Bucklin, Louis P; Sengupta, Sanjit. Organizing successful co-marketing alliances. Journal of Marketing v57, n2 (Apr 1993):32-46. Pub type: Studies; Experimental. Type D 3 AB to see abstract. Copyright American Marketing Association 1993 Recognizing the increasing customer need to improve linkages across different computer systems, IBM Corporation and Apple Computer announced a strategic alliance in July 1991 to share previously proprietary technologies (Bertrand 1992). By this move IBM expands its influence into certain classes of personal computers, which it would be unlikely to achieve through internal development, and Apple Computer obtains improved connectivity to mainstream corporate computing. Though considerably larger than most, this partnership is a fascinating example of what might be termed "complementary product" or "co-marketing" alliances. Co-marketing alliances are a form of working partnership, defined by Anderson and Narus (1990) as the "...mutual recognition and understanding that the success of each firm depends in part on the other firm...." They are contractual relationships undertaken by firms whose respective products are complements in the marketplace. They are intended to amplify and/or build user awareness of benefits derived from these complementarities. They involve coordination among the partners in one or more aspect of marketing and may extend into research, product development, and even production. Unlike buyer-seller or manufacturer-distributor partnerships, co-marketing alliances are lateral relationships between firms at the same level in the value-added chain and represent a form of "symbiotic marketing" (Adler 1966; Vardarajan and Rajaratnam 1986). Vertical relationships receive continued attention in the literature (Frazier, Spekman, and O'Neil 1988; Jackson 1985; Johnston and Lawrence 1988), but relatively little has been written on lateral relationships such as these even though they may now comprise some of the most intriguing and active elements of marketing strategy. This organizational form leverages a firm's unique skills with the specialized resources of its partners to create a more potent force in the marketplace. The results have sometimes been spectacular. Microsoft used its alliance with IBM for the MS DOS operating system in the early 1980s to catapult itself into the position of the dominant PC software firm (Brandt 1989). Apple Computer and Adobe Systems worked together in 1984 to create the tools of desktop publishing, a major market whose evolution was greatly accelerated by this partnership (Clark 1989). Alliances such as these are common in high technology industries in which even the largest firms cannot hope to maintain cutting-edge positions across all technologies of interest to their end users. Large firms such as IBM were once inclined to develop their product lines internally. Now, they aggressively seek partnerships with numerous firms to ensure that their core products are linked advantageously to advance technologies created elsewhere (PC Week 1987). Co-marketing alliances may have become essential to the introduction of radical technologies facing competition from well-entrenched product standards. Pen-based computers entering the market in the early 1990s do so under the aegis of competing consortia of firms, each contributing its unique expertise in terms of hardware, operating systems, and application software (Furger 1991). Only by the linking of multiple firms' resources can new systems be developed with sufficient breadth and sophistication to persuade end users to abandon current investments and upgrade to new technology. Despite their potential contribution, co-marketing alliances pose significant management challenges. The potential for serious conflict is always present as partners often compete with each other in other product lines and, on occasion, in those directly covered by the co-marketing agreement. The potential for opportunism is high as partners may use the alliance only as a means to gain market position at the expense of a partner or to build technological skills from exposure to the partner's intellectual property. We report a study of ongoing co-marketing alliances to aid our understanding of how to manage this new organizational form more effectively. Drawing on the literature from interorganizational exchange (e.g., Cook 1977; Pfeffer and Salancik 1978) and transaction cost economics (Williamson 1975), we formulate a framework for co-marketing alliance effectiveness, establish hypotheses, and define diagnostic measures. We then report primary data collected on these measures from a sample of alliances in the computer and semiconductor industries, test our hypotheses, and review and evaluate our results. We conclude with a discussion of the research and managerial implications of our findings. FRAMEWORK The central theoretical perspective in past studies of interfirm relations is that of interorganizational exchange behavior (Aldrich 1979; Benson 1975; Cook 1977; Frazier 1983; Pfeffer and Salancik 1978). This view holds that given functional specialization and a scarcity of resources, organizations seek to reduce environmental uncertainty by exchanging resources for mutual benefit. ALLIANCE SUCCESS To explore this view, we need a measure of mutual benefit, that is, the degree of alliance success. Quantitative and qualitative performance indicators were considered, but exploratory interviews with 20 co-marketing alliance managers indicated that many benefits are difficult to track quantitatively. In addition to joint marketing efforts by alliance partners, individual efforts and general economic conditions may result in increased sales of products. Tracking what portion of this incremental business is due purely to the alliance is difficult, if not impossible, to accomplish. A qualitative measure of performance, the perceived effectiveness of the relationship, was adopted as an indicator of success. This measure has been developed in organization theory (Van De Ven 1976) and applied to interorganizational relationship dyads (Ruekert and Walker 1987; Van De Ven and Ferry 1980). Because mutual performance is the criterion at issue, perceived effectiveness is defined to be the extent to which both firms are committed to the alliance and find it to be productive and worthwhile. One of the consequences of exchange among organizations is the emergence of resource dependencies among partners. Resource dependence gives rise to power, that is, the alteration of one party's behavior by the other (Gaski 1984). Unequal resource endowments therefore result in power imbalances (Emerson 1962). In an extension of the role of power in exchange, Cook (1977, p. 67) holds that "there is a tendency for actors to prefer exchanges with equally powerful actors because there are fewer costs attached to the exchange process." Such costs arise as the consequence of difficulties in reaching agreement and, echoing Williamson (1975), the potential for exploitation that remains when such unions are formed. According to Cook, the potential for exploitation is greatest where there is imbalance between the participants. Balance is therefore sought by the parties as the means to ensure that neither has incentive to exploit the other. Combining Cook's logic with observations from exploratory interviews, we posit that performance of an alliance ought be dependent on partners' ability to mitigate any power imbalance among them. This notion is central in our conceptual framework under the broader project management factor. Project management involves the balancing of dependencies among the members of the partnership. Figure 1 places this view in context with two other and more traditional factors affecting interorganizational performance, project payoff and partner match. (Figure 1 omitted) Project payoff defines alliance partner ex ante views about market opportunity and cost. Partner match reflects the capability of alliance firms to cooperate and work with each other. PROJECT MANAGEMENT The presence of power imbalance in an exchange relationship creates the potential for conflict. Muller's (1970) expectation that organizations with superior power will act to exploit that power aptly identifies the core of the issue. If dependencies are out of balance in a relationship, the weaker party will take precautions to limit its vulnerability. In the present context, this effort could take the form of competing alliances, subtle efforts to diminish the role of its partner with customers, or failure to employ all of the resources required. Recognizing the potential for this behavior, the more powerful party may similarly be loath to put forth the maximum effort required by the project. The best outcome that can emerge from this type of interaction is a failure to take full advantage of the joint market opportunity. The worst is that the fears lead to a self-fulfilling prophecy and irresolvable conflict from counterproductive efforts, reduced participation, or even outright sabotage. Power imbalance therefore is detrimental to alliance effectiveness. Another dimension of project management is that members of an alliance are likely to be sensitive to the contributions made by their partners. Ouchi (1980) observed the difficulties encountered in the measurement of equity by clan organizations because of ambiguous performance measures for individual members. In these organizations, participant evaluation took place through the subtle reading of signals, such as the effort and time allocated by members to work activities. To the extent that alliances are analogous to clans, participants will look intensely at the resource contributions made by partner firms as a factor affecting their continuing willingness to participate. In the exploratory interviews, executives showed sensitivity to the adequacy of the managerial resources assigned to projects by the partners. Managerial imbalance, or the failure to allocate the expected managerial talent, in numbers or stature, created a concern for equity in "pulling one's load." In an interesting echo of this observation, Achrol, Scheer, and Stern (1990) noted that difference in perceived position or status among the managerial levels at which interaction occurs among alliance partners could lead to cultural and political conflict. Doz (1988) found that differences in the locus of management among partners in technology alliances could lead to communication difficulties. From these arguments, we derive two project management hypotheses. H sub 1 : Power imbalance in a co-marketing alliance is related negatively to the effectiveness of the relationship. H sub 2 : Managerial imbalance in a co-marketing alliance is related negatively to the effectiveness of the relation-ship. It also follows from this discussion that alliance effectiveness ought be affected by the extent of conflict in the relationship. The inability of alliance managers to limit conflict, regardless of its source, is likely to reflect ineffective leadership and exacerbate any power imbalances. Some conflict, of course, is inherent in virtually all exchange relationships (Mallen 1963) and may even contribute to a positive outcome (Assael 1968). Many empirical studies indicate a negative relationship between conflict and perceptual outcomes such as satisfaction (Gaski 1984), whereas others show that conflict resolution results in positive outcomes (Ruekert and Walker 1987). This apparently nonmonotonic relationship can be explained by making a distinction between functional and dysfunctional conflict (Anderson and Narus 1990). Though functional conflict may enhance performance, dysfunctional conflict may attenuate it. We offer the following hypothesis about dysfunctional conflict. H sub 3 : The less the conflict between firms in a co-marketing alliance, the greater the effectiveness of the relationship. PROJECT PAYOFF Interorganizational exchange theory has pointed out that organizations undertake cooperative ventures after careful consideration of costs and returns related to resource deployment (Benson 1975; Schermerhorn 1975). Frazier (1983) put forth a framework of exchange in which expected rewards and required investment in a relationship determined implementation and future outcomes. More specifically, Spekman and Sawhney (1990) note that the motivation for firms to enter into alliances is to obtain strategic advantage such as access to new markets and/or technical information, enhanced product value, and enhanced market reputations. To the extent that the inherent market opportunity of some ventures is greater than that of others, superior results should follow. Resources required to put the co-marketing alliance on stream are a second dimension of project payoff. In the case of joint marketing activities, expenditures include such items as salesforce training and deployment, promotional literature, and advertisements. If joint product development is also required, additional resources must be allocated to develop the necessary technologies. To the extent that the development process stretches over several years, additional expense must be incurred. The time dimension also expands risk because of the potential for changes in market needs. If an alliance must put significant resources to work over time, the potential for return on such investment is attenuated. In this light, project payoff is defined as the strategic value of the alliance net of development cost. The view that alliances with well-identified market opportunities and well-defined costs are more likely to perform well gives rise to the following hypothesis. H sub 4 : The higher the project payoff from a co-marketing alliance, the greater the effectiveness of the relationship. PARTNER MATCH Partner match calls for the creation of alliances in which the chosen partners are similar in management style and company culture. Considerations such as domain similarity and goal compatibility have been found to enhance the effectiveness of interorganizational dyads (Ruekert and Walker 1987; Van De Ven and Ferry 1980). The concept of organizational compatibility has been proposed by Achrol, Scheer, and Stern (1990) to include strategic and cultural compatibilities. Cooperation in alliance relationships is subject to the threat of opportunistic behavior by one or more partners (Spekman and Sawhney 1990; Williamson 1981). If trust and commitment develop between the partners, they counterbalance the potential for adverse forms of behavior. Dwyer, Schurr, and Oh (1987) point out the importance of commitment in their process model of buyer-seller relationships. In their view, commitment is an implicit or explicit pledge of relational continuity between exchange partners. Heide and John (1990) found a positive association between the historical length of an alliance relationship and expected continuity of future interaction. Therefore, a long and stable history of prior business relations can build trust and commitment between partners. These considerations lead to two partner match constructs--organizational compatibility and prior history of business relations. Organizational compatibility reflects complementarity in goals and objectives, as well as similarity in operating philosophies and corporate cultures. Prior history expresses the extent of prior business relationships, a period that would enable potential partners to judge their compatibilities and develop necessary personal relationships to augment their general similarities. Hence: H sub 5 : The greater the organizational compatibility between the firms in a co-marketing alliance, the greater the effectiveness of the relationship. H sub 6 : The longer and more stable the prior history of business relations between partners in a co-marketing alliance, the greater the effectiveness of the relationship. NONCONTROLLABLE VARIABLES Apart from the three general dimensions affecting alliance effectiveness, as shown in Figure 1, certain conditions outside the orbit of management control may also influence performance. In particular, two factors, age and dynamism in the marketplace, have been shown to be important in previous research on interorganizational relationships (Heide and John 1988, 1990; Ruekert and Walker 1987). Levinthal and Fichman (1988) examined the duration of auditor-client interorganizational relationships. They found the failure rate of these relationships to be greatest during an initial, and paradoxically named, "honeymoon" period. After this period, the rate of failure declined continuously with time. "Longer-standing relationships are less vulnerable to threats to their persistence" (Levinthal and Fichman 1988, p. 366). In parallel fashion, we argue that alliances that have already withstood some test of time are more likely to be successful. Early failure would have weeded out alliances that had more limited prospects. Therefore: H sub 7 : The greater the age of a co-marketing alliance, the greater the effectiveness of the relationship. The rate of technological change in the environment is another factor that may affect alliance success. Rapid technological change may render present assets and skills obsolete and, hence, ought have a negative impact on the effectiveness of the alliance. However, firms may enter into alliances because they enable partners to share the development risks of rapidly changing technologies. Hence, the alliance may be a more effective organizational form under conditions of high technological uncertainty. If so, the relationship between the rate of technological change and effectiveness would be positive. We do not offer any hypothesis for this construct, but test the relationship empirically. DETERMINANTS OF POWER IMBALANCE As developed in the preceding discussion, a central dimension of success in a co-marketing alliance is a balance of power between partners. Though this balance may have an impact on performance, the managerial implications are limiting. Given the arguments of Cook (1977), the only clear prescription is that suitors in search of co-marketing partners limit their choice to firms roughly similar in financial resources and market presence. However, such a guideline would severely limit the set of available opportunities. More critically, it would deny the potential for otherwise desirable relationships between mature firms rich in production and marketing skills and startup firms with innovative technologies. What mechanisms are available to help partners whose resource endowments and respective power positions are unbalanced to build fruitful alliances? Considerable aid may come from the law of contract. With contracts, firms have the opportunity to design desired patterns of partner behavior and to extract penalties from failures to perform. Contracts represent historic methods for the resolution of disputes. Moreover, as in relational contracting (Macneil 1980), they provide the opportunity and basis for the partners to renegotiate the agreement in order to adjust to changes in underlying market conditions. Initial interviews with alliance managers provided mixed views on this point. One executive favorably noted how he had placed incentives for cooperation into his agreement with a much larger partner by requiring the latter to purchase a significant stock holding in his company. Conversely, the marketing manager of another company held that contracts were largely ineffectual for his turbulent business, surmising that a contract would constrain actions necessary to adjust to shifts in underlying market conditions. To explore the opportunity for alliance partners to balance their relationships through contractual governance, we identify four characteristics of co-marketing agreements: formality, exit barriers, exclusivity, and financial incentives. They are conceptually similar to the general dimensions of contract structure identified by Stinchcombe (1985)--for example, standard operating procedures, conflict resolution, legitimate authority, and an incentive system. The handling of these dimensions by the partners may redress differences in underlying resource endowments and consequent power imbalances. Formality reflects the degree to which an agreement is subject to highly defined conditions. At one extreme, a relationship could be based on a simple letter of intent. At the other extreme, the agreement may be a long and complex legal document. If the parties have dissimilar endowments, an extensive contract may serve to "level the playing field" by defining rights and powers of redress for the weaker party. By these means, formality may redefine the power relations between parties to create greater balance. One way of increasing control over resources is by forsaking alternative sources and further increasing interdependence between partners (Pfeffer and Salancik 1978). Exit barriers can be raised by designing long-term agreements or by incorporating penalties, such as the forfeit of intellectual property rights involved in the project, for early termination of the agreement. Exclusivity constraints are another method the parties can employ to redress power disparities. The stronger party going into the co-marketing alliance could, for example, be barred from establishing competitive alliances while the weaker party has no such restriction. Grants of exclusivity in vertical marketing systems have a somewhat similar role (Stern and El-Ansary 1988). Supplier provision of exclusive rights to its distributor enhances the latter's position by restricting the options of the former. Anderson and Weitz (1992, p. 19) have interpreted exit barriers and exclusivity as pledges of commitment in channels: "...commitment to a relationship entails a desire to develop a stable relationship, a willingness to make short-term sacrifices to maintain the relationship, and a confidence in the stability of the relationship." The provision of financial incentives is the final component. Equity investments are one option and direct monetary payments are another--for example, a computer firm paying a one-time fee to a software firm to adapt its software to the former's platform. Commissions on joint sales are still another example of direct monetary payments. Incentives may be uni-or bidirectional. From this discussion, the following hypothesis is developed. H sub 8 : Power imbalances in a co-marketing relationship can be reduced by increasing the level of contractual governance (formality, exit barriers, exclusivity, and financial incentives), either individually or jointly. As in the first stage of the framework, other variables outside the direct control of potential partners may also affect power imbalance. Each project, for example, can be expected to involve certain transaction costs (Williamson 1975). When high initial transaction-specific investments must be made in an alliance, the parties are vulnerable to opportunism from changes in market conditions, which in turn may upset current operations and lead to imbalance in the relationship. In addition, when uncertainty about project performance is high or when contacts between the parties are frequent, any balance can be easily disturbed by opportunistic behavior. High transaction costs, in other words, may upset other efforts to create balance. This point leads to the next hypothesis. H sub 9 : The greater the expected transaction-specific investment, expected frequency, and expected uncertainty in a co-marketing alliance, the greater the power imbalance in the relationship. Last, linkages between transaction costs and contractual conditions may also be present. Williamson (1985, p. 185) notes that "...market governance poses hazards when nontrivial transaction-specific assets are placed at risk." The need for some nonmarket form of governance increases as the level of transaction-specific risk increases. Contractual terms may have the greatest impact when transaction-specific investments are high. Joskow's (1987) research on long-term coal contracts between coal suppliers and electric utility customers is intriguing in this regard. He found that where local conditions reduced the availability of buyers and suppliers to one coal supplier and one utility buyer, the parties significantly increased the time duration of the supply contracts. Under these circumstances, the parties had no alternative uses for their investments in mining or electricity-generating assets. This example shows that parties to a transaction will make greater effort to design contracts that serve their respective interests when transaction-specific investments are high. For alliances that call for significant transaction-specific investments to be made, contractual terms effectively safeguard the interests of both parties. Hence, we hypothesize the following interaction effect. H sub 10 : When transaction-specific investments are high, power imbalances in a co-marketing relationship can be reduced by increasing the level of contractual governance: formality, exit barriers, exclusivity, and financial incentives. The second stage of the framework is shown in Figure 2. (Figure 2 omitted) METHOD The co-marketing alliance is the unit of organizational analysis for the study. The dependent variable, perceived effectiveness, refers to the performance of the alliance from the perspectives of both partners. Thus, perceived effectiveness is the extent to which both firms find the alliance productive and worthwhile. To draw a semantic distinction between partners in the alliance, we designate the organization from which data were collected as the focal firm and the other firm in the alliance as the partner. MEASURES For independent variables, the domain of various constructs was specified and a sample of items was generated for each construct (Churchill 1979). These items were reviewed and edited by a panel of marketing researchers. Next, a questionnaire was designed and pretested on four managers similar in profile to the target respondents. The revised questionnaire was used to collect data. In this process, 23 unidimensional constructs were conceptualized. The first 17 were measured by multiple-item Likert scales. The last six were measured by single-item scales. Nine of the multiple-item scales were motivated by previous studies in marketing and organization theory. Reliability scores (Nunnally 1978) for scales motivated by literature are reported in Table 1 in the form of Cronbach's alpha. (Table 1 omitted) New constructs are listed with their Cronbach's alpha scores in Table 2. (Table 2 omitted) The six single-item constructs are described in Table 3. (Table 3 omitted) Items in Tables 1, 2, and 3 were answered on a 5-point Likert scale, except where indicated. Of the single-item measures, age of the alliance was measured in years, so a single item is appropriate. The rate of technological change in the focal product, the rate of technological change in the complementary product, and the development cost to the partner firm were not used by themselves but as inputs to composite constructs. Expected frequency of interaction and expected uncertainty in measuring partner performance were constrained to single-item scales because of constraints on questionnaire length. Churchill and Peter (1984) report that Cronbach's alpha is .5 or greater for 85% of scales used in marketing studies and is .7 or greater for 69%. In Tables 1 and 2, all of the scales have alpha greater than .5 with the exception of the one for exit barriers. A better measure may be needed for that construct. To obtain composite scores on each unidimensional construct in Tables 1 and 2, we standardized individual item scores and summed them over items comprising the construct. Eleven unidimensional constructs in Tables 1 through 3 were used by themselves in the data analyses. From the remaining 12 of 23 unidimensional constructs, five multidimensional constructs were derived by combining the perceptions of the focal firm with attributions by the focal firm about the partner firm. In all, 16 constructs were used in our data analyses. For the alliance as a unit, managerial imbalance, expected transaction-specific investment, the rate of technological change, strategic value, and development cost were made operational as the sum of focal firm scores and its attributions about the partner firm. Project payoff was computed as the difference between strategic value and development cost. Power imbalance was defined as the absolute difference between the power of the focal firm and the attributed power of the partner firm. Factor analysis was performed on all the individual items to establish content validity (Churchill 1979). If individual items load onto factors corresponding to the conceptualized constructs, we have greater confidence in the measures. Factor analysis with varimax rotation was carried out and resulted in satisfactory correspondence between the factors and the constructs. To test the convergent validity of the focal firm responses, we used a second questionnaire to collect data on 16 of 23 unidimensional constructs from partner firms wherever they could be identified. Of the 40 so identified by focal firms, 29 completed this questionnaire. The resulting sample was too small to establish convergent and discriminant validity by using a formal multitrait-multimeasurement (MTMM) model. However, raw correlations across the firms in the dyad showed significant positive correlations for seven of the 16 constructs, thus providing some evidence for convergent validity (Bucklin and Sengupta 1992). DATA COLLECTION The key informant technique (Seidler 1974) was used to collect data. The exploratory study found that alliances were typically directed by a middle-level executive with day-to-day operational management responsibilities. Such boundary-spanning persons were knowledgeable about the roles of both the focal and partner firms and thus well qualified to be key informants. The American Electronics Association Directory (1988) was used to identify target firms in the computer and semiconductor industries. Data collection was limited to these industries to maintain homogeneity within the sample and in anticipation that participation in co-marketing alliances would be greater than in other industries. For this reason, companies with fewer than 40 employees and manufacturers of IBM-compatible personal computers were eliminated because few were likely to have engaged in co-marketing alliances. The resulting sample frame comprised a total of 493 firms. Software firms accounted for 20% of the sample frame and hardware firms for the remaining 80%. A personal letter and the eight-page questionnaire were sent to a senior executive in each selected firm, typically the vice president of marketing. The letter requested the addressee to forward the questionnaire (if necessary) to an executive in the firm who was in charge of a co-marketing alliance that met the following criteria: (1) the alliance involved joint marketing of two complementary products from two different firms, (2) a written agreement governed the alliance, and (3) products from the alliance were commercially available for at least 90 days prior to receipt of the letter. These criteria were designed to exclude from the sample ad hoc verbal agreements that were far from commercialization. They also ruled out joint ventures, separate from their parent organizations, to minimize heterogeneity in the sample. A summary of the survey results was offered to firms as an incentive to participate in the study. After extensive mail, telephone, and fax followup, usable responses representing 98 alliances were obtained. In followup telephone calls, the most-cited reason for not participating was the absence of a co-marketing alliance that met the survey criteria. Hence, the real nonresponse rate for eligible firms is significantly lower than is otherwise apparent. As can be seen in Table 4, responding firms represented a broad spectrum of the computer and semiconductor industries. (Table 4 omitted) Key informants were primarily middle-level managers who on average had worked 5 years in the same firm. They had worked on the co-marketing alliance for which they provided data about 1.6 years on average. They had an average of 5 years' experience in managing all kinds of interfirm partnerships. Of the 70 firms participating in the survey, nine provided data on multiple alliances. However, in each such case, a different key informant provided data on each separate alliance. RESULTS Ordinary least squares (OLS) estimation was used to test the hypotheses derived from Figure 1. (Figure 1 omitted) LISREL was not used for estimation because it assumes the constructs are measured with reflective scales (Bagozzi and Fornell 1982). Five of our 16 constructs are multidimensional, made operational as the sum or difference of underlying unidimensional constructs. These are formative, not reflective, scales and cannot be accommodated in a LISREL model. PERCEIVED EFFECTIVENESS: STAGE 1 OLS results pertaining to perceived effectiveness of the alliance are reported in Table 5. (Table 5 omitted) Because all independent variables were standardized, each variable's coefficient provides a measure of its predictive power. Taken together, independent variables explain 50% of the variance in perceived effectiveness. Project management constructs, power and managerial imbalance, reflecting H sub 1 and H sub 2 are supported at the .05 level of significance or better. As expected, both are related negatively to perceived effectiveness. Additionally, H sub 3 on the negative effect of conflict holds strongly. We note that Anderson and Weitz (1992) found a similar association between conflict and alliance commitment. Project payoff, H sub 4 , tests the role of economic opportunity in determining alliance effectiveness. Again, as anticipated, payoff is related positively to perceived effectiveness and is the second strongest variable in the group. Partner match, the third set of variables in the table, tests H sub 5 and H sub 6 for organizational compatibility and the prior history of the business relationship. These hypotheses are confirmed strongly, with prior history between partners being the strongest predictor in the set. H sub 7 , on the age of the alliance, is supported in the expected direction though at the .10 level of significance. The rate of technological change, used without hypothesis, provides a strong positive result. This finding suggests that co-marketing alliances provide superior strength to survive tempestuous circumstances. POWER IMBALANCE: STAGE 2 Error residuals from separate OLS regressions with perceived effectiveness (stage 1) and power imbalance (stage 2) as dependent variables were found to be uncorrelated. Therefore, independent OLS estimation is appropriate for each (Pindyck and Rubinfeld 1981) and we can use this technique for stage 2. Three stage 2 regressions are shown in Table 6. (Table 6 omitted) Regression 1 gives the results between power imbalance and the four individual dimensions of contract governance (formality, exit barriers, exclusivity, and financial incentives), as well as the three transaction cost variables (expected transaction-specific investment, expected frequency, and expected uncertainty). Surprisingly, regression 1 shows that the four dimensions of contract governance, taken individually, are ineffective in reducing power imbalance. Transaction costs, however, have a stronger influence. Expected transaction-specific investment and expected frequency of interaction have robust, positive associations with power imbalance (H sub 9 ), but expected uncertainty is unrelated. In regression 2, significant interactions among the variables are reported and suggest that the interrelationships between contractual governance and power imbalance are more complex. The combination formality X exit barriers X exclusivity shows a strong negative influence indicating that these three elements of contract governance collectively are effective in reducing power imbalance (H sub 8 ). Expected transaction-specific investment and expected frequency retain significant, though larger, magnitudes (H sub 9 ). It is worth noting that the percentage of variance explained for power imbalance doubles when the two interaction terms are introduced in regression 2. Of particular interest, and in support of H sub 10 is the strong negative interaction effect of formality X exclusivity X incentives X expected transaction-specific investment on power imbalance in regression 2. We interpret this result to show that where these elements of contractual governance are applied under conditions of high transaction-specific investments, the perceived level of imbalance among the partners is sharply reduced. In regression 3, we drop the nonsignificant terms from regression 2, getting stable results for the remaining interaction terms and transaction costs. DISCUSSION Empirical results from our study provide support for previously untested concepts from interorganizational exchange theory and a basis for understanding some of the factors that affect the success of a new form of marketing alliance, one created for the joint management of complementary products controlled by separate firms. It similarly confirms perceptions by executives of the importance of providing a level playing field among the members of this type of partnership. A major tenet of interorganizational exchange theory is that organizations seek external relationships that are balanced (Cook 1977; Spekman and Sawhney 1990) in terms of power. The data suggest that imbalances in power and in the managerial resources that each partner provides are significant drawbacks to alliance operations and, as organizational theorists predicted, have an important role in limiting alliance success. Consequently, a key aspect of project management is understanding the strength of the resource base of potential partners, the managerial contributions that each makes, and how disparity in these conditions might lead to conflict. As in other studies, our data also show that alliance success is sensitive to dysfunctional conflict. The need to manage and contain conflict is further emphasized. Significant, and reassuring, is the finding that well-planned projects with high payouts in relation to cost are most likely to be successful. Mediocre co-marketing alliances are probably not worth the effort. It clearly pays to develop prior relationships with prospective partners before engaging in formal alliances to ensure effective working relations. Similarly, the need for compatibility in terms of partner culture, operations, goals, and objectives is reinforced. Finally, we find that alliances tend to be more successful in turbulent environments. The implication is that alliances provide a superior vehicle for gaining access to new complementary products or technologies without all the risks of internal development. We believe our findings from the second-stage analysis of the determinants of power imbalance extend the horizon of transaction cost economics into lateral dimensions of interfirm relationships. Perceptions of imbalance are accentuated by the presence of transaction-specific investments and high frequency of association between partners. The implication is that co-marketing alliances associated with conditions of high transaction cost are more difficult to manage for success than others not so burdened. The analysis also suggests that problems associated with high transaction costs can be offset through contractual governance procedures. The strongest evidence is found in the interaction among the four defined elements of contractual governance: formality, exit barriers, exclusivity, and incentives. The first three, interacting together, indicate that greater formality, longer lived relationships, and exclusivity in the arrangements may help reduce damaging perceptions of imbalance among the partners. The interaction among formality, exit barriers, financial incentives, and transaction-specific investments further extends empirical findings in transaction cost economics. The significant relationship here implies that contractual governance in co-marketing relationships is especially effective in reducing power imbalance when transaction-specific investments are high. In sum, from a theoretical point of view, our findings extend empirical knowledge in two areas of importance to marketing--interorganizational exchange theory and transaction cost economics. In the case of the former, the importance of power balance in interfirm relations is confirmed in our field research. From transaction cost economics, the importance of transaction-specific investments as a factor affecting success is extended to lateral relationships. MANAGERIAL IMPLICATIONS From a management point of view, our research provides insights on appropriate strategies for firms seeking to form and manage co-marketing alliances. We argue that the management of this new type of inter-organizational alliance can be conceived in terms of three major dimensions. Findings suggest that co-marketing alliances prosper when projects have been well selected, partners chosen carefully, and relationships structured toward balance. The importance of project selection, in terms of both market opportunity and resources required, implies that the process of review of potential alliances must be conducted as carefully as that of new product development. Market objectives must be clearly defined, the value added from improving product complementarities well understood from the end user's perspective, and resource requirements accurately anticipated--not only from the perspective of the focal firm manager, but also from that of the partner because the success of the alliance depends on the results of all parties involved. If clear understanding of these dimensions from the partner's perspective has not been developed at the start of the agreement, project selection work has not been performed adequately. The selection of an appropriate partner is no less important. The findings indicate that compatibility of the partners is critical to alliance success. Hence, managers who perceive that co-marketing alliances are of benefit would be well advised to seek relationships with a variety of potential partners on an exploratory basis. Alternatively, if firms have not worked closely prior to an alliance, frequent interactions between managers at all levels of the firms to explore he mutuality of their interests and their working styles should be encouraged. Such firms would also be wise to initiate relationships with a series of minor projects before attempting to start a project in which major resources must be invested. It will be interesting to follow the recently signed alliance between IBM and Apple Computer (Zachary and Hooper 1991), companies that not only have not worked together, but have been bitter competitors with totally different business cultures. In selecting prospective partners, firms should build relationships with those having roughly similar endowments in terms of resources, market positions, and competitive capabilities. Our results suggest that failure to build relationships in which the partners perceive and treat each other as equals will lead to sub-optimal results. This finding calls into question the wisdom of smaller firms that desperately seek alliances with larger, better established organizations, hoping that such linkages will leverage them into market prominence. Though some extraordinary successes have been achieved, such endeavors are inherently risky. If market prospects are nevertheless inviting for unbalanced partners, our findings suggest that some of the problems can be mitigated by carefully drawing up a contract that would help restore balance. If a partnership requires large transaction-specific investments in the face of unbalanced power bases, managers must seek clear understanding of points of possible future contention and ensure that contract provisions address such problems. A final implication can be drawn from the finding that co-marketing alliances are most likely to prosper in markets subject to considerable technological uncertainty. This result may well provide the rationale for the formation of co-marketing alliances and the reason for their apparent proliferation in the computer and semiconductor industries. In turbulent high technology markets, co-marketing alliances enable firms to provide connectivity across complementary products of importance to consumers without requiring firms to move beyond their core competencies. LIMITATIONS Though our findings are highly suggestive for both the theory and practice of co-marketing alliances, and echo findings of prior theory and research, we must raise certain cautions. The research results are based on perceptions of performance as opposed to hard, quantitative performance measures. Though data of the latter type are unlikely to be available for co-marketing alliances, replication prior to acceptance is prudent given the importance of the findings, especially the findings on the determinants of imbalance. We found contractual governance to be effective, but recent research indicates that other balancing mechanisms are available, such as bonding with customers (Heide and John 1988), the use of pledges (Anderson and Weitz 1992), and the development of relationship norms (Heide and John 1992). Managerial actions based on study findings should proceed with care. At a more detailed level, there are some measurement problems, particularly with the level of reliability of some scales in Tables 1 and 2. These problems may be the consequence of having too few items for some constructs. However, additional work is needed to improve reliability and validity for constructs such as expected transaction-specific investment of partner firm, exit barriers, financial incentives, and managerial imbalance in focal and partner firms. Our results, as indicated, are built on information from only one side of the alliance dyad. Effort to obtain information from both alliance members was undertaken, but many executives were unwilling to identify their partners for (reportedly) strategic reasons. It would be of great value to have independent data from all parties in an alliance. Additional resources are necessary to tackle this type of problem. Notwithstanding these limitations, our findings carry useful implications for managerial practice. Co-marketing alliances give firms the opportunity to reach outside their organizations for help in ways that go beyond traditional vertical arrangements. At a time when increased recognition is being given to the value of outsourcing (Selz 1991), co-marketing alliances offer fresh opportunity for businesses to reach new markets without extending themselves beyond their core competencies. Knowledge of how to manage co-marketing relationships for maximum effectiveness ought be of increasing value. REFERENCES Achrol, Ravi S., Lisa K. Scheer, and Louis W. Stern (1990), "Designing Successful Transorganizational Marketing Alliances," Report No. 90-118 (September). Cambridge, MA: Marketing Science Institute. Adler, Lee (1966), "Symbiotic Marketing," Harvard Business Review, 44 (November-December), 59-71. Aldrich, Howard (1979), Organizations and Environments. Englewood Cliffs, NJ: Prentice-Hall, Inc. American Electronics Association Directory (1988). Palo Alto, CA: American Electronics Association. Anderson, Erin (1984), "The Salesperson as Outside Agent or Employee: A Transaction Cost Analysis," Report No. 84-107. Cambridge, MA: Marketing Science Institute. -- (1988), "Transaction Costs as Determinants of Opportunism in Integrated and Independent Sales Forces," Journal of Economic Behavior and Organization , 9 (April), 247-64. -- and Anne T. Coughlan (1987), "International Market Entry and Expansion via Independent or Integrated Channels of Distribution." Journal of Marketing 51 (January), 71-82. -- and Barton Weitz (1992), "The Use of Pledges to Build and Sustain Commitment in Distribution Channels," Journal of Marketing Research, 29 (February), 18-34. Anderson, James C. and James A. Narus (1990), "A Model of Distributor Firm and Manufacturer Firm Working Partnerships," Journal of Marketing, 54 (January), 42-58. Assael, Henry (1968), "The Political Role of Trade Associations in Distributive Conflict Resolution," Journal of Marketing, 32 (April), 21-8. Bagozzi, Richard P. and Claes Fornell (1982), "Theoretical Concepts, Measurements and Meaning," in A Second Generation of Multivariate Statistics, Vol. 2, Claes Fornell, ed. New York: Praeger Publishers. 24-38. Benson, J. Kenneth (1975), "The Interorganizational Network as a Political Economy." Administrative Science Quarterly. 20 (June), 229-49. Bertrand, Kate (1992), "MacIBM, Will Marketing Bonds Grow as New Product Marriage Matures?" Business Marketing (March), 32-5. Brandt, Richard (1989), "Microsoft Is Like an Elephant Rolling Around, Squashing Ants," Business Week (October 30), 148-52. Bucklin, Louis and Sanjit Sengupta (1992), "Balancing Co-Marketing Alliances for Effectiveness," Report No. 92-120 (August). Cambridge, MA: Marketing Science Institute. Churchill, Gilbert A., Jr. (1979), "A Paradigm for Developing Better Measures of Marketing Constructs," Journal of Marketing Research, 16 (February), 64-73. -- and J. Paul Peter(1984), "Research Design Effects on the Reliability of Rating Scales: A Meta-Analysis," Journal of Marketing Research, 21 (November), 360-75. Clark, Don (1989), "Apple To Sell Its Entire Adobe Stake," San Francisco Chronicle (July 7). Cook, Karen S. (1977), "Exchange and Power in Networks of Interorganizational Relationships," Sociological Quarterly, 18 (Winter), 62-82. Doz, Yves L. (1988), "Technology Partnerships Between Larger and Smaller Firms: Some Critical Issues," International Studies of Management and Organization, 17 (4), 31-57. Dwyer, F. Robert, Paul H. Schurr, and Sejo Oh (1987), "Developing Buyer-Seller Relationships," Journal of Marketing, 51 (April), 11-27. Emerson, Richard M. (1962), "Power-Dependence Relations," American Sociological Review, 27 (February), 1-41. Etgar, Michael (1976), "Channel Domination and Counter-vailing Power in Distributive Channels," Journal of Marketing Research, 13 (August), 254-62. Frazier, Gary L. (1983), "Interorganizational Exchange Behavior in Marketing Channels: A Broadened Perspective." Journal of Marketing, 47 (Fall), 68-78. --, Robert E. Spekman, and Charles O'Neil (1988), "Just-In-Time Exchange Relationships in Industrial Markets," Journal of Marketing, 52 (October), 52-67. Furger, Roberta (1991), "Capitalizing on the Power of the Pen," PC World, 9 (May), 62-6. Gaski, John F. (1984), "The Theory of Power and Conflict in Channels of Distribution." Journal of Marketing, 48 (Summer), 9-29. Heide, Jan B. and George John (1988), "The Role of Dependence Balancing in Safeguarding Transaction-Specific Assets in Conventional Channels." Journal of Marketing, 52 (January), 20-35. -- and -- (1990), "Alliances in Industrial Purchasing: The Determinants of Joint Action in Buyer-Supplier Relationships," Journal of Marketing Research, 27 (February), 24-36. -- and -- (1992), "Do Norms Matter in Marketing Relationships?" Journal of Marketing, 56 (April), 32-44. Jackson, Barbara B. (1985), Winning and Keeping Industrial Customers: The Dynamics of Customer Relationships. Lexington, MA: Lexington Books. Johnston, Russell and Paul R. Lawrence (1988), "Beyond Vertical Integration--The Rise of the Value-Adding Partnership," Harvard Business Review, 66 (July-August), 94-101. Joskow, Paul L. (1987), "Contract Duration and Relationship-Specific Investments: Empirical Evidence From Coal Mines." American Economic Review, 77 (1), 168-85. Levinthal, Daniel A. and Mark Fichman (1988), "Dynamics of Interorganizational Attachments: Auditor-Client Relationships," Administrative Science Quarterly, 33 (September), 345-69. Macneil, Ian R. (1980), The New Social Contract: An Inquiry Into Modern Contractual Relations. New Haven, CT: Yale University Press. Mallen, Bruce (1963), "A Theory of Retailer-Supplier Conflict, Control, and Cooperation," Journal of Retailing, 39 (Summer), 24-32, 51. Muller, Willard F. (1970), A Primer on Monopoly and Competition. New York: Random House, Inc. Nunnally, Jum C. (1978), Psychometric Theory, 2nd ed. New York: McGraw-Hill Book Company. Ouchi, William G. (1980), "Markets, Bureaucracies, and Clans," Administrative Science Quarterly, 25 (March), 129-41. PC Wk (May 27, 1987). "IBM Enlisting Outside Vendors To Broaden Corporate Offerings," 1. Pfeffer, Jeffrey and Gerald Salancik (1978), The External Control of Organizations. A Resource Dependence Perspective. New York: Harper & Row Publishers, Inc. Pindyck, Robert S. and Daniel L. Rubinfeld (1981), Econometric Models and Economic Forecasts, 2nd ed. New York: McGraw-Hill Book Company. Ruekert, Robert W. and Orville C. Walker, Jr. (1987), "Marketing's Interaction With Other Functional Units: A Conceptual Framework and Empirical Evidence," Journal of Marketing, 51 (January), 1-19. Schermerhorn, John R. Jr. (1975), "Determinants of Interorganizational Cooperation," Academy of Management Journal, 18 (December), 846-56. Seidler, John (1974). "On Using Informants: A Technique for Collecting Quantitative Data and Controlling Measurement Data in Organization Analysis," American Sociological Review, 39 (December), 816-31. Selz, Michael (1991), "Small Companies Thrive by Taking Over Some Specialized Tasks for Big Concerns," Wall Street Journal (November 9), B1. Spekman, Robert E. and Kirti Sawhney (1990), "Toward a Conceptual Understanding of the Antecedents of Strategic Alliances," Report No. 90-114 (August). Cambridge, MA: Marketing Science Institute. Stern, Louis W. and Adel I. El-Ansary (1988), Marketing Channels, 3rd ed. Englewood Cliffs, NJ: Prentice-Hall, Inc. Stinchcombe, A. (1985), "Contracts as Hierarchical Documents." in Organization Theory and Management, A. Stinchcombe and C. Heimer, eds. Oslo: Norwegian University Press, 121-71. Van De Ven, Andrew H. (1976), "On the Nature, Formation, and Maintenance of Relations Among Organizations," Academy of Management Review, 1 (October), 24-36. -- and Diane L. Ferry (1980), Measuring and Assessing Organizations. New York: John Wiley & Sons, Inc. Vardarajan, P. Rajan and Daniel Rajaratnam (1986), "Symbiotic Marketing Revisited," Journal of Marketing, 50 (January), 7-17. Williamson, Oliver E. (1975), Markets and Hierarchies. New York: The Free Press. -- (1981), "The Economics of Organizations: The Transactions Cost Approach." American Journal of Sociology, 87 (3), 548-77. -- (1985), "Assessing Contract," Journal of Law, Economics, and Organization, 1 (1), 177-208. Zachary, G. Pascal and Lawrence Hooper (1991), "IBM and Apple Open New Front in PC Wars With Strategic Alliance," Wall Street Journal (July 5). Louis P. Bucklin is Professor, Walter A. Hass School of Business, University of California, Berkeley. Sanjit Sengupta is Assistant Professor, College of Business and Management, University of Maryland, College Park. The tow authors contributed equally to the article. The authors thank Erin Anderson, Gabriel Biehal, Richard Durand, F. Robert Dwyer, Linda Hayes, Jan Heide, Robert Krapfel, Ravi Sohi, and Dave Zaleski for comments on previous drafts. They also gratefully acknowledge the guidance of the editor and two anonymous JM reviewers. The study was funded in part by the Marketing Science Institute.