مدل سازی بازاریابی برای کسب و کار الکترونیکی

The emergence of e-business is opening up new challenges and opportunities for marketing modelers. Drawing on an illustrative pool of recent articles we seek to convey two points in this note. First, that available theories and approaches may be insufficient in tackling many e-business problems. Second, that marketing modeling for e-business can enrich our field quite remarkably in terms of new theories, data and methods.

Few are likely to disagree with the assertion that marketing modeling efforts over the next decade will reflect the Internet's growing influence on consumer behavior and marketing strategy. While marketing issues facing the likes of Folgers, Intel, Microsoft, Nestle, P&G, Sony and Wal-Mart dominated our journal pages, the future is likely to see more of the issues that concern “new age” companies such as Amazon.com, eBay, Netscape, Palm, Priceline, Web-van and Yahoo, whose success is intertwined with the nature and extent of consumers' adoption and use of the Internet. Leeflang and Wittink (2000) offer an insightful critique of extant marketing modeling efforts and propose fascinating avenues for future research. They have also delineated an accessible model building process. Leeflang and Wittink maintain a slant toward models dealing with grocery products that involve UPC scanner data. Grocery products have several distinct characteristics. They are typically in the mature stage of the product life cycle. Marginal costs are not insignificant relative to consumers' willingness to pay. Network externalities are normally absent. The products are repeat-purchased. Data sources are rich, and marketing models in this context are strongly grounded in econometric and statistical methods. Branding, pricing, promotion and physical distribution are key variables in marketing grocery products. Looking ahead, we see Leeflang and Wittink's (2000) blueprint yielding valuable answers for firms such as Coca Cola, IRI, Nielsen, P&G, Peapod and Webvan. We seek to complement Leeflang and Wittink's efforts by emphasizing modeling efforts related to Internet-driven products and activities. (By “Internet-driven” we mean that a significant component of the buying process (e.g., information search, ordering) of a non-trivial segment of consumers occurs on the Internet.) We draw on an illustrative pool of recent marketing articles to highlight the opportunities for exciting new research in the area. We restrict ourselves to B2C marketing. We seek to convey two important and interrelated points. One, existing marketing theories and approaches may be insufficient and/or inappropriate in tackling many emergent problems in e-business. Therein lie the challenges to marketing modeling for e-business. Two, as the Internet powers its way into our lives, the marketing field can gain new insights by way of whole new theoretical extensions, databases and methodologies if more attention is devoted to modeling e-business issues. We categorize key challenges for and new insights from marketing modeling for e-business under four heads. One, those rooted in the setup of the model (Section 2). Two, those related to theory (Section 3). Three, those arising out of new types of data (Section 4). Four, those related to the method (Section 5). To be sure, there could be some overlap among these four categories. In that sense, our demarcation is subjective. The academic articles on e-business that we draw on are summarized in Table 1. We will draw on an article in the section(s) where it is most relevant. We begin with issues in setting up a model.

Leeflang and Wittink (2000) have presented a detailed and compelling roadmap for future modeling work in marketing. Through our short note we have sought to place an added emphasis on marketing modeling for e-business in B2C settings. As illustrated in this note, the rationales for an e-business focus go beyond the growing importance of this sector in our global economy. Let us recap two such rationales. For one, available results from extant models simply may not hold in a large number of e-business situations. For instance, most extant models are based on the objective of profit maximization. Long-term profit maximization is appropriate for e-business firms as well. Yet, conflicting objectives such as increasing customer share and realizing profits and/or cash flow may be important in the short term. Thus, e-business models may have to balance the divergent short- and long-term objectives. Therefore, extant guidelines may not apply. At a minimum, models have to be recast to meet the new objective(s). For another, if preliminary work on e-business is any indication, there is a tremendous opportunity to enrich our field in terms of new theory, data and methods. For example, Bakos and Brynjolfsson (1999) have demonstrated that optimal bundling strategies for digital information goods are different from those for most traditional products. The works of Moe and Fader (2000) and Degeratu et al. (2000) direct us to databases that are likely to interest and challenge modelers more than scanner data that launched a revolution. The new databases contain not only consumers' choice information but also the search processes that they adopted prior to choice (cf. Degeratu et al.'s Peapod data). The few articles that have appeared thus far on e-business have already brought in fresh insights. Yet these represent the proverbial tip of the iceberg. For the future, e-modelers should position their work carefully against traditional models. For example, what is new about their problem in terms of theory, substance and methodology? What theories continue to hold in their study's setting? What theories do not hold and why? Several sources of differences between traditional and e-business problems have been highlighted in this note. This is intended in part to urge caution in applying an available approach to address a new problem in e-business. Recent papers on e-business seem notable more for their theoretical and substantive contributions than for their methodological novelty. At least so far, this trend is unlike what we have seen in the seventies (with conjoint analysis) and the eighties (with scanner data-based choice modeling approaches). Given new types of rich data (cf. Degeratu et al., 2000 and Moe and Fader, 2000), there are many untapped opportunities for developing new approaches. In essence, the dawn of e-business promises great opportunity for marketing modelers. Let us look forward to this future.