Industrial Marketing Management 36 (2007) 219 – 229 Power and interdependence in buyer supplier relationships: A purchasing portfolio approach Marjolein C.J. Caniels *, Cees J. Gelderman 1 ¨ Faculty of Management Sciences (MW), Open University of the Netherlands (OUNL), P.O. Box 2960, 6401 DL Heerlen, The Netherlands Received 27 September 2004; received in revised form 4 July 2005; accepted 3 August 2005 Available online 23 September 2005Abstract Power and interdependence are generally considered to be important concepts for understanding buyer – supplier relationships. Yet, empiricalresearch on power and interdependence in buyer – supplier relationships is still limited. Power and interdependence issues also play an importantrole in Kraljic’s portfolio approach, which is increasingly used by purchasing practitioners for managing different supplier relations anddeveloping appropriate purchasing strategies. In this paper, the concepts of power and interdependence have been quantified for each quadrant ofthe Kraljic portfolio matrix, using data from a comprehensive survey among Dutch purchasing professionals. Several hypotheses have been testedand the findings largely confirm the theoretical expectations. The observed supplier dominance in the strategic quadrant of the Kraljic matrix is anotable finding, which indicates that even satisfactory partnerships are dominated by the supplier. Therefore, the presumed power symmetry ofbuyer – supplier relationships in the strategic quadrant seems no longer valid.D 2005 Elsevier Inc. All rights reserved.Keywords: Buyer – supplier relationships; Power and interdependence; Purchasing portfolio approach; Kraljic matrix; Purchasing strategies1. Introduction models using other classification dimensions (e.g. Bensaou, 1999; Olsen & Ellram, 1997; Van Stekelenborg & Kornelius, Purchasing portfolio models have received much attention 1994). However, the fundamental assumption of all portfolioin recent literature about professional purchasing. Not only did models seems to be the occurrence of differences in power andKraljic’s seminal paper in the Harvard Business Review in dependence between buyers and suppliers (Dubois & Pedersen,1983 have a broad influence on professional purchasing (see 2002). Kraljic (1983) does not explicitly deal with issues ofthe evidence of Gelderman, 2003; Kamann & Bakker, 2004), it power and dependence. However, some of his recommendationshas also inspired many academic writers to undertake further obviously refer to the power structure (Fexploit power_). Othersresearch into portfolio models (e.g. Bensaou, 1999; Croom, are aimed at reducing the dependence on suppliers (Fdiversify_).2000; Dubois & Pedersen, 2002; Dyer, Cho, & Chu, 1998; Moreover, Kraljic (1983: 112) stated that the general idea of theGelderman & Van Weele, 2002, 2003; Lilliecreutz & Ydres- portfolio approach is to ‘‘minimize supply vulnerability andkog, 1999; Nellore & Soderquist, 2000; Olsen & Ellram, 1997; make the most of potential buying power’’. Therefore, powerWagner & Johnson, 2004; Wynstra & ten Pierick, 2000; and dependence play a significant part in the Kraljic approach.Zolkiewski & Turnbull, 2002). Although power and dependence are generally considered Kraljic’s model classifies a firm’s purchased intermediate important for the understanding of buyer –supplier relation-goods into four categories on the basis of two dimensions: (1) ships (e.g. Cox, 2001; Frazier & Antia, 1995), it seems thatprofit impact and (2) supply risk. Recent adaptations and they are still often overlooked factors in conceptual andrefinements of Kraljic’s model have led to alternative portfolio empirical studies (e.g. Cox, 2001; Maloni & Benton, 2000). Little is known about the exact way in which power and dependence in buyer – supplier relationships enter the Kraljic * Corresponding author. Tel.: +31 45 5762724; fax: +31 45 5762103. E-mail addresses: email@example.com (M.C.J. Caniels), ¨ matrix (Dubois & Pedersen, 2002; Gelderman & Van Weele,firstname.lastname@example.org (C.J. Gelderman). 2003). Moreover, the few portfolio models that do discuss 1 Tel.: +31 45 5762590; fax: +31 45 5762103. power and dependence issues in relation to portfolio matrices0019-8501/$ - see front matter D 2005 Elsevier Inc. All rights reserved.doi:10.1016/j.indmarman.2005.08.012
220 M.C.J. Caniels, C.J. Gelderman / Industrial Marketing Management 36 (2007) 219 – 229 ¨generally focus on the strategic quadrant only (Wagner & Kraljic’s approach includes the construction of two portfolioJohnson, 2004). Buyer –supplier relationships in this quadrant matrices. The first matrix classifies a firm’s purchased productscan be characterized as strategic partnerships. However, many on the basis of two dimensions: profit impact and supply risk.studies have acknowledged that not all supplier relationships Each dimension has two possible values: Flow_ and Fhigh_. Thecan or should be strategic partnerships (e.g. Gadde & Snehota, resulting 2 Â 2 matrix consists of four quadrants (see Table 1).2000; Wagner & Johnson, 2004). In fact, firms are found to Depending on the category, Kraljic identifies certain Fmainbenefit from entering into a variety of relationships with tasks_ for the firm in the interaction with its supplier. He alsodifferent suppliers (e.g. Bensaou, 1999; Lilliecreutz & Ydres- identifies the required information and the decision level inkog, 1999). Therefore, undertaking research into power and organizations per category.dependence in all four quadrants of the matrix for all The main purpose of Kraljic’s approach is to identifyrelationship types is critically important. strategic items. The second Kraljic matrix focuses on this The aim of this paper is to empirically test hypotheses that can category. This matrix shows the relative power position of thebe deduced from the literature on power and dependence with company in the corresponding supply markets. Three generalrespect to all quadrants of the Kraljic purchasing portfolio purchasing strategies are distinguished, depending on thematrix. In order to do this we have defined the concepts of power balance of power in the buyer – supplier relationship: exploitand dependence in terms of buyer’s and supplier’s dependence. (in case of buyer dominance), balance (in case of a balancedSubsequently, we have developed constructs for buyer’s relationship), and diversify (in case of supplier dominance).dependence as well as supplier’s dependence. The empirical Note that Kraljic does not pay much attention to strategicanalysis is founded on a survey among 250 purchasing aspects of product categories other than the strategic items.professionals. On the basis of the survey data we have assessed Other scholars have filled this gap (e.g. Bensaou, 1999; Elliott-power and interdependence in buyer –supplier relationships for Shircore & Steele, 1985; Lilliecreutz & Ydreskog, 1999; Olsenall quadrants of the Kraljic matrix. In general terms this study & Ellram, 1997; Syson, 1992; Van Weele, 2000). They refinedcontributes to a better understanding of the (perceived) power the original matrix and elaborated on the Fmain tasks_ forand interdependence in buyer – supplier relationships. bottleneck, non-critical and leverage items. In addition, they The organization of the paper is as follows. In Section 2 a formulated strategic recommendations, resulting in one overallbrief overview of the Kraljic approach is given and, on the basis purchasing strategy for each cell/category.of recent literature, we will identify hypotheses with respect to Although power and interdependence issues underpin thesepower and dependence for each quadrant. In Section 3 the survey purchasing strategies, they are not explicitly discussed in anydesign and the constructs for the key variables are presented. The of the studies referred to above. Yet, the purchasing strategiesresults of the survey are shown in Section 4. Section 5 will for each of the Kraljic quadrants give rise to hypotheses on theconclude and give suggestions for further research. importance of power and dependence in the Kraljic matrix. In Section 2.3 we will revisit these strategies and connect them to2. Conceptual background the power and interdependence balance.2.1. The Kraljic matrix 2.2. Power and interdependence Kraljic (1977, 1983) introduced a comprehensive portfolio Firms always depend, to varying extents, on their tradingapproach as a tool for professional purchasers. With the help of partner. Early studies on dependence focused on the effects forthe portfolio matrix, professional purchasers could optimize the the buyer of its dependence on the supplier, without taking intouse of capabilities of different suppliers (Nellore & Soderquist, account the supplier’s dependence (e.g., El-Ansary & Stern,2000) and thereby effectively manage suppliers. Currently, 1972). More recent studies have incorporated dependence fromKraljic’s matrix is widely used by purchasing professionals. the perspective of the buyer as well as the supplier (Buchanan,Especially in Western Europe the Kraljic approach has received 1992; Geyskens, Steenkamp, Sheer, & Kumar, 1996; Kumar,large-scale recognition and has attained an increasing degree of Sheer, & Steenkamp, 1995). In other words, dependence isadoption. Lamming and Harrison (2001) stated that Kraljic’s mutual.matrix remains the foundation for purchasing strategies of many Mutual dependence and power are closely related concepts.organizations across sectors. In a survey of Dutch companies The buyer’s dependence on the supplier is a source of power forBoodie (1997) found that 44% of the responding purchasingmanagers used the Kraljic matrix for formulating purchasingstrategies. No less than 80% of industrial companies that operate Table 1on a mass production basis use it. Several years later, Bos, Van The Kraljic purchasing portfolio model (modified from Kraljic, 1983: 111)der Heijden, Goedhart, and Notermans (2005) reported in a Profit impact Supply risksimilar study that portfolio usage was increased to 61%. In the Low Highcourse of time the Kraljic approach has entered many textbooks High Leverage items Strategic itemson purchasing and supply management. Gradually Kraljic has Exploitation of purchasing power Diversify, balance, or exploitgained acceptance in other countries, notably in the USA, Low Non-critical items Bottleneck itemsCanada and Northern Europe. Efficient processing Volume assurance
M.C.J. Caniels, C.J. Gelderman / Industrial Marketing Management 36 (2007) 219 – 229 ¨ 221the supplier, and vice versa. A well-known definition is that the possesses much power, it is not likely that either side is goingrelative power of an organization over another is the result of the to use it. The risk of retaliation is often considered as being toonet dependence of the one on the other. If A depends on B more high (Ramsay, 1996). In addition, when total interdependencethan B depends on A, then B has power over A (Pfeffer, 1981). is high, both partners are faced with high exit barriersSimilarly, Bacharach and Lawler define relative power as ‘‘the (Geyskens et al., 1996).dependence of one party compared to the dependence of the From the above can be concluded that the literature makes aother party’’ (1981: 65). Likewise, Dickson (1983) states that clear difference between the concepts of (1) relative power,the power of one party over another is a function of relative which is the result of interdependence asymmetry; and (2) totaldependence. Anderson and Narus (1990) also use the term power, which is the result of full interdependence of both partiesrelative dependence to refer to the difference between a firm’s on each other and which is commonly referred to as totaldependence on its partner and its partner’s dependence on the interdependence. However, empirical research on the impact offirm. The primary consequence of relative dependence is relative power and total interdependence on buyer – supplierindicated as power. relationships is scarce. To be able to fill this void, both concepts Buchanan (1992) conceptualized power-dependence imba- were defined in terms of buyer’s and supplier’s dependence (cf.lances in buyer –supplier relationships as the difference in Bacharach & Lawler, 1981; Pfeffer, 1981). In this study, thevalue that buyers and sellers attach to the relationship. In buyer’s relative power will be measured as the differenceasymmetric relationships, the most independent partner dom- between supplier’s dependence and buyer’s dependence. Sim-inates the exchange. Balanced relationships refer to domination ilarly, the supplier’s relative power will be measured as theof neither party (Buchanan, 1992). Kumar et al. (1995) use the difference between buyer’s dependence and supplier’s depen-term interdependence asymmetry in this respect, which is dence. This conforms Pfeffer’s (1981: 99) viewpoint that thedefined as the difference between the two partner’s levels of relative power of one social actor over another is the result of thedependence. Symmetrical interdependence exists when parties net dependence of the one on the other. In accordance withare equally dependent on each other. Bacharach and Lawler (1981: 61), total interdependence in a Buyer – supplier relationships that are characterized by relationship will be measured by ‘‘the sum of the parties’asymmetric interdependence are believed to be deficient because dependence on one another’’.the independent partner experiences high power and might beattempted to exploit it (Anderson & Weitz, 1989; Frazier & 2.3. Hypotheses on relative power and total interdependence inRody, 1991; Geyskens et al., 1996). McDonald (1999) states in the Kraljic matrixthis respect that power imbalances within a buyer –supplierrelationship can lead to unproductive partnerships. In the long In Section 2.1 we referred to purchasing strategies for eachterm the position of the weaker party will be eroded too much of the four Kraljic quadrants. These strategies will be used toand the partnership will be destroyed. Anderson and Weitz point characterize the relation between the Kraljic quadrants and theout that ‘‘imbalanced channel relationships are characterized by power-dependence balance in buyer – supplier relationshipsless cooperation and greater conflict’’ (1989: 312). However, (see also Van Weele, 2000).note that an unbalanced relationship does not automatically Strategic products 2 represent a considerable value to theinvolve actual misuse of power (Provan & Gassenheimer, 1994). organization in terms of a large impact on profit and a highPower can provide an effective coordination of exchange supply risk. Examples are engines and gearboxes for automo-relationships, as the distribution of power has become legitimate bile manufacturers, turbines for the chemical industry andover time (Frazier & Antia, 1995). Maloni and Benton (2000) bottling equipment for breweries. Often strategic products canfound empirical evidence, which indicates that power asymme- only be purchased from one supplier (single source), causing atry can be used as a tool to promote supply chain integration and significant supply risk. In order to counterbalance this risk,to induce high levels of performance. firms will aim at building a partnership relationship with its Various researchers have argued that a comprehensive view supplier (Elliott-Shircore & Steele, 1985). The mutual trust andof the interdependence of a dyadic relationship should include commitment that comes with the intensified relationship isnot only interdependence asymmetry (or relative power), but likely to reduce the supply risk to a minimum. A close andalso total interdependence (or total power), for example lasting cooperation with suppliers will lead to improvements inBacharach and Lawler (1981), Frazier and Antia (1995), product quality, delivery reliability, lead times, product devel-Gundlach and Cadotte (1994), Geyskens et al. (1996), Kumar opment, product design, and it will lead to cost reductionet al. (1995). Total interdependence refers to the intensity of a (Hadeler & Evans, 1994; Tuten & Urban, 2001). This situationrelationship. A high level of total interdependence is an can be characterized as one with balanced power. Buyers andindicator for a strong, cooperative long-term relationship in suppliers are both heavily involved in the partnership, thereforewhich both parties have invested. Mutual trust and mutualcommitment will characterize those relationships (Geyskens et 2al., 1996). Besides this loyalty towards the other partner and Although Kraljic (1983) identified three strategies for this quadrant (exploit, balance and diversify), our discussion will be limited to the balance strategy,the accompanying desire to continue the relationship, there is which is in accordance with the approach adopted by many others, e.g. Elliott-an alternative motivation for both firms to keep the partnership Shircore and Steele (1985), Hadeler and Evans (1994), Olsen and Ellramintact. In the case that both parties know that the other party (1997), and Van Weele (2000).
222 M.C.J. Caniels, C.J. Gelderman / Industrial Marketing Management 36 (2007) 219 – 229 ¨mutual dependence is expected to be high. Total interdepen- Table 2dence is high as well, since the relationship is very intense. Expectations on basis of the literature with respect to relative power and total interdependence in the Kraljic matrixHypothesis 1a. The strategic quadrant of the Kraljic matrix is Profit Supply riskcharacterized by balanced power. impact Low HighHypothesis 1b. In the strategic quadrant of the Kraljic matrix High Leverage items Strategic items Buyer dominated: BD < SD Balanced power: BD = SDtotal interdependence is higher than in each of the other Moderate level of High level of interdependence:quadrants. interdependence: (BD + SD) in (BD + SD) in the strategic the leverage quadrant < (BD + SD) quadrant > (BD + SD) in each Bottleneck products have a moderate influence on the in the strategic quadrant and of the other quadrantsfinancial results of a firm, however, they are vulnerable with (BD + SD) in the leverageregard to their supply. Suppliers have a dominant power position quadrant > (BD + SD) in thefor these products (Kempeners & van Weele, 1997). The non-critical quadrantpurchasing strategy is therefore primarily focused on assurance Low Non-critical items Bottleneck items Balanced power: BD = SD Supplier dominated: BD > SDof supply, if necessary even at additional cost. Keeping extra Low level of interdependence: Moderate level ofstocks of the materials concerned or developing consigned stock (BD + SD) in the non-critical interdependence: (BD + SD)agreements with suppliers are examples of this strategy. Firms quadrant < (BD + SD) in each in the bottleneckcan make a risk analysis to determine the most important of the other quadrants quadrant < (BD + SD) in thebottleneck products and the consequences hereof. Contingency strategic quadrant and (BD + SD) in the bottleneckplanning might be a possibility for dealing with unexpected bad quadrant > (BD + SD) in thesituations. Since the buyers and suppliers are not highly involved non-critical quadrantin the relationship, total interdependence in this quadrant is Note that BD refers to buyer’s dependence and SD denotes supplier’sexpected to be lower than in the strategic quadrant. dependence.Hypothesis 2a. The bottleneck quadrant of the Kraljic matrix ischaracterized by supplier dominance. From a purchasing point of view, these items cause only few technical or commercial problems. As a rule of thumb routineHypothesis 2b. In the bottleneck quadrant of the Kraljic matrix products require 80% of the purchasing department’s time, whiletotal interdependence is higher than in the non-critical they often represent less than 20% of the purchasing turnover.quadrant. The handling of these products requires a purchasing strategy In general leverage products can be obtained from various aimed at reducing the logistic and administrative complexitysuppliers. These products represent a relatively large share of the (Olsen & Ellram, 1997). Systems contracting is generallyend product’s cost price in combination with a relatively low advised as the way of doing business with suppliers of routinesupply risk. As a consequence, this segment is buyer dominated products (Elliott-Shircore & Steele, 1985; Kempeners & van(Kempeners & van Weele, 1997). The buyer has many possibilities Weele, 1997). The main idea is to enhance purchasing power byand incentives for negotiation, since small percentages of cost standardization and bundling of purchasing requirements. Thesavings usually involve large sums of money (Olsen & Ellram, routine character of the transaction implies that the mutual1997). At the same time the supply risk is minimal. These dependence between buyers and suppliers is balanced. Totalcharacteristics justify an aggressive approach to the supply market interdependence is low, since the buyer’s dependence and the(e.g. Van Weele, 2000). Frequently, a purchasing strategy on the supplier’s dependence will both be quite low.basis of principles of competitive bidding is pursued. Since Hypothesis 4a. The non-critical quadrant of the Kraljic matrixsuppliers and products are interchangeable, there is no need for is characterized by balanced power.long-term supply contracts. In general, a coordinated purchasingapproach is adopted that has the form of a centrally negotiated Note that it is not necessary to propose a separate hypothesisumbrella agreement with preferred suppliers. Call-off orders are about total interdependence in the non-critical quadrant. Thethen placed as an administrative formality. The buying power is combination of hypotheses 1b, 2b and 3b makes a separateactively used to get better deals with interchangeable suppliers. hypothesis redundant.Total interdependence is expected to be moderate. Although the In sum, the quadrants in the Kraljic matrix correspond to foursupplier’s dependence is expected to be high, the buyer’s basic power-and-dependence positions. Table 2 gives andependence is expected to be quite low. overview of our expectations for the balance of power and the level of total interdependence in each of the four quadrants.Hypothesis 3a. The leverage quadrant of the Kraljic matrix ischaracterized by buyer dominance. 3. MethodologyHypothesis 3b. In the leverage quadrant of the Kraljic matrix 3.1. Survey design, sample and responsetotal interdependence is higher than in the non-critical quadrant. Non-critical products usually have a small value per unit. The hypotheses were tested in a survey among 250Many alternative suppliers can be found for these products. purchasing professionals. For this purpose we translated each
M.C.J. Caniels, C.J. Gelderman / Industrial Marketing Management 36 (2007) 219 – 229 ¨ 223of the purchasing strategies discussed in Section 2.3 into (NEVI) in three stages. Purchasing professionals were speci-comprehensive descriptions of real-life situations (scenarios). fically targeted with this survey. Purchasing professionalsThe description of the four scenarios is given in Table 3. undertake negotiations with suppliers on a daily basis. There- The survey adopts a repeated measures design, i.e. respon- fore, they have vast experience, expertise and insight into thedents had to evaluate a series of identical questions for each of power and dependency issues in the relationship with theirthe four scenarios from their own perspective, i.e. the suppliers. In the first two stages, the questionnaires were sent outperspective of the buyer. The distinct advantage of using a by postal mail. The last stage consisted of an electronic mailing.repeated measures design instead of other experimental designs In each round only those respondents were addressed that hadis that the potential bias caused by individual differences among not responded to an earlier mailing. Tests on the likelihood of agroups of respondents is taken away. Each of the respondents has non-response bias indicated no statistical significant differencesto answer all questions in each of the four scenarios. Therefore, between the first wave and the second wave of respondents.specific characteristics of the respondents, such as I.Q., A total number of 248 responses were received, resulting in aeducation and motivation, do not differ across the four groups response rate of 21.5% (248 / 1153). Questionnaires that werethat reply to the questions in each scenario. In sum, by adopting a completed for less than 90% were declared invalid. Hence, arepeated measures design, variability among groups of respon- total of 10 reactions were discarded due to incompletedents is removed from the error term, which makes the design information. In addition, all responses were removed that scoredmore powerful than randomized designs (Stevens, 2001). on average less than 3 for recognition on a 5-point Likert scale. A potential drawback of the adopted research method is that In this way, another 22 responses were omitted, resulting in anrespondents might not be able to fully visualize themselves in the effective response rate of 18.7% (216 / 1153). This can beproposed descriptions, resulting in unreliable answers. This considered as quite satisfactory for an industrial mail survey inshortcoming was countered by including an entry for recognition Europe (Erdogan & Baker, 2002; Frohlich, 2002), especiallyof the scenario, i.e. respondents were asked to assess the degree when we take into account that the questionnaire was very long.in which they recognize the described situation. In the analysis of Whereas Yu and Cooper (1983) have shown that an optimalthe data, we removed the survey results for respondents with low response is received on questionnaires that contain 40 to 50scores on Frecognition_ from the database. In this way we items, the questionnaire contained 8 pages with 175 items inensured the validity of the results. total. The survey procedure included a pilot study aimed atenhancing the reliability and the validity of the questionnaire. 3.2. Developing constructs for buyer’s and supplier’sThe pilot study entailed discussions with six purchasing dependenceprofessionals in four manufacturing companies in the Nether-lands. Several issues were discussed during these interviews, In Section 2.2 we have explained that relative power andsuch as the clarity of the questionnaire items, the recognizability total interdependence were defined in terms of buyer’s andof the scenarios, the time needed to fill in the questionnaire and supplier’s dependence (cf. Bacharach & Lawler, 1981; Pfeffer,issues for further improvement of the items. On account of the 1981). There is no general consensus in the literature on howpilot study several improvements have been made in the these two concepts should be operationalised. We havedescription and layout of the questionnaire items, response examined relevant academic contributions to the literature tooptions and scenarios. derive several aspects of organizational dependence that could The final questionnaire has been administered to 1153 be used to operationalise buyer’s dependence and supplier’smembers of the Dutch Association of Purchasing Management dependence.Table 3Description of the scenarios corresponding to the Kraljic quadrantsScenario Characterization Corresponding Kraljic quadrant Scenario description1 Maintain partnership Strategic quadrant Consider a product with a high purchasing risk and a high financial value. You consider the supplier as an important partner with whom a satisfactory strategic relationship exists. The performance of the supplier is excellent. Both parties have an interest in continuing the relationship and the parties have a good mutual understanding.2 Keep safety stocks Bottleneck quadrant Consider a product with a relative low financial value, but a high purchasing risk. Your firm is vulnerable regarding the supply of one supplier. You try to ensure a constant supply by keeping high stocks.3 Partner of convenience Leverage quadrant You have a very favorable negotiating position with this product. The purchasing risk is low, while the product has a relatively high financial value. Negotiations are fierce. You let those suppliers prevail that offer the lowest price while guaranteeing quality and prompt delivery. Competitive bidding is one of your tactics. You only allow short-term contracts.4 Pooling of requirements Non-critical quadrant Consider a product that has a relative low financial value and a low purchasing risk. The product is not very critical for your company, but still it has to be purchased. You choose to buy the product as a part of a package of similar products from a certain supplier. In this way it is possible to have only one supplier for several products.
224 M.C.J. Caniels, C.J. Gelderman / Industrial Marketing Management 36 (2007) 219 – 229 ¨ Jacobs (1974) introduces two concepts from economic alternative suppliers and low switching costs between supplierstheory to describe dependence, namely Fessentiality_ and are much more important to the buyer than the relative amount ofFsubstitutability_. He points out that it is of primary importance money that is involved. Hence, we have excluded Ffinancialto the concept of dependence whether A can do without B magnitude_ from the construct of buyer’s dependence and(essentiality of a resource) or whether other sources are included it in the supplier’s dependence construct.available (substitutability of the resource). Scholars in Re- There is a difference in the perspective of buyers andsource Dependence Theory refer to Fessentiality_ as Fthe suppliers with regard to the second component of dependenceimportance of a resource_, which is said to be determined by as well. The criticality of a resource refers to the degree in which(1) the relative financial magnitude of the resource and (2) the the organization is able to continue its business processes in thecriticality of the resource (Pfeffer & Salancik, 1978). On the absence of the resource. In essence, however, the concept is two-other hand, substitutability can be subdivided into (1) the fold in nature. On the one hand it refers to a need foravailability of alternative sources and (2) the level of relation technological expertise of the partner, on the other hand itspecific investments (i.e. the costs involved with switching points to issues of logistical indispensability (Cagliano, Caniato,between suppliers) (e.g. Bourantas, 1989). & Spina, 2002). The need for technological expertise is critical In addition to these conceptual studies, several empirical for both parties, buyer and supplier. In an industrial context,studies pay attention to the concept of organizational depen- companies rely more and more on technologically advanceddence. Most of these studies regard organizational dependence (key) suppliers. From the supplier’s perspective a similaras an explanatory variable instead of trying to find the factors argument holds. Companies increasingly need the criticalthat have an influence on it. The few studies that try to explain expertise and specialized knowledge of their (industrial)the factors that influence organizational dependence (e.g. customers. Logistics-based dependence, on the other hand, isNooteboom, de Jong, Vossen, Helper, & Sako, 2000; Sriram, less an issue to the supplier than it is to the buyer. The buyer isKrapfel, & Spekman, 1992) find similar factors as indicated by mainly interested in receiving the goods in a way that isthe conceptual studies above. In all empirical studies logistically compatible with its own production system. InFimportance_ and Fsubstitutability_ determine dependence, contrast, the supplier will deliver the goods in any logistic waywhereby importance is found to have a positive relationship that is required, as long as the buyer will pay for it. The buyer’swith dependence and substitutability is found to have a negative main concern is the correct delivery of the goods, whereas therelationship. The empirical studies do not provide decisive supplier’s main concern is of a financial nature. On the basis ofanswers concerning the statistical significance of the dimensions these considerations we have redefined the concept of resourceof dependence. Only some tentative empirical evidence can be criticality in the construct of supplier’s dependence to solelyfound that importance and substitutability are significantly include the need for the buyer’s technological expertise. Therelated to dependence. construct for buyer’s dependence includes the logistical In summary, the analysis of conceptual and empirical studies indispensability of the supplier, in addition to the need for ashows that organizational dependence contains four key supplier’s technological expertise.characteristics: With respect to the availability of alternative sources and switching costs the dependence positions of buyers and suppliers (1) the financial magnitude of the exchanged resources; are symmetrical. The buyer depends as much on the supplier as (2) the criticality of the resources; the other way around. Both buying and supplying organizations (3) the availability of alternative sources; invest in the relationship with their trading partner. When the (4) switching costs, incurred when replacing a trading supplier develops and uses dedicated equipment assigned partner. exclusively to one customer this will result in high switching costs if the relationship deteriorates. On the other hand, buying With these characteristics in mind we have set up organizations also face relation specific investments, makingconstructs for buyer’s dependence and supplier’s dependence significant investments in suppliers.as follows. For obvious reasons the overall dependency on the other The Resource Dependence Theory posits a positive re- party is also included in both variables, the construct of buyer’slationship between the (financial) magnitude of a resource and dependence and the construct of supplier’s dependence. Table 4the mutual dependence of the trading partners. It is reasonable toassume that the first characteristic of organizational dependence, Table 4the relative financial magnitude of transactions, particularly Aspects that compose buyer’s dependence and supplier’s dependencerelates to supplier’s dependence. In the view of the supplier, a Construct Buyer’s dependence Supplier’s dependencerelatively important buyer (in terms of financial magnitude of thetransaction) will have a very powerful position in negotiations, Aspects Logistical indispensability Financial magnitude (items) Need for supplier’s Need for buyer’scausing the supplier to be dependent on the buyer. For the buyer technological expertise technological expertisethe financial magnitude of the transaction with a certain supplier Availability of alternative Availability ofis much less crucial for it’s dependence position. If switching suppliers alternative buyerscosts are low, the buyer will not experience any dependence on Switching costs buyer Switching costs supplierthe supplier. Therefore, factors such as the availability of Overall buyer’s dependence Overall supplier’s dependence
M.C.J. Caniels, C.J. Gelderman / Industrial Marketing Management 36 (2007) 219 – 229 ¨ 225Table 5 4. ResultsReliability analysis: Cronbach’s alphas Construct A comprehensive view of the dyadic nature of buyer –Category in the Kraljic matrix Buyer’s dependence Supplier’s dependence supplier relationships should include the assessment of (1) theStrategic quadrant 0.64 0.74 difference between buyer’s and supplier’s dependence (netBottleneck quadrant 0.61 0.76 dependence) which corresponds to the relative power of eachLeverage quadrant 0.64 0.69 party; and (2) the sum of buyer’s and supplier’s dependenceNon-critical quadrant 0.67 0.72 (total interdependence) which indicates the intensity and development phase of the relationship between parties.summarizes the aspects of buyer – supplier relationships that Note that the repeated measure design of the survey waspertain to buyer’s dependence and supplier’s dependence. Note accounted for when analyzing the data. We used Estimatedthat each aspect was measured by a corresponding item in the Marginal Means (EMMEANS) in SPSS (General Linearquestionnaire, using a 5-point Likert scale. Model-repeated measures) for the comparisons of the means in different scenarios. Bonferroni adjustments were made to3.3. Validity analysis control for Type I errors (Tabachnick & Fidell, 2001). Table 6 shows the average score for the construct variables A reliability analysis using Cronbach’s alpha was performed supplier’s dependence and buyer’s dependence in the fourto ensure the internal consistency of the items that constitute scenarios. The last two columns in Table 6 show the resultingeach construct as shown in Table 4 (Cronbach, 1951). Table 5 findings for the power balance and total interdependence. Notepresents the results of the reliability analysis. The table shows that column 3 indicates the relative power position of the buyer,the values of Cronbach’s alpha for each of the four Kraljic i.e. the difference between supplier’s dependence and buyer’squadrants. The coefficients of Cronbach’s alpha are all higher dependence. Therefore, a negative sign for the power balancethan 0.60, indicating an acceptable internal consistency and shows that the supplier dominates the relationship, whereas areliability of the constructs. positive sign points to buyer dominance. Total interdependence Additional correlation analysis in each quadrant showed that (last column in Table 6) is measured by the sum of buyer’sitems that should be related, are strongly correlated, indicating dependence and supplier’s dependence. On a scale that runsconvergent validity. In addition it was found that items that from +2 (minimal interdependence) to +10 (maximum interde-theoretically should not be related, did not correlate (discri- pendence) we consider values below 5 as low, between 5 and 7 asminant validity). moderate and above 7 as high. Furthermore, we used explanatory factor analysis (principal Several points emerge from Table 6. Column 1 in Table 6components analysis with varimax rotation) to identify a shows a relatively high buyer’s dependence on the right handpossible underlying factor structure (see the table in the side of the Kraljic matrix (bottleneck and strategic quadrant),Appendix for the detailed results of the factor analysis). For and a relatively low buyer’s dependence on the left hand side ofeach quadrant of the Kraljic matrix a separate factor analysis has the matrix (non-critical and leverage). These findings are inbeen executed. The factor solutions confirmed our expectations, accordance with our prior expectations. Column 2 in Table 6i.e. in each quadrant two components were found: one contain- indicates that the findings for supplier’s dependence are noting the items that were expected to indicate buyer’s dependence clear-cut. Remarkably, the supplier’s dependence in the leverageand the other containing the items that were expected to point to quadrant is lower than we might have expected in advance (2.68supplier’s dependence. Almost all items had factor loadings that on a 5-point scale). The same holds for the strategic quadrant,exceeded the recommended level of 0.50 (Hair, Anderson, where the supplier’s dependence is medium (3.31 on a 5-pointTatham, & Black, 1998). Note that none of the items cross- scale). However, the relatively low supplier’s dependence in theloaded on both factors. In other words, the items that should not bottleneck and the non-critical quadrant confirm our expecta-be related were indeed not related. We can safely conclude that tions on the basis of the literature.buyer’s dependence and supplier’s dependence in each of the These findings have notable implications for the hypoth-four quadrants are dissimilar constructs. eses. With respect to the power balance (column 3) it wasTable 6Power and interdependence in the Kraljic matrix (n = 216) Scenario Buyer’s Supplier’s Relative Total interdependence dependence (1)* dependence (2)* power (2) À (1)** (1) + (2)***Strategic quadrant Maintain partnership 4.03 3.31 À0.72a 7.34Bottleneck quadrant Keep safety stocks 3.66 2.32 À1.34a 5.97Leverage quadrant Partner of convenience 2.41 2.68 +0.27a 5.09Non-critical quadrant Pooling of requirements 1.90 1.97 +0.07 3.87a Difference between supplier’s dependence and buyer’s dependence is significant at p < 0.05.*Supplier’s and buyer’s dependence are measured by taking the average score on the items of each construct respectively.**Power is measured on a scale from À4 (maximum supplier’s dominance) to +4 (maximum buyer’s dominance).***Total interdependence is measured on a scale from +2 (minimum interdependence) to +10 (maximum interdependence).
226 M.C.J. Caniels, C.J. Gelderman / Industrial Marketing Management 36 (2007) 219 – 229 ¨found that buyer – supplier relationships are dominated by the Table 8supplier in the strategic quadrant, rejecting Hypothesis 1a. We Differences with respect to interdependence between quadrants of Kraljic matrixwill come back to this unexpected result in the next Results with respect to interdependenceparagraph. The results in column 3 indicate that buyer –supplier relationships are also dominated by the supplier in (1) INDEPstrat À INDEPbottle +1.37a (2) INDEPstrat À INDEPnon +3.47athe bottleneck quadrant, confirming Hypothesis 2a. In (3) INDEPstrat À INDEPlev +2.25acontrast, the buyer is dominant in the leverage quadrant, (4) INDEPnon À INDEPbottle À2.10asupporting Hypothesis 3a. The t-tests showed that, with the (5) INDEPnon À INDEPlev À1.22aexception of the non-critical quadrant, all quadrants presented a Significantly different from zero at p < 0.05.statistically significant differences between supplier’s andbuyer’s dependence. The insignificant difference between When we look at the underlying data of the supplier’ssupplier’s dependence and buyer’s dependence in the non- dependence construct (Table 7) we find that respondentscritical quadrant refers to a significant power balance in this reported that they:quadrant, thereby confirming Hypothesis 4a. The results for the strategic quadrant lead to a rejection of - have more need for the supplier’s technological expertiseHypothesis 1a that suggested a balanced power relationship than vice versa; andfor this quadrant. The literature puts a lot of emphasis on the - face higher switching costs than the suppliers; andidea that buyer – supplier relationships in this quadrant are - have fewer alternative trading partners than the suppliertypically characterized as satisfactory relationships based on does.trust, commitment and open communication (e.g. De Ruyter,Moorman, & Lemmink, 2001; Morgan & Hunt, 1994). The last column of Table 6 shows the level of totalHowever, the results suggest that such satisfactory relation- interdependence in the various quadrants.ships in the strategic quadrant have an asymmetric power From the literature on the Kraljic matrix we are expecting:balance. From the buyer’s perspective the supplier dominatesthe relationship. When we examine the data in greater detail, - high levels of interdependence in the strategic quadrantthis result can be traced back to the medium level of (Hypothesis 1b);supplier’s dependence (3.31 on a 5-point scale) and the high - moderate levels of interdependence in the bottleneck andlevel of buyer’s dependence (4.03 on a 5-point scale) that leverage quadrant (Hypotheses 2b and 3b respectively); andwas reported by the respondents in the survey. This deviates consequentlyfrom what is expected from the literature, which indicates - low levels of interdependence in the non-critical quadrant.that supplier’s dependence is high as well in the strategicquadrant. Several separate tests have been undertaken to determine Presumably, once a buyer has entered a partnership this whether total interdependence in one quadrant significantlyensures a disproportionate raise in the dependence of the buyer differs from total interdependence in all other quadrants. Theon the supplying partner. Cox, Lonsdale, Watson, and Qiao results are reported in Table 8, where INDEPstrat, INDEPbottle,(2003) give a possible explanation for this phenomenon. In INDEPlev and INDEPnon refer to total interdependence in thetheir belief a pre-contractual situation of balanced power shifts strategic, bottleneck, leverage and non-critical quadrantto a post-contractual situation of supplier dominance, when the respectively.supplier refuses to offer balance in the relationship and locks The results largely confirm our prior expectations. Totalthe buyer in. Other writers suggest that unbalanced relation- interdependence has its highest value in the strategicships may not always be troublesome: a known distribution of quadrant, confirming Hypothesis 1a. As expected, totalpower between both partners could provide effective co- interdependence is moderate in the bottleneck and theordination of the exchange relationship (Frazier & Antia, leverage quadrant, thereby giving confirmation of Hypotheses1995). A relative power position can be used to enhance the 2b and 3b respectively. Consequently, the non-criticalnature of relational exchange between trading partners. It quadrant contains the lowest value of total interdependence.seems that the distribution of power can become legitimated We conclude that the empirical findings confirm the hithertoover time, so that both social actors expect and value a certain untested theoretical notions about total interdependence inpattern of influence. buyer – supplier relationships.Table 7Item scores and standard deviations (between parentheses) for the buyer’s dependence and the supplier’s dependence construct in the strategic quadrantBuyer’s dependence Supplier’s dependenceLogistical indispensability 4.63 (0.541) Financial magnitude 3.95 (0.761)Supplier’s technological expertise 4.08 (0.868) Buyer’s technological expertise 3.32 (1.038)Alternative suppliers 2.81 (1.168) Alternative buyers 3.27 (1.079)Switching costs buyer 4.08 (1.074) Switching costs supplier 3.36 (1.131)Overall buyer’s dependence 4.22 (0.757) Overall supplier’s dependence 3.25 (0.998)
M.C.J. Caniels, C.J. Gelderman / Industrial Marketing Management 36 (2007) 219 – 229 ¨ 2275. Summary, conclusions and suggestions for further justified in the case of channel studies (manufacturer –research distributor), where relations often revolve around one major supplier. However, the method is also often used in studies Purchasing practitioners maintain a variety of supplier relating to industrial relationships, in which the limitation torelationships with different suppliers, and the management of the dominant supplier is not a self-evident point of departure.these relationships is increasingly based on a portfolio frame- Alternatively, many studies invite respondents to answerwork. Power and dependence are generally considered to be questions referring to Ftheir suppliers_ in general. On theimportant for understanding buyer– supplier relationships, yet basis of this study we conclude that neither approach providesthese factors are often overlooked in empirical studies. Little is a comprehensive insight into buyer – supplier relationships. Weknown about the role of power and dependence in buyer – have found evidence of the existence of four buyer – suppliersupplier relationships from a purchasing portfolio perspective. relationships that differ significantly with respect to relativeIn this study we filled this gap by first deducing hypotheses from power and interdependence. This result confirms the notionthe literature on Frelative power_ and Ftotal interdependence_ in that companies maintain a portfolio of differentiated suppliereach of the four Kraljic quadrants. Second, we empirically tested relationships. Therefore, neither reference to Fa key supplier_these hypotheses using data from a comprehensive survey nor reference to Fsuppliers_ in general takes into account theamong Dutch purchasing professionals. full variation in the actual supplier base of a company. This A comparison of the hypotheses with the empirical findings common practice among researchers should only be used ifleads to the following results. The hypotheses that concerned the the research question specifically requires such a samplingrelative power position of the buyer and the supplier in each method. In all other cases it should be discarded. Future andquadrant (1a – 4a) were confirmed, except Hypothesis 1a. That further surveys can no longer ignore the large variety inis, we observed supplier dominance in the strategic quadrant, supplier relationships.where one would expect a balanced power situation on basis of The findings also have managerial relevance. They offerthe literature. This provocative result sheds new light on the evidence that in the strategic quadrant a supplier-dominatedbuyer’s view on issues of power and dependence. It indicates that partnership is perceived to be satisfactory from the perspectivesuppliers are perceived to dominate satisfactory partnerships. of the buyer. Industrial marketers should be aware that All hypotheses concerning the level of total interdepen- professional purchasers feel dominated by them, even indence in each quadrant (1b –3b) were confirmed. Table 9 satisfactory relationships. Possibly this finding could enticesummarizes the differences between the expectations on the marketers to try to exploit their power and to skim off thebasis of the literature on power and dependence, and the market surplus. However, marketers should be aware of theresults from this study. negative effects of the exploitation of their position. In fact, a The theoretical contribution of this paper lies primarily in situation in which the buyer feels dominated yet satisfied isthe notion that future research can no longer assume that desirable, since the buyer will not search for alternativesbuyer –supplier relationships in the strategic quadrant of the (suppliers or products). Therefore, marketers should nurtureKraljic matrix are necessarily characterized by symmetric these kinds of relationships.power positions. When we combine this result with the high For professional purchasers the managerial implication oftotal interdependence reported in this quadrant, we arrive at a this study is that they should be aware that dependencemore refined theoretical implication of this study. Future implies vulnerability. Buyers should ask themselves whetherstudies should take into account that a relationship which is there are sufficient benefits attached to the relationship tocharacterized by a high involvement of buyers and suppliers offset the obvious disadvantages of such a vulnerable anddoes not necessarily imply a balanced power position between dependent position towards a supplier. In addition, purchasersthe parties, but can yet be satisfactory, at least from the point of should assess the risks that are harbored in this kind ofview of the buyer. relationship, and explore possibilities that might increase the Furthermore, this study holds an implication for future bargaining power of their company. In other words, even inresearch, which pertains to the sampling method in survey- satisfactory relationships, buyers should explore the marketbased studies. Many studies ask respondents to express their by scouting for alternative suppliers and determining theiropinions on their relationship with a single (type of) supplier, competencies. Furthermore, professional purchasers shouldusually the key or the major supplier. This approach is become aware of their own power basis. They should investigate to what extent the perceived supplier dominanceTable 9 is based on an objective assessment of the relationship.Comparison of relative power and total interdependence in the Kraljic matrix: Despite the rigor of the analysis, this study contains severaltheory and practice limitations that might entice further research. One of the Relative power Total interdependence limitations of the study concerns the fact that the survey was Expected Observed Expected Observed confined to the perspective of the buyer. Yet, buyer – supplierStrategic Balanced Supplier dominance Highest Highest relationships are dyadic in nature. Suppliers might have differentBottleneck Supplier dominance Supplier dominance Moderate Moderate opinions on the power and interdependence structure of theLeverage Buyer dominance Buyer dominance Moderate Moderate various buyer –supplier relationships. Therefore, there is a needNon-critical Balanced Balanced Lowest Lowest to more fully study the supplier’s perspective in order to
228 M.C.J. Caniels, C.J. Gelderman / Industrial Marketing Management 36 (2007) 219 – 229 ¨establish whether or not both parties perceive each other’s power by the supplier in the strategic quadrant calls for furtherposition in the relationship in the same way. research to identify the circumstances under which supplier Another limitation concerns the sample, which was drawn dominated relationships provide a problem for the buyerfrom a list of members of the Dutch Association of Purchasing (Frazier & Antia, 1995). Additionally, the operationalisationManagement (NEVI). Although the sample included a wide in this study could serve as a promising point of departure forrange of industry sectors, the generalizability of the results further quantitative research to the issues of power andwould benefit from the inclusion of firms in the service sector, dependence in buyer – supplier relationships. The level ofas well as other Dutch and international companies. This might relative power might be related to the sizes of the buying andbe the object of a further study. the supplying companies. Alternatively, network positions or With respect to additional issues for further research we the positions in the supply chain could be included as apropose the following. The finding that buyers feel dominated determining factor of relative power.Appendix A. Results of the factor analyses (factor loadings) Leverage quadrant Strategic quadrant Non-critical quadrant Bottleneck quadrant Factor 1 Factor 2 Factor 1 Factor 2 Factor 1 Factor 2 Factor 1 Factor 2 Buyer’s Supplier’s Buyer’s Supplier’s Buyer’s Supplier’s Buyer’s Supplier’s dependence dependence dependence dependence dependence dependence dependence dependenceItems of buyer’s dependence:Logistical indispensability 0.445 À 0.144 0.654 À0.058 0.529 À0.035 0.556 À0.091Supplier’s technological expertise 0.692 0.205 0.697 0.237 0.791 0.288 0.709 0.107Alternative suppliers À0.544 À 0.232 À0.484 À0.037 À0.617 0.064 À 0.470 0.334Switching costs buyer 0.721 0.177 0.545 0.330 0.586 0.216 0.654 0.139Overall buyer’s dependence 0.757 0.005 0.790 0.054 0.643 0.142 0.676 0.179Items of supplier’s dependence:Financial magnitude À0.170 0.661 0.325 0.534 À0.027 0.695 0.036 0.831Buyer’s technological expertise 0.194 0.683 0.223 0.710 0.330 0.573 0.067 0.874Alternative buyers À0.083 À 0.451 0.023 À0.570 À0.215 À0.499 0.105 0.766Switching costs supplier 0.105 0.717 À0.722 0.793 0.103 0.727 À 0.046 À0.450Overall supplier’s dependence À0.053 0.799 0.188 0.783 À0.323 0.824 À 0.037 0.731References Cox, A., Lonsdale, C., Watson, G., & Qiao, H. (2003). Supplier relationship management: A framework for understanding managerial capacity and constraints. European Business Journal, 15(4), 135 – 145.Anderson, J. C., & Narus, J. A. (1990). A model of distributor firm and Cronbach, L. J. (1951). Coefficient alpha and the internal structure of tests. manufacturer firm working partnerships. Journal of Marketing, 54, 42 – 58. Phychometrica, 16, 297 – 334.Anderson, J. C., & Weitz, B. A. (1989). Determinants of continuity in Croom, S. R. (2000). The impact of web-based procurement on the conventional industrial channel dyads. Marketing Science, 8, 310 – 323. management of operating resources supply. Journal of Supply ChainBacharach, S., & Lawler, E. (1981). Power and politics in organizations. Management, 36(1), 4 – 13. San Francisco’ Jossey-Bass. De Ruyter, J. C., Moorman, L., & Lemmink, J. G. A. M. (2001).Bensaou, M. (1999). Portfolios of buyer – supplier relationships. Sloan Antecedents of commitment and trust in customer – supplier relation- Management Review, 40(4), 35 – 44. ships in high technology markets. Industrial Marketing Management,Boodie, M. L. J. (1997). World class purchasing in Nederland is fictie en 30(3), 271 – 286. helaas nog geen werkelijkheid. Berenschot Inkoopenquete 1997. Utrecht’ ˆ Dickson, P. R. (1983). Distributor portfolio analysis and the channel Berenschot Inkoopmanagement. dependence matrix: New techniques for understanding and managing theBos, R. L., Van der Heijden, G., Goedhart, E. I., & Notermans, R. M. M. channel. Journal of Marketing, 47, 35 – 44. (2005). World-class purchasing: Grip op Inkoop. Berenschot Inkoopen- Dubois, A., & Pedersen, A. (2002). Why relationships do not fit into quete 2004/2005. Utrecht’ Berenschot Group B.V. ˆ purchasing portfolio models: A comparison between the portfolio andBourantas, D. (1989). Avoiding dependence on suppliers and distributors. industrial network approaches. European Journal of Purchasing and Long Range Planning, 22(3), 140 – 149. Supply Management, 8, 35 – 42.Buchanan, L. (1992). Vertical trade relationships: The role of dependence and Dyer, J. H., Cho, D. S., & Chu, W. (1998). Strategic supplier segmentation: symmetry in attaining organizational goals. Journal of Marketing The next Fbest practice_ in supply chain management. California Research, 29, 65 – 75. Management Review, 40(2), 57 – 77.Cagliano, R., Caniato, F., Spina, G. (2002). Supply strategy configurations: El-Ansary, A. I., & Stern, L. W. (1972). Power measurement in the distribution Linking operational integration with supply management practices. Paper channel. Journal of Marketing Research, 9, 47 – 52. presented at the 11th International IPSERA Conference, Enschede, The Elliott-Shircore, T. I., & Steele, P. T. (1985, December). Procurement Netherlands. positioning overview. Purchasing and Supply Management, 23 – 26.Cox, A. (2001). Understanding buyer and supplier power: A framework for Erdogan, B. Z., & Baker, M. J. (2002). Increasing mail survey response rates procurement and supply competence. Journal of Supply Chain Manage- from an industrial population. Industrial Marketing Management, 31(1), ment, 37(2), 8 – 15. 65 – 73.
M.C.J. Caniels, C.J. Gelderman / Industrial Marketing Management 36 (2007) 219 – 229 ¨ 229Frazier, G. L., & Antia, K. D. (1995). Exchange relationships and interfirm Nellore, R., & Soderquist, K. (2000). Portfolio approaches to procurement: power in channels of distribution. Journal of the Academy of Marketing Analysing the missing link to specifications. Long Range Planning, 33, Science, 23, 321 – 326. 245 – 267.Frazier, G. L., & Rody, R. C. (1991). The use of influence strategies in interfirm Nooteboom, B., de Jong, G., Vossen, R. W., Helper, S., & Sako, M. (2000). relationships in industrial product channels. Journal of Marketing, 55(1), Network interactions and mutual dependence: A test in the car industry. 52 – 69. Industry and Innovation, 7(1), 117 – 144.Frohlich, M. T. (2002). Techniques for improving response rates in OM survey Olsen, R. F., & Ellram, L. M. (1997). A portfolio approach to supplier research. Journal of Operations Management, 20(1), 53 – 62. relationships. Industrial Marketing Management, 26(2), 101 – 113.Gadde, L. E., & Snehota, I. (2000). Making the most of supplier relationships. Pfeffer, J. (1981). Power in organizations. Massachusetts’ Pitman Publishing Industrial Marketing Management, 29(4), 305 – 316. Inc.Gelderman, C. J. (2003). A portfolio approach to the development of Pfeffer, J., & Salancik, G. R. (1978). The external control of organizations—a differentiated purchasing strategies. Eindhoven’ Eindhoven University of resource dependence perspective. New York’ Harper and Row Publishers. Technology. Provan, K. G., & Gassenheimer, J. B. (1994). Supplier commitment inGelderman, C. J., & Van Weele, A. J. (2002). Strategic direction through relational contract exchanges with buyers: A study of interorganizational purchasing portfolio management: A case study. International Journal of dependence and exercised power. Journal of Management Studies, 31(1), Supply Chain Management, 38(2), 30 – 37. 55 – 68.Gelderman, C. J., & Van Weele, A. J. (2003). Handling measurement issues and Ramsay, J. (1996). Power measurement. European Journal of Purchasing and strategic directions in Kraljic’s purchasing portfolio model. Journal of Supply Management, 2(2/3), 129 – 143. Purchasing and Supply Management, 9(5 – 6), 207 – 216. Sriram, V., Krapfel, R., & Spekman, R. (1992). Antecedents to buyer – sellerGeyskens, I., Steenkamp, J. E. M., Scheer, L. K., & Kumar, N. (1996). collaboration: An analysis from the buyers’ perspective. Journal of The effects of trust and interdependence on relationship commitment: A Business Research, 25(4), 303 – 320. trans-Atlantic study. International Journal of Research in Marketing, 13, Stevens, J. P. (2001). Applied multivariate statistics for the social sciences. 303 – 317. Portland’ Lawrence Erlbaum Associates.Gundlach, G. T., & Cadotte, E. R. (1994). Exchange interdependence and Syson, R. (1992). Improving purchase performance. London’ Pitman. interfirm reaction: Research in a simulated channel setting. Journal of Tabachnick, B. G., & Fidell, L. S. (2001). Using multivariate statistics. Marketing Research, 31, 516 – 532. Needham’ Allyn and Bacon.Hadeler, B. J., & Evans, J. R. (1994). Supply strategy: Capturing the value. Tuten, T. L., & Urban, D. J. (2001). An expanded model of business-to- Industrial Management, 36(4), 3 – 4. business partnership formation and success. Industrial Marketing Manage-Hair, J. F., Anderson, R. E., Tatham, R. L., & Black, W. C. (1998). Multivariate ment, 30, 149 – 164. data analysis with readings. Upper Saddle River, New Jersey’ Prentice Van Stekelenborg, R. H. A., & Kornelius, L. (1994). A diversified approach Hall. towards purchasing and supply. In C. Walter, & F. Kliemann (Eds.),Jacobs, J. (1974). Dependency and vulnerability: An exchange approach to the Evaluation of production management methods (pp. 45 – 55). Amsterdam’ control of organizations. Administrative Sciences Quarterly, 19(1), 45 – 59. Elsevier Science Publishers.Kamann, D. F., & Bakker, E. F. (2004). Changing supplier selection and Van Weele, A. J. (2000). Purchasing management: Analysis, planning and relationship practices: A contagion process. Journal of Purchasing and practice. London’ Chapman and Hall. Supply Management, 10(2), 55 – 64. Wagner, S. M., & Johnson, J. L. (2004). Configuring and managing strategicKempeners, M., & van Weele, A. J. (1997). Inkoopportfolio: Basis voor supplier portfolios. Industrial Marketing Management, 33(8), 717 – 730. inkoop-en marketingstrategie. In H. W. C. Van der Hart, & A. J. van Weele Wynstra, J. Y. F., & ten Pierick, E. (2000). Managing supplier involvement in (Eds.), Dynamiek in Commerciele Relaties. Bunnik’ F & G Publishing. ¨ new product development: A portfolio approach. European Journal ofKraljic, P. (1977). Neue Wege im Beschaffungsmarketing. Beschaffung Aktuell Purchasing and Supply Management, 6(1), 49 – 57. (pp. 20 – 26). Yu, J., & Cooper, H. (1983). A quantitative review of research design mailKraljic, P. (1983). Purchasing must become supply management. Harvard survey response behavior. Public Opinion Quarterly, 55, 613 – 639. Business Review, 61(5), 109 – 117. Zolkiewski, J., & Turnbull, P. (2002). Do relationship portfolios and networksKumar, N., Sheer, L. K., & Steenkamp, J. E. M. (1995). The effects of provide the key to successful relationship management? Journal of perceived interdependence on dealer attitudes. Journal of Marketing Business and Industrial Marketing, 17(2), 575 – 597. Research, 32, 348 – 356.Lamming, R. C., & Harrison, D. (2001). Smaller customers and larger Marjolein C.J. Caniels is Assistant Professor Supply Chain Management in ¨ suppliers: The potential for strategic purchasing approach: A case study. the Faculty of Management Sciences of the Open University of the Nether- Proceedings of the 10th International IPSERA Conference, Jonkoping, ¨ ¨ lands. Her research interests include buyer – supplier relationships, supply chain Sweden (pp. 595 – 610). management, technological change and economic growth. She has publishedLilliecreutz, J., & Ydreskog, L. (1999). Supplier classification as an enabler for among others in the Cambridge Journal of Economics and in Industrial and a differentiated purchasing strategy. Global Purchasing and Supply Chain Corporate Change. Management, 11, 66 – 74.Maloni, M., & Benton, W. C. (2000). Power influences in the supply chain. Cees J. Gelderman is Associate Professor of Marketing and Purchasing Journal of Business Logistics, 21(1), 49 – 73. Management in the Faculty of Management Sciences of the Open University ofMcDonald, F. (1999). The Importance of power in partnership relationships. the Netherlands. His research interests include buyer – supplier relationships, Journal of General Management, 25(1), 43 – 59. purchasing portfolio management, and power and dependence. He hasMorgan, R. M., & Hunt, S. D. (1994). The commitment – trust theory of published among others in the Journal of Supply Chain Management and the relationship marketing. Journal of Marketing, 58, 20 – 38. Journal of Purchasing and Supply Management.