A Transaction-Efficiency Analysis of an Internet Retailing Supply Chain in the Music CD Industry*

Decision Sciences, Winter 2003 by Rabinovich, Elliot, Bailey, Joseph P, Carter, Craig R

As data sources for the first two indicators, Amazon's measures allow for generalizability because its music CD market share surpasses that of any other Internet site (Collura & Applegate, 2000, 2001). Sales rankings at Amazon are specific to CDs and are updated regularly. The lop 10,000 sellers are updated each hour to reflect sales over the preceding 24 hours and the next 100,000 are updated daily. Thus, the ranking yields an instantaneous popularity measure, where a change in demand immediately causes a title to move in the sales rankings. On the other hand, customer reviews at Amazon have been collected since Amazon opened for business. The number of customer reviews at Amazon, for any given CD, therefore represents a historic measure of CD popularity. Prior title popularity increases are directly reflected in the number of customer reviews for that item at Amazon.

MySimon is a shopping robot that enables comparison shopping for a variety of products, including CDs, at more than 2,000 online stores. MySimon tracked the largest possible number of industry-recognized retailing sites at the time data was collected ("Shop Bots Rule," 2000). Thus, the number of merchants offering a given CD title and listed at MySimon account for an industry-wide measure of popularity. An increase in demand popularity for a given CD title is expected to increase the number of merchants listed at MySimon offering that specific CD title.

Market scope

Internet traffic, as a measure of market scope, is increasingly becoming a scarcer and, consequently, a more valuable resource to Internet sellers (Hanson, 2000). Therefore, a wider market scope may better position sellers in the supply chain to extract demand-side transaction costs and offer, on average, lower list price discounts for CD titles or to source the products sold to consumers from up-stream echelons in the supply chain (Bailey, Faraj, & Yao, 2002; Shapiro & Varian, 1999).

On the other hand, increases in market scope, as a function of Internet traffic (Brynjolfsson & Smith, 2000) and the number of product segments in which Internet retailers compete (Coughlan, Anderson, Stern, & El-Ansary, 2001; Mallen, 1973), may yield lower inventory acquisition costs to Internet retailers. This effect will ultimately yield greater list price discounts stemming from reductions in both Internet retailing marginal costs and, under competitive conditions, prices.

Furthermore, a wider market scope leads to lower supply chain distribution costs, per unit, resulting from statistical economies in inventory-carrying costs and scale economies in transportation and ordering costs (Rabinovich, 2001). This increase in efficiency is expected to generate lower marginal costs and greater average list price discounts for CDs. As well, market scope-related distribution efficiencies are anticipated to enable entities upstream in the supply chain to source products in echelons closer to the consumer echelon.

From the opposing arguments outlined above, and in terms consistent with the empirical model in this research: