【正文】 s in making their MRO purchases more efficient (see, for example,Purchasing, 2001a,b,c and d). One example of what seem tobe a mon tendency is MRO purchasing at Boeing, whereit ‘‘is moving from a nonstrategic, ad hoc, decentralisedpurchasing activity to a more strategic and virtually centralised purchasing practice’’ (Purchasing, 2001d).According to Bechtel and Patterson (1997), purchases ofMRO items were previously based strictly on obtaining thebest price. Hence, buying firms tried to lower purchasingcosts by focusing on direct costs for individual transactions.When it became clear that substantial indirect costs acpany this purchasing orientation, firms began to makeinternal rationalisations of these purchases. In particular,the huge amount of paperwork generated appears to havebeen considered a problem. By introducing blanket ordersand system contracting the paper flow could be reducedconsiderably. In this way, the buying firms made it possibleto reduce the time spent on these purchases (Bechtel amp。 Ha?kansson, 2001). However, for most manufacturing firms, thevariety in their purchasing needs implies that the totalnumber of suppliers would remain large even if they concentrated on one supplier per product. In this respect, it is notalways obvious exactly what is to be purchased. This isparticularly problematic because most existing models ofpurchasing are based on specified purchasing needs, whichare thus considered to be ‘‘givens’’.A. Dubois / Industrial Marketing Management 32 (2003) 365–3743673. Cost rationalisation through supplier relationshipsFor most panies, the costs of purchased goods andservices have e to account for the majority of total costs.The primary driving force of this development is the increasing reliance on outsourcing, making efficient purchasingactivities crucial to the financial performance of panies.Therefore, Monczka and Morgan (2000) argue that theremust be an ‘‘a(chǎn)bsolute linkage of sourcing, purchasing, and thesupply chain—to the financial plan’’ and thus to the financialoute of the operations of the firm. The authors concludethat increasing attention to cost management is the mostimportant factor for the future of purchasing and supplymanagement. They also refer to numerous examples wherepanies have been able to reduce purchasing costs. In mostcases, these efforts have concerned direct purchasing costs,such as price, transportation costs, etc. However, sometimes,the indirect costs associated with purchasing (., costs ofwarehousing, administration of paper work, and supplierhandling costs) represent greater potential for rationalisation.The problem is that these costs are more difficult to affectbecause most of them can be tackled only through jointefforts with suppliers. Consequently, these costs are oftenneglected, and Monczka and Morgan (2000, p. 55) conclude:[And] regardless of all talk about strategic cost management, most panies are only at the tip of the iceberg interms of actual practices—in terms of looking at wherecosts reside, looking at cost drivers, building crossenterprise strategies and share the results. All of thesepractices are going to need to be refined. Crossenterprisecost is, and will increasingly be, critical to a firm’s success.It has been argued that purchasing of MRO items is one ofthe areas where actions to reduce costs across firm boundaries are most evident (see, for example, Avery, 1997。Ha?kansson, 2001).Strategies based on high involvement, on the other hand,are assumed to result in ‘‘cost benefits in terms of reducedcosts in production processes and material flows as well asimproved service levels and flexibility. Furthermore, it ispossible for the customer to take advantage of supplier skillsand capabilities to improve the quality of its own productsand services, which, in turn, has revenue benefits’’ (Gadde amp。 MRO。 Ha?kansson, 2001。 Pedersen,2002). There are typically two dimensions that determine thechoice of strategy for each product or purchasing situation:(1) ‘‘the importance of purchasing’’ and (2) ‘‘the plexityof the supply market’’ (Kraljic, 1983).If the buyer cannot avoid dependence on a particularsupplier, the portfolio models typically advise partnership,where the buyer and supplier bee mutually dependent.Hence, high involvement relationships or ‘‘partnerships’’with suppliers are suggested as means of dealing with aplex purchasing situation. However, as it entails reciprocal dependence between the buyer and the supplier,‘‘partnerships’’ are mainly seen as problematic and thussituations that should be avoided if possible.The number of suppliers is of importance in bothpurchasing strategies but for different reasons and in different ways. High involvement is monly associated with asingle sourcing policy and low involvement with dual ormultiple sourcing (Gadde amp。Patterson, 1997). A further improvement of internal purchasing routines was made when purchasing cards wereintroduced. Using these cards meant significant administrative processing savings, since purchasing requisitions andpurchase orders could be eliminated.However, internal efforts to reduce cost while maintaininga large number of suppliers can only take the firm to a certainpoint after which there are no additional cost rationalisationsto be obtained inhouse. Therefore, panies seem to beabandoning this philosophy, favouring consolidation tofewer suppliers (Gadde amp。Kim, 1986), the number of suppliers used, or tendersrequired, has to be balanced against the potential gains inobtaining a low price.When a high involvement strategy is pursued, the numberof suppliers is of importance simply because it is not possiblefor a firm to cooperate closely with too many suppliers.Therefore, single sourcing is often suggested as a me