Traditional and IS-Enabled Customer Acquisition on the Internet

Geographic variation in consumer use of Internet retailers is partly explained by variation in offline shopping costs. Explanations for geographic variation in the efficacy of different customer acquisition methods including traditional methods of offline word-of-mouth (WOM) and magazine advertising and information systems (IS)-enabled methods of online WOM and online search remain unexplored. We estimate a multivariate negative binomial distribution (NBD) model on zip code-level customer counts from a leading Internet retailer and provide new insights into factors explaining geographic variation in the success of these methods. First, we show that target customer density explains geographic variation over and above the impact due to the number of potential customers. Moreover, the effect of density is greatest for offline and online WOM acquisitions; this suggests that density contributes to contagion, connectivity, and a hypothesized "social multiplier." Second, when senders and recipients of WOM share consumption benefits, WOM is more powerful and compelling. We find that location-based convenience benefits have stronger effects on location-dependent offline WOM acquisitions than on location-independent online WOM acquisitions. Third, acquisition channels contribute differently to the total customer pool--offline WOM acquisitions are clustered, whereas magazine acquisitions are dispersed. Finally, separate click-to-conversion data from Coremetrics.com indicates that using the model-based predictions to target specific markets delivers a twofold improvement in actual click-to-order rates. This paper was accepted by Sandra Slaughter, information systems.