Logistics strategy: Does it exist?

Logistics strategy is not a new topic. It has been discussed in the trade and academic literature for twenty years. Despite that fact, most articles concerning logistics strategy are either quite general and descriptive or tightly focused on a specific aspect of logistics strategy (e.g., warehousing in post-1992 European Market). Few authors confidently stake out a specific position and elaborate on a chosen strategy. One exception to this is the Bowersox and Daugherty typology, which put forth the "Process/Market/Information/ Other" classification scheme.' This typology is clearly defined and understandable. However, from a research standpoint, little work has been done in relation to the validity of the typology.

In order to provide empirical evidence regarding the existence and characteristics of the Bowersox and Daugherty logistics strategy typology, this article uses American and Canadian survey data that relate firm practice to strategy specification. The specific objectives are: (1) to identify which of the Bowersox and Daugherty strategies are employed by North American firms; (2) to identify the behavior or constructs determinant of each alternative; and (3) to compare the characteristics of each strategy to determine the nature and magnitude of the differences. Following the background section, the paper reviews the statistical results, including reliability measures, exploratory factor analysis, and analysis of variance, used to identify the latent constructs of each strategy. The final two sections discuss the comparative results and propose conclusions regarding the validity of the Bowersox and Daugherty logistics strategy typology.

BACKGROUND

Bowersox and Daugherty, using a qualitative research design, investigated linkages between organization structure and logistics strategy.2 They identified three distinct strategies: ( 1 ) process, (2) market, and (3) information. The process strategy consists of "a broad group of traditional logistics functions managed as a value added system. The market strategy "typically involves a limited group of traditional logistics activities which are managed across business units."4 The information strategy is most frequently found in a logistics organization that typically has responsibility for traditional and nontraditional (e.g., real estate, data processing) logistics activities. As such, the information strategy focuses on "interorganizational coordination and uses logistics to achieve cooperation and collaboration."5 In a later study, information strategy was renamed channel strategy. To avoid confusion, channel strategy will be used throughout the remainder of this article.

Bowersox and Daugherty emphasize that the process/market/channel classification is not absolute. Firms can exhibit, to varying degrees, aspects of each strategy. They did note, however, that advanced logistics organizations shared certain characteristics. These "commonalties" suggest that certain factors can be associated with advanced logistics practices or strategies. Table 1 details the ten most frequently mentioned commonalties. The identification of these characteristics or commonalties clearly suggests a number of underlying factors that contribute to successful implementation and execution of logistics strategy. What is unclear is how each is related, if at all, to process, market or channel strategies.

Based on this initial work, Bowersox et al. continued investigating logistics strategy in the "Leading Edge" initiative.6 The process/market/channel classification was published in Leading Edge Logistics: Competitive Positioning for the 1990s. Table 2 reproduces the formal definition of each strategy presented therein.

The Leading Edge Logistics study used survey data from U.S. manufacturers. Using the definitions indicated above, firms were asked to indicate their firm's logistics strategy. Table 3 reports the strategic self assessment of 375 U.S. manufacturers according to the process/market/channel typology. Process strategy is identified as the most frequently employed logistics design.

In an effort to identify more clearly the basis of distinct logistics strategies, McGinnis and Kohn conducted a factor/cluster analytic study.7 Their factor analysis identified four primary factors: (1) customer service commitment, (2) integrated computer systems, (3) coordinated logistics, and (4) integrated customer service. Using these four factors, McGinnis and Kohn identified the following logistics strategies: (1) intensive logistics, (2) integrated logistics, (3) low integration, and (4) low effectiveness logistics.

In a later study, McGinnis and Kohn abandoned their 1990 classification scheme, opting for the process/market/channel scheme.8 In this later work, McGinnis and Kohn examined logistics strategy in relation to organizational environment and time competitiveness. A stated objective was to determine if the Bowersox and Daugherty framework could be used to assess logistics strategy.

McGinnis and Kohn conducted a two-phase study during 1989 and 1990. Using the Bowersox and Daugherty typology, McGinnis and Kohn formulated a series of questions related to each of the three strategies. Their results provided tentative support for process and market strategies. Channel strategy failed to achieve an acceptable reliability level and was excluded from further analysis.

In addition, a series of questions was developed concerning customer service commitment, integrated computer systems, logistics coordination effectiveness, company/division competitive responsiveness, logistics system responsiveness, environmental munificence, environmental dynamism, environmental heterogeneity, and environmental hostility. Using a composite score based on these nine areas, plus the strategy variables, McGinnis and Kohn conducted a cluster analysis. Process and market strategies were cluster analyzed, revealing three logistics strategy clusters. These clusters were identified as: (1) intense logistics strategy, (2) balanced logistics strategy, and (3) unfocused logistics strategy. Table 4 reproduces the pertinent results.

It must be stressed that these three strategies should be viewed as substrategies within the process and market classification scheme. The substrategies suggest logistics responses based on environmental and competitive conditions. For example, an intense logistics strategy is identified with an emphasis on customer service and logistics coordination in a very competitive (i.e., hostile) and moderately unpredictable environment. As indicated in Table 4, this is the highest-ranked choice for both process and market strategies. Thus, if pursuing a process or market strategy under a highly competitive and unpredictable environment-and logistics coordination were a priority-then an intense logistics strategy would be an appropriate response within the overall strategy. But as such, the three substrategies do not explain factors distinguishing process and market strategies from each other.

In their discussion, McGinnis and Kohn note that the process/market/ channel classification scheme appears to be "a promising approach for assessing logistics strategy."9 That suggests, however, that much more needs to be understood about each strategy. Additionally, given the McGinnis and Kohn results, it remains to be seen if the channel strategy actually is a verifiable strategy. With that in mind, this research now examines whether underlying factors can be identified for the process/market/channel classification scheme.

STRATEGY CLASSIFICATION RESEARCH

The strategy classification research applies a data base collected at Michigan State University as part of the World Class Logistics (WCL) project. The overall WCL replicates and extends the earlier Leading Edge and Logistics Excellence studies. As a result of using the Bowersox and Daugherty typology in the WCL, comparisons can be made with results from the earlier studies.

A pilot study was conducted in an effort to position and direct the broader research. A three-part questionnaire was developed. It was designed to gather: (1) perceptions regarding industry trends, (2) industry demographic information, and (3) company-specific practices of manufacturers and merchandisers (i.e., wholesalers and retailers). The survey instrument drew upon questions from the previous Michigan State studies, published scales and questions in the logistics, marketing, organizational science, and strategic management literature, and expert opinion of an external advisory board composed of logistics consultants and senior-level executives.

Whereas the Leading Edge study limited strategy investigation to U.S. manufacturers only, the present effort employs a broadened sampling frame. In addition to U.S. manufacturers, U.S. merchandisers and Canadian manufacturers and merchandisers are also included.

The questionnaire was mailed to the general membership of the Council of Logistics Management (CLM) during the summer of 1993. The Canadian Association of Logistics Management (CALM) was identified as the leading Canadian logistics association and comparable to CLM. CALM agreed to participate in the research and distributed the survey to its membership during the fall of 1993. Table 5 provides a breakdown of mailing and response rates for the U.S. and Canada.

The questionnaire requested respondents to identify their logistics strategy according to the process/market/channel/other classification scheme. Respondents were given a standard definition of each strategy. Respondents were also asked a series of forty-three (43) questions regarding their information technology, channel alliances, performance measurement, management command and control, and emerging issues (e.g., "green" concerns) practices. Each item measured responses on a 5-point Likert scale (e.g., I=strongly agree, 5=strongly disagree). Appendix I lists each item. The relationship between logistics strategy and the operating practices offers an empirical basis for determining the relevance of the classification scheme.

U.S. manufacturers were initially grouped according to their strategic orientation before assessing reliability. This approach ensures that each of the three strategies is adequately measured by the questionnaire. Item reliabilities were evaluated using the recommended coefficient alpha measure. to For exploratory research, reliabilities of .50 to .60 are adequate." As indicated in Table 6 the reliability results for U.S. manufacturers' responses are satisfactory. No significant changes from the Leading Edge research are evident.

The expanded sample base of U.S. and Canadian manufacturers and merchandisers (N=818) was then examined for reliability. As the results in Table 7 indicate, the coefficient alphas were satisfactory for each group, suggesting that the questionnaire items are reliable across different channel members and across the U.S. and Canadian participants. Based on this finding, it is appropriate to expand the Leading Edge findings beyond U.S. manufacturers.

Initial Factor Analysis

Factor analysis is a data reduction technique designed to aid in the identification of underlying factors. Exploratory Factor Analysis (EFA) is appropriate when the primary objective is to identify latent constructs and little knowledge exists concerning the amount of unique or error variance.'2 In this case, it was not clear a priori how the behaviors and perceptions would load to form the underlying constructs. Since this research does not hypothesize the construct relationships, EFA is used to identify the underlying dimensions of logistics strategy based on the forty-three (43) original variables.

The goal of EFA is to examine the underlying patterns and relationships in the variables to determine whether or not the information can be condensed or summarized into a smaller set of factors or components. The identification of these key factors and their underlying dimensions facilitates management understanding of the major determinants of logistics strategy.

According to Hair et al. factor loadings greater than .30 are considered significant.'3 For this initial analysis, however, a slightly more stringent cutoff of .40 was used. This higher level enables a clearer factor definition to emerge. Thus, after the initial common factor analysis and an orthogonal rotation (VARIMAX), twenty-two (22) variables were retained.14

The number of factors were restrained using the latent root criterion of 1.0.15 Any factors with eigen values of less than 1.0 are not considered. Using this criterion five factors were retained. The results are shown in Table 8. Non-significant loadings are blocked out.

The factor patterns exhibited in Table 8 indicate a clear structure. Factor 1 has significant loadings on those items related to alliances. Factor 2 groups items concerning information systems. Factor 3 distinctly loads on items dealing with specific EDI practices and applications. Factor 4 can be termed an "inventory" factor as it groups items relating to inventory level and deployment. The final factor, Factor 5, loads significantly on two items relating to restructuring and reengineering.

Based on these results, it is apparent that there is an underlying structure to the items in the questionnaire. Each identified factor addresses a topic common to the general discussion of logistics. It is now of interest to investigate whether or not differences exist between the specific strategies when evaluated in the context of these factors.

Figure 1 illustrates the relationship being investigated. If there are differences, it can be concluded that logistics strategy is operationalized through different combinations of these factors. If no differences exist at this aggregate level one might conclude that any differences between the strategies are minor-that, in fact, the strategies have more commonalties than differences.

MANOVA

MANOVA is an appropriate multivariate technique to use when evaluating groups of variables. A number of metric dependent variables can be grouped together and their overall composite measure can be tested for differences across independent categorical groups. This approach specifically addresses the problem of multicollinearity among variables that may not be apparent in a series of ANOVAs."

Significant items for each factor, identified through the previous factor analysis, were grouped and evaluated against the process/market/channel classification scheme. The Student-Newman-Keuls multiple range test was used to determine whether or not mean differences exist.

The results of the MANOVA application indicate that some differences do exist between the three strategies. No differences were evident for Factor 3 (EDI Applications) and Factor 4 (Inventory Levels and Deployment). Consequently, those factors are not discussed. Table 9 illustrates the means and significant differences for Factor 1 (Alliance Practices).

Internal and external co-location deal with a company's employees working off-site at a different firm as part of a cooperative agreement. According to the results, those respondents indicating a channel strategy differ in this respect from process and market strategies. Firms employing a channel strategy are more likely to exchange employees with suppliers and customers. There appears to be general agreement, however, in the other alliance practice characteristics.

Table 10 summarizes the results regarding the information systems factor. Based upon the results of items Q30 (Interdepartmental Operating Goals), Q31 (Logistics Information Capability), and Q33 (Logistics Information Satisfaction) significant differences exist among the three strategies in terms of the adequacy and capability of existing information systems. The results suggest that firms employing a channel strategy tend to rate their information systems capabilities lower. This could be because channel strategy firms are more demanding thus resulting in lower ratings. More likely, however, firms employing an effective channel strategy require more sophisticated and integrated information capabilities. The integration required by firms employing a channel strategy is more difficult to achieve so channel firms are less satisfied.

Table 11 contains the results of the reengineering/restructuring factor. Although all participants either strongly agree or agree that major changes have occurred in their logistics organization, Channel strategy respondents have apparently seen less change than their counterparts.

The results from the MANOVA analysis suggest commonalties and differences among and between the three strategies, particularly on highly loaded items. Given these results it is worthwhile to probe further in an effort to determine how the specific strategies overlap and diverge.

Factor Analysis by Strategy

Table 12 presents a comparison of responses based on the process/market/channel classification scheme. In the 1987 Leading Edge study, U.S. manufacturers selected process strategy as the most common logistics strategy.' The 1993 GLR results for U.S. manufacturers reveal that there has been no change in the relative rank of the strategies between 1987 and 1993. However, among U.S. manufacturers there has been a distinct increase in the popularity and deployment of Process strategy. This increase has come at the expense of market strategy.

Combined results for U.S. and Canadian manufacturers and merchandisers indicate nearly identical figures with U.S. manufacturers alone. This finding suggests that U.S. merchandisers and Canadian manufacturers and merchandisers do not differ greatly from U.S. manufacturers in the selection of logistics strategies. Based on this finding and the sufficient reliability discussed earlier, it is reasonable to use the expanded sample in the EFA independently for each logistics strategy.

Once again, given the exploratory nature of this effort and the primary goal of identifying underlying dimensions of logistics strategy, EFA was performed on the original forty-three (43) variables. For this analysis the loading cutoff was again .40. Although the intent is now to identify factor structure unique to each strategy, factor interpretability is still a key issue. Therefore, the higher cutoff is appropriate.

An initial EFA with an orthogonal rotation (VARIMAX) was performed. Eight variables were identified as not significant for any strategy. These variables were eliminated and a subsequent factor analysis was repeated.

The fact that a greater number of variables were retained at the individual strategy level than at the aggregate level is not surprising. This was expected as the aggregate level masks items unique to a single strategy. Consequently, several variables that were not used at the aggregate level were used in the exploratory strategy factor analysis.

A VARIMAX rotation method was again used in an effort to simplify factor interpretation. The rotated results for each strategy are shown in Tables 13,14, and 15 respectively. Only significant scores are noted. The number of factors are restrained using the latent root criterion of 1.0. Any factors with eigen values of less than 1.0 are not considered.

As indicated in Table 13 five factors emerged for process strategy Variables loading on Factor 1 strictly include alliance-related items. For shorthand purposes this factor will be called ALLIANCE1. Factor 2 primarily deals with the area of information technology utilization within the firm. This factor is called INFOTECH. The third factor underlying process strategy concerns the deployment of information applications and the resources devoted to support this deployment. This factor is labelled INFOAPP1. The variables loading on the fourth factor deal with restructuring issues and changing logistical arrangements. Consequently, this factor is called CHANGE 1. The final factor underlying Process strategy highlights inventory levels past and present as well as inventory costing policies. This factor is termed INVENTORY 1.

Table 14 displays the results for market strategy. Seven factors emerged for market strategy. As was the case with process strategy, the most important factor in market strategy is related to alliances. One additional question (Q24) loads in market strategy's alliance factor, but it logically fits with elements of alliance formation, structuring, and monitoring. To highlight the slight difference with process' ALLIANCE 1 factor this factor will be called ALLIANCE2.

The second and third market strategy factors contain the variables that loaded on one factor in process strategy (INFOTECH). The distinctions within the market strategy are between those information items concerning the volume and standards of information transactions (Factor 2) and the competitive advantage perspective of technological applications (Factor 3). These factors are respectively, INFO and COMPETE1. The fourth factor in market strategy coalesces around management decisions and related performance measurement standards. The theme of this factor is decision-making and therefore the factor is labelled DECISION. The fifth factor is also concerned with performance measurement. However, it differs from the previous factor in the sense that the items loading on the fifth factor relate to absolute numbers, past and present, of performance measures tracked within the firm. Additionally, this factor acknowledges the role of customer input and its relationship to number of measures tracked. This factor is called PERFORM.

The sixth factor contains the major variables present in process' INVENTORY1 factor. The only difference is that the inventory factor in market strategy also contains a reverse logistics variable (Q7). This factor is called INVENTORY2. The seventh and final factor under market strategy is similar to the CHANGE1 factor in process strategy. However, two of the four variables loading on this factor in market strategy are different from those found in CHANGE 1. The difference is in the comprehensiveness of the elements found in the market strategy focus. This factor goes beyond change in logistics process and organization (i.e., as found in process strategy) and encompasses changes in information systems and the importance of an "organizing" mission statement. In order to highlight this distinction this factor will be labelled REDESIGN.

Seven factors meet the minimum eigen value criterion for channel strategy. The factor loadings are displayed in Table 15. The first factor in channel strategy contains all the variables found in the previous two alliance factors. However, it goes far beyond alliance-related matters and includes many more variables. At first glance there appears to be no discernible pattern to this factor. Closer examination reveals an internal/external orientation. Variables related to internal performance measurement (Q30/Q32) are tied to solicitation of external customer input (Q7/Q13/Q34) through the mechanism of alliance-based relationships (Q11/Ql2/Q14) which in turn reflects operating procedures (Q8/Q36). In short, this factor entails a broadly based planning focus. For that reason this factor is called PLAN.

The second channel factor resembles the COMPETE 1 factor of Market strategy. However, the EDI emphasis is placed elsewhere in channel strategy. To underscore this difference this factor is labelled COMPETE2. The third factor differs slightly from process' CHANGE1 factor. Consistent with the interorganizational focus of channel strategy, channel's change-related factor embraces external as well as internal co-location and does not emphasize restructuring. In order to indicate the similarities while still highlighting the differences this factor is called CHANGE2. The fourth factor, INVENTORY3, contains some elements of previous inventory factors. But it also introduces performance and cost of capital measures, perhaps again stressing the interorganizational considerations.

The fifth factor shares a common factor with market strategy-that being INFO. The sixth factor again links channel and process strategies. In this case, channel strategy does not use all the elements of process' INFOAPP1. This reduced set is called INFOAPP2. INFOAPP2 lacks the tie to performance measures and operating goals found in process strategy. Finally, the seventh factor is unique to channel strategy and is labelled INFOMGMT. This factor ties together senior logistics executives and their role in developing and upgrading information systems.

DISCUSSION

Table 16A summarizes the factor names in order of importance for each strategy. Table 16B rearranges the same factors according to overlaps and similarities. Factors at the same level but with different numerical suffixes generally differ by only one or two items. As one reviews the tables several key points are apparent.

The results provide support for the commonalties (Table 1) originally conceptualized by Bowersox and Daugherty in 1987.18 The factors identified at the strategy level support six of the ten commonalties: (1) emphasis on planning (PLAN/DECISION/INFOMGMT); (2) formal performance measurement (PERFORM); (3) emphasis on managing relationships (ALLIANCE1 and 2/PLAN); (4) significant users of management information systems (INFO/INFOTECH/INFOAPP1 and 2); (5) seek to coordinate operations (INVENTORY 1-3); and (6) frequent organizational change (CHANGE1 and 2/REDESIGN). This finding indicates direct linkages between the organization of a firm's logistics practices and its logistics strategy.

A second key point is illustrated in Table 16B. It is clear that significant overlap exists across the three strategies. Initially, this is disconcerting as it implies a lack of true differences between each of the strategies. However, an alternative explanation is that logistics fundamentally must perform the same activities (e.g., managing inventory levels, moving and storing goods). Therefore, overlap is to be expected across strategies as logisticians contend with the basic demands placed upon logistics.

What emerges then is not necessarily clear-cut differences in underlying factors but rather differences in emphasis and intensity. For example, the increasing use of information technology in logistics requires every strategy to invest in information systems and applications. But whereas a process strategy lumps technological applications and the related competitive aspect into a single factor, INFOTECH, market strategists distinguish between the two concepts, resulting in INFO and COMPETE factors. Thus, the same variables are important in each strategy-but the level of importance differs. Rather than an "either/or" choice, it may be more appropriate to consider specific logistics strategies as a "bundling" of common tactical and operational variables.

The results emphatically demonstrate a third point-it is difficult to ignore the importance of alliances. In each of the three strategies, alliance variables either entirely or greatly contribute to the initially extracted factor. This finding supports logistics academics focusing research efforts in this area. But it also underscores the need for additional research linking logistics strategy to alliances. The fact that this study indicates firms are concerned with the complexities of alliance creation and maintenance within a strategy context is important and should be pursued.

A fourth and final point is that taken as a composite view, the results suggest additional insights into each of the three strategies. Consistent with the definition put forth in Leading Edge, the factors underlying process strategy can be viewed as more internally focused than the factors underlying either market or channel strategies. The process strategy illustrates the fewest constructs with the major characteristics focused on efficiency. The technology constructs (INFOTECH and INFOAPP1) stress the ideas of internal process integration and efficiency. Similarly, CHANGE1 focuses on changing logistical requirements along with network and process restructuring. For example, internal co-location appears in the process CHANGE factor but external co-location does not. The information-related factor INFOAPP further illustrates this internal focus as interdepartmental operating goals and integrated information applications are tightly linked to items measuring process' respondents evaluation of their firms' logistical information systems capability and resource level.

Market strategy contains many of the same items prevalent in process strategy. The difference is that additional items not associated with the process strategy give the market strategy an obvious external focus. Such items as shared mission statements, seeking customer input, external co-location, and reverse logistics reflect this difference. More importantly, each of these items brings the external focus to factors that in the process strategy are strictly limited to an internal perspective. For example, the ALLIANCE factor is the most important factor in both the process and market strategies. In the latter strategy, however, external co-location is an additional variable associated with the creation and monitoring of alliance policies-suggesting a greater appreciation of the external firm's position. Similarly, reverse logistics is associated with the INVENTORY factor in the market strategy but not in the process strategy. Reverse logistics must include external firms, but it can also impact internal inventory practices. Recognition of this fact is apparent in the more externally focused market strategy.

Other differentiating characteristics of the market strategy are illustrated through the information and decision-making constructs. The COMPETE and INFO constructs demonstrate that market strategy firms view their logistics information systems as a source of competitive advantage. This view of information systems as a means to identify (DECISION), monitor (PERFORM), and retain (COMPETE1) profitable customer segments is quite different from the Process focus on efficiency.

Finally, channel strategy appears to adhere to its name. The linking of firms, first and foremost, in an organized and coordinated fashion, is most evident in the PLAN factor. Within this factor the alliance items are coupled with distinct customers, customer input, the measurement of customer satisfaction, reverse logistics, and consistent interdepartmental operating goals. The channel strategy also evidences strong consideration of ultimate consumer considerations and the requirement for completing reverse logistics requirements. This internal and external boundary spanning posits a channel-wide perspective that goes well beyond the external focus of the market strategy. In addition, the strong influence of information utilization across several factors suggests the use of information technology to support this boundary spanning viewpoint.

To put a management perspective on the research, all three strategies have the common objective of trying to manage the logistics process. However, each strategy demonstrates a somewhat different emphasis. The process focus emphasizes internal integration and efficiency. While the process focus does incorporate the ALLIANCE construct, it appears to be more from a development and control perspective rather than demonstrating a strong indication of integrated operations.

The market strategy focuses on identification, monitoring, and delivery of products and services to meet the needs of specific customer segments. The market-based strategy increases the scope and depth of information system requirements specifically with respect to maintaining a competitive advantage and decision making. Firms employing a market strategy must have enhanced information systems to effectively service and monitor each customer segment's requirements. While alliances are an important dimension for both the market and process strategies, market strategies seem to demonstrate a stronger sense of relationship as demonstrated by consideration of issues like employee co-location and reverse logistics.

The channel strategy exemplifies a stronger focus on integrated planning and operations. The themes include integration of operations, measures, information technology, and information sharing capability. There is also evidence of integration in the focus on the consumer for channel redesign.

The fact that channel strategy exhibits only two unique factors, PLAN and INFOMGMT, is interesting. The strong overlap with process and market strategies may perhaps explain why McGinnis and Kohn were unable to test Channel strategy due to low reliability. The characteristics of channel strategy may often be confused with other strategies, thereby confounding interpretation and results. Further investigation in this area of the Bowersox and Daugherty typology is warranted.

CONCLUSIONS

Does logistics strategy, as defined by the Bowersox and Daugherty typology, exist? The results of this study lead to a cautious optimism, similar to McGinnis and Kohn's belief, that the process/market/channel classification scheme is "promising." A number of the commonalties originally posited in 1987 are supported and the factors identified with each strategy in this study point to specific differences between the strategies, consistent with the definitions used in the 1989 Leading Edge research. Although overlaps are apparent, it appears they are logical when viewed in the context of logistics work practices. While the results indicate that there are differences in emphasis for the firms that employ each strategy, the results also raise some issues.

It seems that today's business environment with its customer and supply chain focus would flow directly to market or channel strategies. However, the increase in the relative number of firms reporting use of a process strategy indicates that this is not the case. This suggests either that managers do not believe that their firms are moving toward a more market or channel-focused strategy or there is some confusion over what a processfocused strategy is. These results suggest that managers seem to equate a process strategy primarily with internal considerations. This confusion may result from definitional differences regarding the meaning of a process strategy or from the differences in the perspectives of various managerial levels (managers versus vice presidents). Further research is necessary before an unqualified "yes" can be given to the question, "Does logistics strategy exist?"

Two limitations exist. First, if the more lenient +.30 level were used in the exploratory factor analyses the factors would not exhibit a clean structure. Considerable overlap would then exist between factors, thereby confounding factor interpretability. This problem leads, in part, to the second limitation, which is that the limited question set certainly has not exhausted potential variables. Perhaps other variables exist that would aid in factor interpretability and might reveal greater distinctions between the strategies. For example, is strategy dependent on the type of product and its attributes or can the same strategy adapt to any product? In short, the richness of logistics strategy variables is far from exhausted.

Future research in this area must concentrate on the items composing the underlying constructs. Lack of this information makes it difficult to state conclusively and definitely any absolutes about logistics strategy. Despite that fact, this study provides further evidence of the existence of differentiated logistics strategies and should encourage other researchers to move forward in this area.

NOTES

1Donald J Bowersox, Patricia J. Daugherty, Cornelia L. Dr/ge, Dale S. Rogers, and Daniel L. Wardlow, Leading Edge Logistics: Competitive Positioning for the 1990s (Oak Brook, Ill.: Council for Logistics Management, 1989).

2Donald J. Bowersox and Patricia J. Daugherty, "Emerging Patterns of Logistical Organization," Journal of Business Logistics 8 (Winter 1987): 46-60. 3Same reference as Note 1, p. 51. 4Same reference as Note 1, p. 52. Same reference as Note 1, p. 53. 6Same reference as Note 3.

7Michael A. McGinnis and Jonathon W Kohn, "A Factor Analytic Study of Logistics Strategy," Journal of Business Logistics 11 (Spring 1990): 41-63. 8Michael A. McGinnis and Jonathon W Kohn, "Logistics Strategy, Organizational Environment, and Time Competitiveness," Journal of Business Logistics 14 (Spring 1993): 1-23.

9Same reference as Note 8, p. 16.

'10Gilbert A. Churchill, "A Paradigm for Developing Better Measures of Marketing Constructs," Journal of Marketing Research 16 (February 1979): 64-73. "II Nunnally, Psychometric Methods (New York: McGraw-Hill, 1967). '2Joseph E Hair Jr., Rolph E. Anderson, Ronald L. Tatham, and William C. Black,

Multivariate Data Analysis with Readings, (New York: Macmillan Publishing Company, 1992), p. 231.

13Same reference as Note 12, p. 239.

14Oblique rotation is recommended as more realistic when high correlations exist between variables. Although that is the case with this data set an oblique rotation (Promax) failed to yield interpretable results. For that reason the orthogonal rotation was performed.

15Same reference as Note 12, p. 237. 16Same reference as Note 12, p. 157. "Same reference as Note 1. 'Same reference as Note 2.

ABOUT THE AUTHORS

Steven R Clinton is a doctoral candidate at Michigan State University. He has published in journals and numerous conference proceedings. His research interests include international supply chain management, import/export channel relationships, and selection and evaluation of service providers.

David J. Closs is a professor of marketing and logistics in the Eli Broad College of Business at Michigan State University. He is co-author of Logistical Management and has authored numerous articles in the areas of logistics strategy, systems, modeling, inventory management, and forecasting.

Copyright Council of Logistics Management 1997
Provided by ProQuest Information and Learning Company. All rights Reserved