Joseph Bailey

<bailey@rpcp.mit.edu>

Technology, Management and Policy

Massachusetts Institute of Technology

Erik Brynjolfsson

<erikb@stanford.edu>

Sloan School of Management

Massachusetts Institute of Technology

and

Graduate School of Business

Stanford University

presented at

The Twenty-Fifth Annual

Telecommunications Policy Research Conference

September 27-29, 1997

Radisson Plaza Hotel

Alexandria, Virginia

We quantitatively and qualitatively discuss early pricing strategies for Internet commerce by examining prices for several near-commodity products from retailers on the Internet and a matched set of firms selling identical products through conventional channels. In total, we have collected and analyzed total of 23,789 observations from 52 Internet and conventional retailers for 337 distinct titles of books, music compact disks and software for the period from June 1996 through August 1997.

If has been argued that the Internet lowers search costs and makes physical location increasingly irrelevant, so that retailers of near-commodity products should be more likely to face Bertrand price competition to determine who will dominate the market. Low (marginal cost) prices and minimal price dispersion typically characterize Bertrand price competition.

However, our preliminary analysis suggests that prices on the Internet are generally higher than prices of identical products sold by retailers with physical stores in all of the product markets we examine. Furthermore, in two of the three markets, price variance is higher for Internet retailers. We consider several reasons for these observations: 1) demographics and convenience: the decision to buy goods online is a signal of a higher willingness-to-pay for the products we tracked, 2) market immaturity: firms are still experimenting with strategies for setting prices and developing customer relationships; 3) high search costs: the Internet does not currently reduce search costs significantly; and 4) seller collusion: electronic commerce facilitates certain types of collusive equilibria among sellers.

There is some empirical evidence in support of each of these hypotheses: 1) Online consumers are generally wealthier and better educated than the rest of the population; 2) Internet retailers appear to change their prices much more frequently than physical world retailers and there are (obviously) far more new entrants in this market; 3) the online market for software, which is characterized by generally more technologically-sophisticated consumers and which has a well-developed search intermediary, appears to exhibit less price variance than the other markets, and 4) the pricing strategies of Amazon books appears to have become surprisingly less aggressive after the Barnes & Noble created a significant web presence.