collis 1991; Ghoshal and Moran 1996), the robust empiricalfinding tbat specialized activities tend to be governed hier-archically (internalized) can be reinterpreted as follows: In-ternalizing an activity facilitates the deployment of morefirm-specific skills, language, and routines against that ac-tivity. Therefore, the firm internalizes those activities (orcompetencies) that contribute to creating superior value.Furthermore, internalization protects them from imitation.Rather than market failure on account of transaction costs, itis the advantage of hierarchical governance forms over mar-ket governance in using specific skills for value creation thatdrives the decision to invest in more specific assets. As astrategic choice, firms expand (internalize) into areas tbatare adjacent to their existing (internalized) specific assets.Notice that this ignores the TCA emphasis on comparativeanalysis and does not ask whether such expansion might notbe achieved by contractual means, such as licensing, ratherthan internal expansion. Parenthetically, we note that two re-cent tests of these competing explanations (Poppo andZenger 1998; Silverman 1998) support the TCA causalprocess over the RBV argument.

From a strategic marketing standpoint, TCA offers valu-able insigbt into the sticky factor of specific investmentsthat is also implicated in the strategy models just reviewed.However, the shortcoming of the TCA work is that specificassets are not treated fully endogenously. Studying market-ing strategy decisions effectively requires that we endoge-nize specific investment decisions more explicitly.

Facilitating adaptation to uncertainty. Often, parties toan exchange find tbat circumstances require tbem to changepreviously planned courses of actions and decisions involv-ing existing assets and/or abandon previous investments infavor of striking out in new directions.'^ As before, there isan opportunity cost of value not generated by a failure toadapt, as well as out-of-pocket costs of communication andhaggling over revisions in tbese circumstances.

To the extent that revisions can be foreseen or are small,it suffices to devise contingent contracts (i.e., do A if X oc-curs, do B if Y occurs), and alternative governance formslose their significance. However, when these revisions areunforeseeable and expected to be large, contingent contract-ing is not feasible or is at least incomplete. Given noncon-

^"Knowledge-based" and "resource-based" are used inter-changeably in the strategy literature,

"*We stress that both revisions of effort and obsolescence of in-vestments are at issue.

134 / Journal of Marketing, Special Issue 1999

traclability, governance forms must be devised to align withrevision and/or exit needs.

Specifically, the likelihood of obsolescence of previousinvestments calls for governance forms tbat facilitate fasterabandonment, whereas a greater need to revise effort callsfor governance forms that facilitate faster revision. Again,observe that the alignment is undertaken to ensure valuemaximization (accomplished through transaction cost mini-mization). This argument also is found most clearly in theformali/ed models that are emerging. For example, Wemer-Iclt (1997) develops a model in which value is maximizedby aligning hierarchica! governance forms with exchangesin which frequent or diverse changes are very important.

A large number of empirical studies exist on this matter.They are largely corroborative (see Rindfleisch and Heide1997), but they all test two reduced-form predictions; that is,faster obsolescence is associated with shorter duration mar-ket governance forms, wbereas greater revisions of effortarc associated with longer duration relational governanceforms. These reduced-form tests cannot distinguish the op-portunity cost of the forgone value of a failure to adapt fromilie out-of-pocket costs of haggling. Even those rare studies[hat attempt to measure transaction costs inevitably trackonly the out-of-pocket costs {e.g., Noordewier, John, andNevin 1990). Tbe value creation aspects again are obscured.

From a strategic marketing standpoint, TCA offers in-sight into effective ways to create and claim value in uncer-tain environments, but as previously, tbis is not treated as afully endogenous choice. Choices about served markets af-fect the need to revise and exit, yet this choice is never ac-counted for. This gap is reminiscent of the organizationalihcory literature linking structure to environmental uncer-(iiinly that likewise failed to account for the possibility thatlinns consciously may choose to avoid or accept uncertain-ly hy selecting their task environment. Thus, the GVA mod-el must cndogenize the uncertainty that arises irom strategicchoices about markets and customers.

Accommodating performance measurement difficulties.It is fundamental to market governance that the parties to anexchange measure value given and received. Absent effec-tive measurement, the high-powered incentives of marketgovernance fail. We define per;fbrniance measurement diffi-culty as the degree to which the value of an actor's contri-hution is not verifiable by ex post inspection of output.

Performance measurement difficulties entail a reductionin value creation in two respects. First, there is an opportu-nity cost of the failure to identify and motivate the right ef-lort (i,e.. misdirected effort). Second, there areout-of-pocket costs of shirking (i.e.. insufficient effort) as-sociated with monitoring, enforcement, and termination.

In TCA, exchange parties are motivated to devise gov-ernance forms that accommodate measurement difficultiescommensurate with value creating activities. Absent suchgovernance forms, the parties must either shift to alternativeactivity sets that generate lower value but are easier to mea-sure or incur large out-of-pocket costs. Notice, however, it isnot the pursuit of performance measurement (or the mini-mization of these costs) per se that matters but the pursuit ofvalue through adequate measurement.

Again, it is the reduced-form prediction that is tested inthe empirical work; namely, performance measurement dif-ficulties are associated with more hierarchical governanceforms. Rindfleisch and Heide's (1997) review finds strongsupport; however, such support cannot distinguish the op-portunity costs of misdirection from the out-of-pocket costsof shirking. Tbe language in these studies focuses almost ex-clusively on the costs of monitoring and ignores the oppor-tunity cost of forgone trade and value creation.

From a strategy standpoint, the challenge is clear.Strategic choices about markets and products profoundly af-fect the level of measurement difficulties. As previously not-ed, firms do not pursue minimization of performancemeasurement difficulties per se. They choose a level of per-formance measurement difficulty that maximizes value, giv-en their strategic choices of markets and products. Thus, toaddress tbese matters, the GVA model must link firms'choices to performance measurement difficulties explicitly.

Summary. Three points emerge from our brief review.First, the empirical work lends support to the robustness ofthe core TCA model. This gives us confidence to use it as afoundation. Second, value maximization is integral to TCA.Unfortunately, the tests of reduced-form predictions drawattention only to the out-of-pocket costs of inefficient ex-change (tbe friction costs), which creates the impression thatTCA has a single-minded obsession with cost minimization.This is unfounded. Tbird and most important, tbe effects ofstrategic choices and heterogeneous resources of firms onthe alignment of exchange with governance are conspicu-ously absent in the standard TCA model. Our expandedframework redresses tbese imbalances.

The Governance Value AnalysisModel

Basic StructureIn Figure I. we show that the GVA model consists of fourcore constructs. It begins with the familiar constructs fromthe TCA model reviewed previously (i.e., attributes of ex-change and governance form) and incorporates two addi-tional constructs, positioning and resources.

Positioning is defined as the particular bundle of bene-fits selected by the firm to be created and delivered to thetarget customer. Popularly called ihe "value proposition,"this benefit-target pair is arguably the most fundamental de-cision in setting a marketing strategy. Given the plethora ofpositioning strategies, it is not possible to generate a catego-rization of generic positioning possibilities. Instead, we usethe multiattribute benefit space model, lamiliar to mar-keters, to characterize positions. Thus, we talk about highperformance (or ease of use) as a position.

Resouives are defined as the scarce and imperfectly mo-bile skills or assets owned by a party to an exchange. Previ-ous research efforts to catalog tbese resources areremarkably similar in the categories identified. For example.Day and Wensley's (1988) list can be categorized into man-ufacturing (engineering and design skills, availability of as-sembly lines, and so on), channel marketing (breadtb of

Governance Value Analysis /135

FIGURE 1Governance Value Analysis Model

Positioning

#1

Exchange Attributes• Specific investments• Uncertainty• Performance measurement

#5Governance Form• Market• Hierarchies• Relational

Firm Resources• Technology• End-customer• Supply chain

coverage, sales force skills, cooperative linkages within thechannel, and so on), and end-customer marketing (branding,level of advertising expenditure, level of service provided)capabilities. Simon and Sullivan (1993) organize them intotechnological (patents, trademarks, research and develop-ment, knowledge), marketing (brand equity, advertising),and industry (market structure, capital intensity) factors. Us-ing these efforts as a basis, we categorize resources into (1)technological, including unique equipment, processes, andpatents; (2) end-customer, including brand equity, customerloyalty, switching costs, and market share; and (3) supplychain (channel}, derived from extant relationships with trad-ing partners and including trust and goodwill of channelpartners. We note that our categorization is by no means de-finitive but rather provides a useful point of departure forour discussion. For our purposes, it is sufficient that thethree resource categories are qualitatively different in theirimpact on value creation and claiming. Technological re-sources give the firm access to superior production andquality, which reduces the cost of delivering customer ben-efits and enables a firm to claim value through uniquenessand scarcity of the product. End-customer resources give thefirm access to positions that are more highly desired hy endcustomers and insulate the value of these positions fromcompetitive erosion. Finally, supply chain resources give the

firms access to superior partners, which is critical whencomplementary components come together to create and de-liver customer benefits.

Our core thesis is that the four constructs in GVA inllu-enee one another mutually. Put differently, the "best" strate-gic choice is the positioning option that is best matched tothe resource position of the firm, the exchange attribute lev-els, and the governance forms deployed to manage its sup-ply chain and end-customer exchanges. The real value of themodel lies in the explication of specific value maximizingmatching rules. However, we refrain from articulating anoverall global prescription for strategy similar to the one de-scribed in Nickerson's (1997) TCA model of strategy. We donot view the firm as making one giant transaction andmatching it to a single governance structure." Instead, GVAis a middle-range model that addresses individual marketingstrategy choices in individual product markets. Firms makemany such choices at different points in time, but these in-dividual strategy decisions are tied together because the re-sources of tbe firm transcend these individual situations.Resources typically operate at the strategic business unit(SBU) or higher (firm) level, whereas positioning operatesat the product market level within an SBU. Attributes of the

"We thank an anonymous reviewer for this language.

136 / Journal of Marketing, Special Issue 1999

exchange operate al or even below the level of the productmarket (eg., segment level), a.s does governance form. Con-sequently. GVA is a mixed-level model.

As we noted previously, positioning, resources, ex-change attributes, and governance form are chosen simulta-neously and interactively, but this level of analysis is animpossibly complex matter to unpack, so we focus on a sub-set of the causal links. Although we recognize that position-ing and resource.s influence each other (Figure 1, #1), weorganize our discussion into separate sections on positioningand resource effects.12

In the positioning section, we illustrate its direct influ-ence on exchange attributes (Figure 1. #2) and governance(Figure 1. #4), as well as on exchange-governance align-ment (Figure 1, #5). Likewise, in the resource section, we il-lustrate its direct influence on exchange attributes (Figure I.#3) and governance (Figure I. #6). as well as on ex-change-governance alignment (Figure 1. #5). We proceedhy illustrating empirically refutable matching rules. Whenpo.ssible. we discuss supply chain and end-customer ex-changes separately because their governance form possibil-ities are markedly different. For example, verticalintegration of an end customer by the firm is not a logicallyfeasible option, whereas vertical integration is a fundamen-tal choice in supply chain exchanges.

Positioning EffectsSpecific investments by supply chain partners. Suppose

that an automobile firm were to position its products as highon reliability and fit/finish attributes. For example, Toyota'scars are often described in magazine reviews as exemplify-ing unexceptional designs and providing high reliability andexcellent til/finish. Whitney (1992) notes that high reliabil-ity and tit/finish require the firm to innovate on assemblyprocesses because reliability depends greatly on the qualityof assembly. It is critical to note that innovations in manu-facturing prtKesses are much more specific to an assemblyline than the components of an automobile, so the transac-tions between this firm and potential sources of theseprocess innovations will need to be supported by high levelsof specific investments.

Another example is the airline marketplace. An airlinepositioned to offer seamless, end-to-end travel requires ahuh-and-spoke operation, which involves making muchgreater specific investments than an airline that positions it-self to compete on a city-pair basis (Nickerson 1997). Thisoccurs because hub assets are far less redeployable thanplanes and other airline assets.'-^

is a controversy about the mulual influence of resourcesand p<isitioning: RBV cniphasi/es the primacy of resources in de-termining positioning, whereas Porter (1996) emphasizes the dis-cretion in making the positioning choice. We do not join this debatehecause no new theoretical leverage is afforded from the TCA per-spective on this matter.

"Indeed, the absence of specific investments by airlines in theprehub era gave regulators the confidence to deregulate airlines de-spiic highly concentrated city-pair routes. Recent analysis points tothe sunk costs of a hub as the principal competitive weapon and en-iry-deterring device in the contemporary airline marketplace.

We reiterate that these positions are endogenous choic-es. Thus, two firms competing in the same prtxluct marketmay choose different positions with different implicationsfor attributes of their exchanges with supply chain partners.To return to the automobile example, one firm might posi-tion itself on reliability, whereas a competing auto firmmight select a high-performance position that dcemphasizesreliability and fit/finish. A sports car manufacturer might usethe latter position, which demands that the firm organize touse the best available technology and components from out-side sources. Its component suppliers must be persuaded tobuild higher performance components that are likely to becustomized for the original equipment manufacturer(OEM). This requires specific investments in the OEM-sup-plier interface. In contrast, because its assembly proces.sesneed to be supported by fewer specific investmenis, the or-ganization of this part of its supply chain will be quite dif-ferent than the previous firm's.

Whitney (1992) provides empirical evidence about thedramatic differences between General Motors (GM) andToyota that is consistent with our argument. Toyota is high-ly vertically integrated into its manufacturing processes, in-cluding computer-aided design and manufacturing softwareand equipment such as machine tools and assembly robots,whereas GM is much less integrated in these aspects. Thisflows naturally from the need for Toyota's assembly andmanufacturing processes to be supported by much higherlevels of specific investments. The causal sequence is alsoconsistent with the idea that, even if GM desired to empha-size fityfinish, il would be less successful in realizing such apositioning choice unless its supply chain exchanges for in-frastructure items were realigned appropriately.'''

Specific investments by end customers. The influence ofpositioning on specific investments extends to the end-cust()mer exchange as well. To illustrate, a key positioningdecision in many high-tech markets pertains to compatibili-ty (or lack thereof) with competing products. Basically,firms must choose between a closed or proprietary positionat one extreme and an open standard at the other extreme.With a completely proprietary position, a firm designs itsproduct without any concern for interoperability with othercompeting and/or complementary products, Its interfacesare kept secret or patented. In contrast, strict adherence to anopen standard requires firms to design their products in con-formance with perfonnance and/or interface specificationsthat are openly available to everyone.

These compatibility decisions vary across firms in thesame market and arc based on their respective technologyand design choices and their differing ahility to convert cus-tomers to their proprietary standard. To illustrate. AppleComputer chose to position itself in a proprietary (closed)standard versus the de facto open standard position chosenby IBM and its many subsequent clones (Wilson, Weiss, andJohn 1990).

'••Parenthetically, it is interesting to note that this is a role rever-sal for these two firms from their better known differences in com-ponent purchasing, Here. GM is much more vertically integratedthan Toyota's famed subcontractor network (keiretsu).

Governance Value Analysis/137

When a firm chooses a proprietary position, its end cus-tomers' expenditures in people, processes, and equipment tosupport and use the product become vendor-specific invest-ments. Expenditures of the same magnitude are much lessvendor-specific when the seller chooses an open standardposition. To observe this, compare an end customer's in-vestment in Apple-specific software and peripherals thatcannot be used with any other vendor's computers, unlikesoftware and peripherals that can be moved from one ven-dor's IBM-compatible machine to another.

The familiar TCA safeguarding principle applies withfull force here: Proprietary positions require stronger safe-guards for customers' investments. For example, strongercommitments by the vendor may be needed to reassure cus-tomers that spares and support for peripherals and productswill continue even after the particular item is no longer sold.In contrast, these matters can be left to third parties for openstandard products. This illustrates an important trade-off.Because the stronger commitments are costlier governanceforms, the proprietary position that necessitates them is on-ly worthwhile if tbe position delivers a much higher level ofbenefits to customers. Thus. Apple computers must delivergreater levels of desired attributes compared with IBM com-patibles to support their more complex and costlier customergovernance forms.

To summarize, positions emphasizing process-derivedattributes such as reliability and more proprietary locationsrequire greater specific investments by supply chain part-ners and end customers, respectively (Figure 1, #2). Gover-nance forms must be aligned with these investment needs(Figure 1. #5), or competitive disadvantages will followpredictably.

The second critical exchange attribute in GVA is uncer-tainty in exit and revision respects. Positioning choices af-fect the uncertainty attribute of supply chain exchanges, aswell as end-customer exchanges (Figure 1, #2). They alsoinfiuence governance choice directly (Figure I, #4) and in-directly (Figure I, #5). We turn to these effects next.

Uncertainty in supply chains. The cola market has beencharacterized in recent years by greatly increased promo-tional campaigns by the two market sbare leaders. Constant-ly changing advertising campaigns by Coca-Cola and Pepsithat are tightly coordinated with complex promotional cal-endars stand in contrast to those by smaller firms that haveengaged in far fewer campaigns and less complex promo-tional calendars. In short. Coke and Pepsi are positioned asheavily promoted brands. Although their list prices typical-ly exceed those of the smaller share brands, their promo-tional efforts ensure that one of the two leaders is availableat a promoted price at virtually any given time in a store.

These intense marketing activities have increased theneed for the leaders' supply chain exchanges to be preparedto revise coordinated actions often and adapt quickly rela-tive to their smaller competitors. Muris, Scheffman. andSpiller (1992) argue that the ongoing buyout of independentbottlers by the two market share leaders represents a shift ingovernance form to align better with the shift in positioning.They estimate that the two firms currently own or have anequity interest in bottling operations that account for ap-proximately two-thirds of their volume. The standard TCA

prescription of increased hierarchical governance in re-sponse to greater adaptation requirements accounts for thisshift. Notice that if the smaller share firms also chose to in-crease their promotional marketing intensity, they too wouldneed to realign their governance forms. However, thesesmaller firms are disadvantaged because the economies ofscale in bottling preclude them from using captive bottlersas a governance form.

An historical anomaly in this market pinpoints the com-petitive advantage of aligning positioning with appropriategovernance forms. Although Coke and Pepsi have virtuallyidentical marketing systems for bottled and canned softdrinks, their systems are quite different in the fountain mar-ket. Coke historically has employed a vertically integratedsystem for fountain sales, whereas Pepsi's independent bot-tlers have enjoyed perpetual exclusive franchises. If adapta-tion needs have increased because of the positioning shifts,we would expect tbat Pepsi increasingly would be disad-vantaged in the fountain market because of its reliance on aless well-aligned governance form (independent bottlers).This disadvantage is magnified because fountain customerssuch as fast-food restaurants often cross bottler territoryboundaries, and coordinating multiple bottlers to servicesuch national accounts is difficult.

The market share differences are remarkable. Pepsi hascaught up with Coke in the bottled and canned market, butCoke continues to enjoy a huge market share advantage inthe fountain market. We conclude this is due to its continu-ing advantage in governance fonn.15

To summarize, positioning decisions targeting multi-site, national account end customers and emphasizing high-frequency promotional calendars require facilitating higherlevels of adaptation (Figure I, #2) by the partners in thesupply chain compared with their competitors. Governancemechanisms must be aligned with these adaptation needs(Figure I, #5). or competitive disadvantages will followpredictably.

Performance measurement of supply chain partners. Thethird and final critical exchange attribute in GVA is the per-formance measurement dimension. The ability of the partiesto measure the benefits and cost of each other's contribu-tions varies enormously across different positions (Figure I.#2). We identify two pervasive measurement issues linkedto positioning choices. First, cost reduction efforts in supplychains pose far fewer measurement problems than do bene-fit enhancement efforts. Second, costs are far more verifi-able than benefits, because the latter is often in the eye of thebeholder

For example. Lincoln Electric positions itself as a high-end supplier of welding and related equipment. Many low-er-priced providers offer functionally similar equipment. Allof these products can perform the minimal task expected ofthem, and this capability is readily measurable. In contrast,the high-end positioning of Lincoln Electric emphasizes lessverifiable attributes such as more reliable operation and bet-ter quality. Lincoln transforms these benefits into a cost-

' Alternative explanations abound for individual case histories.We posed this explanation to a senior Pepsi executive who insistedthat the fountain market was not profitable anyway to either firm.

138 / Journal of Marketing, Special Issue 1999

reduction position in the following way: Instead of the sim-ple equipment sales contract normally used in this market.Lincoln uses a complex capitation program called "Guaran-teed Cost Reduction."'^ Its Held engineers first audit a cus-tomer site and develop a target cost for welding operations.Lincoln then offers to install and service the required equip-ment, with its revenues tied to attaining and beating thesetargeted costs. Similar performance contracts are emergingin a variety of settings, ranging from management consult-ing firms to third-party logistics vendors. From a GVAstandpoint, such contracts as more valuable to competitorspositioned at the high end. as Lincoln Electric is.

Many contemporary procurement practices, such as just-in-time, efficient consumer response, and early supplier in-volvement, offer husiness customers the promise oflowering their cost of doing business as well as increasingtheir revenue stream (value enhancement). We conjecturethat the relatively easier task of measuring cost reductionsleads the participants in these initiatives to emphasize thisaspect, often at the expense of larger hut more nebulous val-ue enhancing effects. From a GVA standpoint, large strate-gic gains are available to marketers that can convert thesevalue enhancement promises creatively into measurableoutcomes.

Performance measurement by end customers. Sellers po-sition themselves at various locations in multiattribute ben-efit space. Positions at low leveis of the attribute dimensionsin such spaces pose fewer measurement problems than high-end positions. Verifiable perfonnance. certification, andstandards are all more robust at the low threshold level of anattribute.

End cu.stomers' difficulties in measuring benefits re-ceived for bigh-end product positions can be found in the lit-erature spurred by Klein and Leffier's (1981) model. Theyshow that a particular market governance form aligns wellwith positions that involve substantial levels of nonverifi-able product attributes. Briefiy. brand expenditures in con-junction with a price premium constitute a marketgovernance form that safeguards buyers against quality de-basement by the supplier. From a GVA standpoint, it is eru-ciat to note that the amount of nonverifiable attributes isminimal at some threshold level of the attribute(s) and in-creases with positions at higher levels. Positioning at thebaseline level of such an attribute or close to it does not re-quire as much spending on brands and price premiums tocreate and deliver value to the customer compared with thecrucial role of these costly governance forms for firms posi-tioned far above the baseline. Quantitatively, the cost of thegovernance forni increases dramatically (nonlinearly) as wemove away from the verifiable (enforceable) baseline (KleinandLeffier 1981).

The GVA mtxlel draws some further implications aboutthis market governance form. One implication is that anygiven brand is only credible in a certain price range and acorresponding range of the unverifiable attribute. This isconsistent with the price bands observed in retail store as-sortments. Another implication is that multiple brands with-

details are taken from a 1996 Lincoln Electric market-ing brochure titled "Guaranteed Cost Reduction."

in a corporate umbrella are needed to implement multipleprice tiers.

To summarize, positions emphasizing end-customerbenefits require greater accommodation of performancemeasurement difficulties than end-customer cost reduction(Figure I. #2). These difficulties are magnified for positionslocated at higher levels of the end-customer attribute(s) inquestion. Govemance mechanisms must be aligned withthese measurement problems (Figure 1, #5), or strategic dis-advantages predictably will follow.

Resource Effects

Technology resources. Access to superior technology isoften the basis for positioning products. Such a position isimmensely valuable in markets in which the technologydifferences are material to the performance of the product.According to GVA. firms that employ their superior techni-cal resources to position themselves as state-of-the-art sup-pliers faee more severe adaptation challenges (Figure 1. #3)compared with firms positioned as low-cost suppliers. Tocreate the value from their intended position, the gover-nance forms chosen by these firms (Figure I. #5) shouldvary accordingly.

To illustrate, consider technology-intensive capital equip-ment markets in which superior technology is the key resourcefor sustaining profitability. Firms that position themselves a.sstate-of-the-art suppliers in these markets often find they needto incorporate frequent engineering changes and upgrades tothe installed base of equipment at customer sites. Why?

Because capital equipment is relatively long-lived, signif-icant technology advances occur within the economic life ofa particular generation or version of equipment. Existing cus-tomers would be disadvantaged if technology developmentswere embodied in the newest models only and the installedbase was left unchanged. Such developments are relativelyless frequent for firms that position themselves as low-costproviders of capital equipment for two reasons. First, theirpace of innovation is slower. Second, it is expensive to main-tain the service staff to perform the field upgrades, and theirlow-cost position does not lend itself to such a strategy.

Our expectation of greater adaptation needs conforms toMasten and Synder's (1993) reanalysis of the famous Unit-ed Shoe antitrust case. Using court documents, they studiedthe use of leasing versus selling from a governance perspec-tive. Leasing aligns better with tnore severe adaptation prob-lems, because the retention of ownership by tbe suppliergives it the hierarchical authority to implement field up-grades. Consistent with this governance argument, their da-ta show that state-of-the-art suppliers such as United Shoetend to lease their shoemaking equipment, in contrast, itslow-cost competitors invariably sell their machines outright.The data also show that United Shoe machines were up-graded much more frequently. A more fine-grained predic-tion also is supported in this data. Adaptation needs are mostsevere for the most complex machines sold by the state-of-the-art suppliers because their technology advances are mostfrequent. Accordingly, state-of-the-art suppliers lease onlytheir more complex, technically advanced machines.

The positioning of IBM as a state-of-the-art vendor inthe mainframe market also illustrates this effect. Us selling

Governance Value Analysis/139

strategy in the mainframe market was characterized histori-cally by a strong reliance on leasing. Field upgrades of soft-ware and hardware improvements commonly wereundertaken across the entire installed leased base. In the1980s. IBM departed from this strategy, supposedly becauseof its high cost. This coincided with fewer field upgradesand more difficulty in getting new technology to market,which were attributed by knowledgeable observers to theproblem of convincing owners of expensive mainframes tobudget for frequent upgrades. Recently, there has been ashift back toward leasing.'^

To summarize, state-of-the-art technology resource-based positions increase adaptation demands (Figure 1, #3)and subsequently the governance form (Figure I, #5).

End-customer resources: Brand equity. Recall that end-customer resources describe those factors that give a firmaccess to positions that appeal more highly to customers.Brand equity is such a resource. It increases the appeal of afirm's product such that customers are willing to pay a high-er price for the branded product. Firms in the same marketvary in their brand equity, and higher margins earned by onebrand compared with other brands in the same category area good indication of its greater brand equity.

The GVA model argues that brand equity influencesgovernance directly (Figure 1, #6). Specifically, strongerbrands are less able lo use relational govemance forms withsuppliers, whereas weaker brands are more capable of usingit. Conversely, stronger brands are more able to use marketgovernance, whereas weaker brands are handicapped in us-ing it. We illustrate these principles next.

Suppose that specific investments'^ are required to sup-port the development and manufacture of a component sup-plied to an OEM. As per standard TCA reasoning, thesespecific investments must be secured with appropriate gov-ernance forms (Figure 1. #5). In many cases, the preferredgovernance form consists of relational ties. Why is the abil-ity to secure investments with relational governance poor-er when the OEM has larger margins in the downstreammarket?

The downstream appeal of the OEM's product is a func-tion of the appeal of the focal component, as well as the at-tractiveness of the rest of the product, including OEM brandequity. When the OEM has greater brand equity and the itemis tailored to the OEM. customer appeal is enhanced andthere is superior value creation. However, such an OEM isin a more vulnerable position at the value claiming stage. Allthe margins earned by the downstream product, includingthose generated by the OEM's own brand appeal, and unre-lated to the activities of the supplier are nevertheless "on thetable." They become subject to the bargaining process in re-lational governance that determines the value claimed. No-tice that in the limit, if the OEM were to earn onlycompetitive returns, the hargaining only would occur overthe joint value created in the exchange because there wouldbe no additional margins on the table.

''This account is derived from two persona! interviews withmarketers with long experience in the mainframe industry.

"*The same argument would hold for adaptation and measure-ment, the other two drivers, as well.

More precisely, the problem is that the incremental val-ue to be split between the parties (TIJ - KQ) should consistideally of the value generated from the exchange-specificactivities. It should not include value generated from otherresources such as brand equity. Unfortunately, these arecommingled, to the detriment of the party with the greaterbrand equity. One party cannot appeal to ex ante bargainingpower to solve this problem because ex post bargainingpower is what matters in TCA. Ex post bargaining power isreduced as brand equity increases, as was shown previously.

As always, if contingent contracts were feasible, theOEM could "seal off these other unrelated margins, Firm,fixed prices are one such means. This is unrealistic becauserelational governance would not be considered in the firstplace if such contracts sufficed. In summary, relational gov-ernance forms available to weaker brand equity OEMs areunavailable to stronger brand equity OEMs. In effect, thelatter firms fear their downstream profits will be siphonedaway if they develop close ties with suppliers. Helper andLevine (1992) first drew attention to this trade-off and con-tended it explained why a firm such as Chrysler, with rela-tively smaller margins, was quicker to adopt closer, valueenhancing keiretsuWke ties with its suppliers compared withstronger brands such as GM and Ford.

As a corollary. OEM buyers with stronger brands profitfrom using a larger number of competing component sup-pliers compared with a weaker OEM. all else being equal.Why? Observe that the importance of the suppliers' partner-specific investments relative to the importance of the OEMbrand asset increases with weaker brands, all else beingequal. Reducing the number of competing suppliers giveseach remaining supplier a better hargaining position withwhich to secure their investments. Offering such security loinduce these investments is more valuable to an OEM buy-er with a relatively weak brand. Thus, such finns shrink thenumber of vendors more aggressively than firms withstronger brand equity.

In contrast to the disabling effect of hrand equity on re-lational governance. GVA contends that it has an enablingeffect on market governance. Suppose that end customersneed to make vendor-specific investments to use a firm'sproducts more effectively. For example, a customer migbtbuild a library of training videos in a single proprietary for-mat (e.g., tbe U-Matic video format) that might reduce theirtraining and maintenance costs. In this scenario, the cus-tomers' investments must be secured against lock-in haz-ards. Otherwise, at the margin, some customers will scaleback or forgo such investments, with the familiar conse-quence that some value creation opportunities are lost andoverall sales are reduced.

Previous TCA analysis (e.g., Dutta and John 1995)shows that such customer investments can be secured with aparticular market governance form called "second-sourcing." Here, the focal firm licenses its proprietary for-mat free of royalties (or at nominal cost) to other competingfirms.1^ The "invited" competition between the licensor and

is is distinguished from the act of licensing out technologyor other intellectual property for a revenue stream of license fees.

140/Journal of Marketing, Special Issue 1999

second-source licensee safeguards the customer's value cre-ating investments.

This is a mixed blessing. Plainly, such invited competi-tion can wipe out profits to both firms. From a GVA stand-point, the potential value creation from second-sourcingmust be balanced with a concern for claiming value withsecond-source competition. As it turns out, tbis balancing oftensions favors parties with greater brand equity in the useoi this governance form.

Specifically, if the originating firm does not have suffi-cient brand equity (or other unique technical features) to dif-ferentiate itself after compatible products from thesecond-source have appeared, the value created by second-sourcing will be claimed almost totally by the user commu-nity. Xerox in the Ethernet market. IBM in the personalcomputer market, and AT&T in the UNIX market illustratethis effect. In each case, the additional value created forusers was huge, as was evidenced by the enormous size ofthe markets created (Varian and Shapiro 1998). Neverthe-less, little value was claimed by the firms that created the defacto standards (and "invited" competition from clones) togrow these markets. Put differently, strongly differentiatedfirms are more likely to be able to employ this market gov-ernance torni profitably.

A related balancing effect between value creation andclaiming occurs when the core item per se is valued highlyby its customers. Here, lock-in hazards deter only a smallnumber of potential customers. In these circumstances, thefocal firm might forgo the additional value created from sec-ond-sourcing to claim a larger piece of a slightly smaller-sized value pie. Sony's experience with the Beta tape formatis a gocxi example. This proprietary format succeeded in theprofessional marketplace because of its highly valued posi-tion in tliese custttmers" eyes, despite the insecurity of theirinvestments. In contrast. Beta was overwhelmed in the con-sumer market by the "clones" adhering to the VHS standard.The intrinsic superiority of the Beta format over VHS wasnot t)f sufficient value to casual consumers to offset the an-cillary benefit of a greater number of prerecorded moviesexpected in the de facto standard VHS format.

To summarize, the brand equity of a firm infiuences thegovernance form in very direct ways (Figure 1. #6). Strongerbrand equity makes it less feasible to exploit value creatingopportunities that require relational governance forms butmore feasible to exploit value creating opportunities tbat re-quire securing investments with market governance.

End-customer resources: Market share. The marketshare of a firm is a prominent source of heterogeneity acrosscompetitors in virtually every market, and any model pur-ptirting to inform strategic decisions should be capable ofanalyzing market share effects. The GVA model carries thesame burden, and we turn to this task next. Two size-relatedeffects can be isolated that are principally applicable to sup-ply chain exchanges.

Williamson (1989) provides an early TCA treatment ofsize effects. Assuming some economies of scale in produc-tion in an industry, he reasons that larger competitors aremore able to use hierarchical governance forms at lower lev-els of sjiecific investment compared with smaller competi-tors (Figure 1, #6). His logic is that the larger competitors

sacrifice fewer scale efficiencies from internal production,so they are better able to use this safeguard to secure specif-ic investments (or to adapt to uncertainties or better measureperformance).

A related effect of size is that hierarchical governanceforms are less feasible for smaller firms, on account of set-up costs. The classic marketing example is the minimumlevel of sales needed to support the setup costs of a compa-ny sales force. Therefore, smaller-share firms often mustrely on independent representatives. Using the GVA model,we conclude that smaller firms that are unable to use the hi-erarchical governance form should reengineer activities toscale back the specific investments required to realize theirselected position.-^ Otherwise, their disadvantages aremagnified.

For example, smaller industrial firms might redesigntheir products to deemphasize company-specific tooling andparts needed to fix them. Similarly, smaller-share consumerpackaged goods firms that rely on independent brokersmight use fewer nonprice in-store promotions than theirlarger competitors with company sales forces. Otherwise.the lower capability of their market governance forms toprovide sufficient safeguards will create insufficient invest-ments by their partners and diminish their ability to realizea selected position.

To summarize, smaller market share firms face greaterconstraints in using hierarchical governance but no specialconstraints with market and relational governance (Figure1, #6).

Supply chain resources: Sales force. Selling resourcessuch as a sales force are among the stickiest and least mo-bile of a firm's resources because the sales force's knowl-edge and ties to customers accrue over time. As such, theyare highly path-dependent. A firm's choices are constrainedat one point in time, and in turn, this exerts an influence onits choices in subsequent periods. The principal source ofthis resource effect lies in the cost of switching from onegovernance form to another if and when changed conditionswarrant a switch (Figure I. #6). This can be advantageous ordisadvantageous.

Many product lines begin at a small volume, and thus, afirm may be forced to use an independent representativeforce initially. Weiss and Anderson (1992) show that firmssometimes do not switch to company sales forces, evenwhen they believe a company sales force is better alignedwith the transaction costs drivers. They continue to use theoriginal representative system because of anticipatedswitcbover costs. Terminated representatives may retaliateby disparaging their new products and spreading the notionthe firm is untrustworthy.

In a case study of a leading implantable medical devicefirm, we found that this firm historically used independentrepresentatives to influence surgeons" choices. Althoughsurgeons had a strong infiuence on the purchase, they werenot particularly skilled in the electronics and programmingaspects of these devices. They relied on these representa-tives to be their trusted gateway into the firm's technical

20AItemalively, smaller firms can tum to relational governanceto "simulate" hierarchical governance but without the setup costs.

Governance Value Analysis /141

personnel. This is a high-cost selling approach, with which

these representatives earned hundreds of thousands of dol-lars. In recent years, cosi-conscious health plans and hospi-tals have moved to a committee-oriented approach to buyingthese products. The surgeons' preferences no longer dictatematters. Multifunctional sales teams are more effective atselling to these buying committees.

However, switching to these sales teams is complicatedbecause the representatives object to the erosion of their in-fluence and can credibly threaten to leave and take a con-siderable volume of sales with them. Therefore, many of thelargest firms have not switched to the more appropriate gov-ernance form. This places them at a competitive disadvan-tage to newer entrants that are not faced with theseswitchover costs.

Supply chain resources: Vertical control. A firm's abili-ty to control a large number of downstream partners in asupply chain is a valuable resource because It enables thefirm to implement systemwide changes rapidly in responseto market conditions. Franchise chains can acquire this ver-tical control resource by offering margins greater than theopportunity cost of capital to their downstream channel part-ners. For example. Kaufmann and Lafontaine (1994) docu-ment substantial (hundreds of thousands of dollars peroutlet) ex ante premiums-' earned by McDonald's fran-chisees. By allowing franchisees to earn and keep ex antepremiums, a franchiser gets the franchisee to cede control.Wealth-constrained franchisees are motivated to cooperatebecause of the capital loss they would sustain if terminated.These premiums generate the vertical control resource.

In a twist on the usual advantages that accrue to early en-trants, maintaining vertical control poses a greater challengeto older franchise chains compared with newer entrants.Older chains face a potential erosion of this resource whentheir initial franchisees retire and sell to new owners at mar-ket value. The new franchisees earn smaller (or no) premi-ums because they paid a fully capitalized market price forthe expected stream of earnings. In turn, this makes themless receptive to rapid changes in marketing efforts by thefranchiser. Two strategic directions are implicated. Olderfranchise chains might simply scale back coordinated,chainwide promotional activities. This is unattractive be-cause it undercuts the fundamental value creating rationalebehind franchise chains. A more appealing direction is tobuy back franchises at market value and refranchise them atless than the market clearing level. This should generatequeues for franchises, which is not uncommon tor success-ful chains. Although casual empiricism suggests this occursin some older franchise chains, it cries out for systematicempirical work.

To summarize, channel resources impose path-depen-dent constraints on govemance form choices (Figure 1. #6).

that ex ante premiums are different than the ex post pre-miums featured in Klein and Leftler's(l98I) approach. The formerare "real" profits greater than the cost of capital, whereas the latterarc follow-up period "quasi" profits greater than capital costs thatonly make up for early period negative returns.

Discussion and ConclusionMotivating Further Research

Although the study of many marketing phenomena has beeninfluenced profoundly by TCA, its utility for informing mar-keting strategy analysis has been questioned. We have pre-sented an extension to the core model by developing theinteractions between the creation and claiming of value anddisclosing the influence of positioning and resources on thechoice of governance forms. Using our GVA model, wehave addressed hitherto unanswered questions raised bycritics. For example, we show that firms in the same indus-try use different approaches to bring products to market be-cause of their resource differences. Similarly, we show thatdifferent positioning choices across firms profoundly infiu-ence the design of their supply chain and end-customer gov-ernance forms. We illustrate our conceptual argumentslargely through anecdotal examples because of the lack ofsystematic empirical work on these issues. To encouragesuch work, we have stated several refutable implications inthe form of summary statements throughout the article andtrust that empirical researchers will find them useful.

GVA Biueprint for Strategic Marketing Decisions

In Figure 1, we depict the complete set of links that must beassessed simultaneously in the GVA model. It is Impossiblycomplex to evaluate all of these tensions simultaneously inthe real world. What should a manager make of this com-plexity? One reaction is to reject it in favor of tried-and-trueapproaches to marketing strategy decisions, such as key suc-cess factors; strength, weakness, opportunity, and threatanalysis; and so on. Our goal is to close this endemic gap be-tween contemporary thinking in the literature (research or-thodoxy, as it were) and the approaches that dominatepractice (practice orthodoxy, as it were). To this end, we of-fer a staged decision process that captures the spirit of thecomprehensive GVA model but simplifies it to a consider-able degree.

In Figure 2, we outline the sequence of steps involved. Itis important to note that the starting point is merely a con-venience because the full model treats all four constructs asfully endogenous. In principle, we could start at any point inthe process, loop through the stages, and converge on thesame results.

Step 0 begins the process with an identification ofprospective positions in multiattribute space. We offer no al-gorithm for identifying prospective positions. A usefulheuristic from the RBV literature involves starting from ex-isting resources and positions and seeking out adjacent po-sitions at which these resources might be leveragedadvantageously. This must be tempered with an assessmentof the presence of customers who desire products located atthese adjacent positions.

Regardless of the search process, there should be only afew viable strategic positions that emerge after we apply thecommitment-intensity criterion to purge those positions thatdo not involve difficult-to-re verse aspects. We reiterate thatthis process of identifying prospective strategic positions isnot intended to be the kind of exhaustive search over the en-tire multiattribute space that is featured in optimization

142 / Journal of Marketing, Special Issue 1999

FIGURE 2GVA Blueprint for Marketing Strategy Decision Making

Step 0:Identify activitiesto implementprospectiveposition '

Step 1:Identify impact ofactivities onexchange attributelevels

Step 2:Identify tentativegovernance form tomatch exchangeattributes

Step 5:Verify consistencyof implementedactivities withintended position

Step 3Identify resourceconstraints ongovernance form

Step 4:Modify govemanceform to matchresources

models of product positioning. When the few promising po-sitions have been identified, a rigorous assessment of each isundertaken, as follows.

In Step I, we identify the impact of the prospective po-sition on the three critical exchange attributes (specific in-

vestments, uncertainty, and performance measurement diffi-culties). To do this, the investments and activities required torealize the position must be specified.

In Step 2. we tentatively identify the governance formthat aligns with the investments and activiiies previously

Governance Value Analysis /143

identified. To do this, we apply the three traditional TCArules: Are intended investments sufficiently secured? Isnecessary adaptation facilitated? Is performance ambiguityaccommodated?

In Step 3. we identify the customer, technological, andchannel resources of each partner. To do this, we apply theRBV principle that these resources must be sticky, scarce,and appropriable in the context of the governance form athand.

In Step 4, we modify the tentative choice of governancefrom Step 2 to arrive at a revised choice of govemance form.To do this, we match up the resources identified in Step 3against the tentative governance choice to assess whethermodifications are necessary. Modifications are disclosed byapplying the constraints imposed by resource differences ongovemance form choices (e.g.. strong brand equity disablesrelational forms but facilitates market governance, state-of-the-art technology increases adaptation demands, channelresources impose path-dependent constraints).

In Step 5. we check whether the investments and activi-ties specified in Step I to realize the selected position willbe implemented, given the modified governance form. To dothis, we reapply the three TCA principles used in Step 2about securing investment, facilitating adaptation, and ac-commodating performance ambiguity. If the investments

and activities are not consistent with the positioning, we it-erate back to Step 0.

When all the prospective positions have been evaluat-ed in this fashion, the one with the largest profit is select-ed.^^ It is important to note that this decision model makesno claims about global optimality. Instead, it is a compar-ative exercise. The model is equally useful, if not morevaluable, in making modest comparative judgments, suchas a comparison of the status quo position with a singlenew position.

We also stress that this process is not a one-time exer-cise. As circumstances change, the resources of the firmand its partners, the attributes of the exchange, and thegovernance choice set are likely to change as well. Like-wise, proactive changes in positioning may be motivatedby customer and competitive changes. At any point intime, the process can be reapplied. Using GVA to revisitprevious decisions draws particular attention to the path-dependent constraints that arise from prior strategic choic-es. We trust that managers will find this blueprint useful.

use own-firm profits instead of joint profits, which is per-missible because own-tirm profits will align with joint profitswhen the matches among govemance, positioning, resources, andexchange attributes have occurred.

REFERENCESBakos. Yannos J, and Erik Brynjolfsson (1993), "From Vendors to

Partners: The Role of Information Technology and IncompleteContracts in Buyer-Supplier Relationships," Journal of Organi-zational Cotnfmiing. 3 (3). 301-28.

Brandenburger, Adam M. and Barry J. Nalebuff (1997). 'TheAdded-Value Theory of Business." Strategy and Business. 9(4), 4-6.

Collis. Daniel J. (1991), "A Resource-Based Analysis of GlobalCompelilion," Strategic Management Journal.. 12(1). 49-68.

Day. George S. and Saul Klein (1987). "Cooperative Behavior inVenical Markets: The Influence of Transaction Costs and Com-petitive Strategies," in Review of Marketing, Michael J. Hous-ton, ed. Chicago: American Marketing Association. 39-66.

and Robin Wensley (1988). "Assessing Advantage: AFramework for Diagnosing Competitive Superiority." Journalof Marketing, 52 (April), 1-20,

Dutta. Shantanu and George John (1995), "Combining Lab Exper-iments and Industry Data in Transaction Cost Analysis: TheCase of Competition as a Safeguard." Journal of Law. Econom-ics and Organization. II (I) , 189-204.

Dwyer, F. Roberl and Sejo Oh (1988). "A Transaction Cosi Per-s[>ective on Vertical Contractual Structure and interchannelCompetitive Strategies." Journal of Marketing, 52 (April).21-34.

Ghemawat, Pankaj (1991). Commitment. New York: The FreePress.

Ghoshal. Sumantra and Peter Moran (1996). "Bad for Practice: ACritique of the Transaction Cost Framework." Academy ofManagement Journal, 21 (1). 13-47.

Grossman. Sanford and Oliver Hart (1986). "The Costs and Bene-fits of Ownership: A Theory of Vertical and Lateral Integra-tion." Journal of Political Economy, 94 (August). 691-719.

Helper, Susan and David Levine (1992). "Long-Term Supplier Re-lations and Product-Market Structure," Journal of Law. Eco-nomics and Organization, 8 (3), 561-81.

Hunt. Shelby D. and Robert M. Morgan (1995). "The ComparativeAdvantage Theory of Competition." Journal of Marketing, 59(!). I-I5.

Kaufmann, Patrick J, and Francine Lafontaine. (1994), "Costs ofControl: The Source of Economic Rents tor McDonald's Fran-chisees," Journal of Law and Economics, 32 (October).417-53.

Klein, Benjamin and Keith J. Leftler (1981). 'The Role of MarketForces in Assuring Contractual Performance." Journal of Polit-ical Economy, 89 (4), 615-41.

Masten. Scott E. and Edward A. Syndcr (1993). "United StatesVersus United Shoe Machinery Corporation: On the Merils."Journal of Law and Economics, 36 (April). 33-70.

Muris, Timothy J , David Scheffman. and Pablo T Spiller (1992)."Strategy and Transaction Costs: The Organization of Distribu-tion in the Carbonated Soft Drink Industry." Journal of Eco-nomic.^ and Management Strategy. 1 (1). 85-128.

Nickerson. Jack A, (1997), 'Toward An Economizing Model ofStrategy," working paper. Olin School of Business. WashingtonUniversity.

Noordewier, Thomas G.. George John, and John R, Nevin (1990),"Pertormance Outcomes of Purchasing Arrangements in Indus-trial Buyer-Vendor Relationships," Journal of Marketing, 54(October). 80-93.

Peteraf. Margaret A. (1993). "The Cornerstones of Competition: AResource-Based View." Strategic Management Journal, 14 (3),179-91.

Poppo, Laura and Todd Zenger (1998), 'Testing Alternative Theo-ries of the Firm: Transaction Cost, Know I edge-Based, andMeasurement Explanations for Make-or-Buy Decisions in In-formation Services," Strategic Management Journal. 19 (9).853-77.

Porter. Michael E. (1985), Competitive Strategy. New York: TheFree Press.

(1996). "What is Strategy?" Harvard Business Review, 74(November/December). 61-78.

144 / Journal of Marketing, Special Issue 1999

Rindlleisch. Aric and Jan B, Heide (1997). "Transaction CostAnalysis: Past, Present and Future Applications," Journal ofMarketing, 61 (October). 30-54.

Silverman, Brian S. (I99S). 'Technological Resources and the Di-rection of Corporate Di versification: Toward an Integration of theResource-Based View and Transaction Cost Economics," Work-ing Paper #99-019, Cambridge, MA: Harvard Business School.

Simon, Carol J. and Mary W. Sullivan (1993), 'The Measurementand Determinants of Brand Equity: A Financial Approach,"Marketing Science, 12(1), 28-52.

Varian, Hal R. and Carl Shapiro (1998), Information Rules: AStrategic Guide to the Network Economy. Boston. MA: HarvardBusiness School Press.

Webster. Frederick E. (1992), "The Changing Role of Marketing inthe Corporation," Journal of Marketing. 56 (October), 1-17.

Weiss. Allen M, and Erin Anderson (1992). "Convening from In-dependent to Employee Sales Forces: The Role of PerceivedSwitching Costs," Jottrnal of Marketing Research. 29 (Febru-ary). 101-15.

Wernerfelt. Birger (1994), "An Efficiency Criterion for MarketingDesign." Journal of Marketing Research, 31 (November).462-70.

(1997). "On the Nature and Scope of the Firm: An Adjust-ment-Cost Theory of the Firm." Journal of Business, 70 (4),489-514.

Whitney. Daniel E. (1992), '"State of the Art in Japanese CADMethodologies for Mechanical Products-Industrial Practice andUniversity Research." Office of Naval Research Asian OfficeScientific Information Bulletin, 17(1). 89-1I I .

Williamson. Oliver E. (1989), "Transaction Cost Economics." inHandbook of Industrial Organization. Richard Schmalenseeand Robert Willig, eds. Amsterdam: North-Holland, 135-82.

(1991), "Strategizing, Economizing and Economic Orga-nization." Strategic Management Journal. 12 (Special Issue).75-94.

(1996), The Mechanisms of Govemance. New York: Ox-ford University Press.

Wilson. Lynn O., Allen M. Weiss, and George John (1990), "Un-bundling of Industrial Systems." Journal of Marketing Re-search, 30 (May), 123-38.

Zajac. Edward J. and Cyrus P. Olsen (1993). "From TransactionCost to Transaction Value Analysis: Implications for the Studyof Interorganizational Strategies," Journal of ManagementStudies, 30 {I), 1 3 1 - 4 5 .

Governance Value Analysis /145