By Troy J. Strader
Assistant Professor, Management Information Systems, Iowa State University
E-mail: tstrader@iastate.edu
URL: Iowa State College of Business

Troy Strader is Assistant Professor of Management Information Systems at Iowa State University. He received his Ph.D. from the University of Illinois at Urbana-Champaign in 1997. His research interests include electronic commerce, strategic impacts of information systems, and information economics.

Recent studies have shown that consumers pay less for products in Internet (electronic) markets than they do in traditional markets such as retail stores (Lee, 1998; Strader, 1997). This result was hypothesized in earlier research (Bakos, 1991). The question is why. One theory is that Internet markets reduce the costs associated with consumer's search for price information. Search can be characterized as a buyer canvassing various sellers to ascertain the most favorable price. For search costs, the cost involved would be determined by the value the buyer places on their time and effort (Stigler, 1961). With Internet markets, consumers are not limited to shopping locally, but may easily compare prices from around the world. In this article I report the findings of a study designed to test the relationship between prices and search costs for three different market types (retail stores, card shows, and Internet markets) in the sports trading card industry.

The Internet market that I studied is supported by USENET, an Internet-based newsgroup system. The six separate newsgroups studied include one for discussion of issues outside of buying and selling (rec.collecting.cards.discuss) and five for buying and selling baseball (rec.collecting.sport.baseball), basketball, football, hockey and miscellaneous items. On average there are more than ten thousand buy, sell and trade advertisements posted per week to these groups.

There are incentives for collectors to participate in this Internet market because it enables them to find each other without an intermediary (such as a specialty sports card shop). Internet markets have the potential for a major impact in the sports trading card industry because there are a large number of buyers and sellers, and there is a widespread and accepted standard for describing the products. It also involves product prices and transaction values comparable to a wide range of industries.

To gather the data for this study a survey was sent electronically to 400 randomly selected active participants in the Internet market. The population surveyed included adults who had purchased sports trading cards through an Internet market in 1996. They were chosen for the study because they are able to compare their experiences in traditional markets and Internet markets. Of the surveys sent, 58 were completed and returned for a response rate of 14.5%. The sample was predominately male Americans, but this was expected based on the consumers in the sports trading card industry.

The first hypothesis addressed by this study is that prices in Internet markets are lower than in traditional markets (retail stores, specialty card shops, and so forth). To test this hypothesis consumers were asked to report the average percentage of the full retail prices, as reported in the monthly Beckett price guides, they paid for items during 1996. The Beckett magazines report retail prices for a wide range of sports cards and it is a widely accepted source of price information throughout the industry. Consumers reported that on average they paid 81.2% of full retail prices in retail stores (for example Wal-Mart), 72.4% in specialty card shops, 53.8% at card shows (for example several dealers gather at a mall) and 45.5% in Internet markets. From the prices reported by consumers it was found that prices in Internet markets are significantly lower than in other sales channels. The question then is why? One explanation is that it is easier to search for price information in Internet markets than in other sales channels. This leads to the second hypothesis that prices paid by consumers are positively correlated with sales channel marginal search costs.

To test this hypothesis we go back to the data above. In each sales channel there is an associated marginal search cost. This can also be characterized as the level of "convenience" for consumers associated with a sales channel. Marginal search costs can be defined as one divided by the number of sellers in close proximity to each other in a sales channel. Close proximity means it takes little time or effort to identify another seller's price for a product. The assumption is that searching Internet markets, particularly using search capabilities in computer software, reduces the cost to search for price information such that a greater number of sellers can be canvassed within the same total search cost.

For this study total consumer search cost for identifying all prices in close proximity in a market is defined as a cost of one. For retail stores and specialty card shops the number of sellers in close proximity is one. Thus, the marginal search cost is high, 1 / 1 = 1.000. For one unit of search a consumer could find one price for a product. Searching is "less convenient" for consumers in these channels. From the survey results I found that there are an average of 51.3 dealers at any particular card show. Assuming that each dealer would have any particular product (for illustration purposes) the marginal search cost is lower, 1 / 51.3 = 0.020. For the same amount of total search cost, a consumer could potentially find fifty times as many product sellers, and prices, at a card show than at a card shop or retail store.

For the Internet market, the number of sellers in the same place is potentially very high. Searching is "more convenient" for consumers in this channel. This is especially true as software used to search these markets, such as the World Wide Web, improves. The average number of sellers of a particular item is hard to gauge. For illustration purposes I use the approximate average number of buy and sell ads posted in the USENET market for one sport in a day, 350. Thus, the marginal search costs associated with comparing prices is very low, 1 / 350 = 0.003. I assume that the Internet market is a competitive market. Statistical analysis shows that the data support the hypothesis that product prices are positively correlated with search costs. Retail stores and specialty card shops result in the highest prices, card shows have prices in the middle, and Internet markets result in the lowest prices. The correlation between marginal search costs and percentage of retail prices paid is significant.

Why then aren't all products purchased through Internet markets? One answer is that there are varying degrees of risk associated with sales channels (and individual sellers). For example, retail stores have the advantage of allowing the consumer to physically see and touch the product and carry it home themselves. Internet market transactions, like mail order, involve risks associated with payment, distribution, and opportunism on the part of the seller, among others. Because of these risks it is common for consumers to incur certain costs to minimize their risk. This leads to the next hypothesis that higher prices may be paid by consumers to offset risk.

To test this hypothesis consumers were asked three questions. They answered the questions on a three point scale with 1 = false, 2 = unsure, and 3 = true. It was found that good deals make consumers more likely to buy again from the same seller if the price for the cards is similar to other sellers, and that they may pay a higher price to a seller if they know they can trust them. It was also found that consumers do not always buy items from the seller with the lowest price. Therefore the data support the hypothesis that price is not the only determinate of where consumers buy. Other non-price factors that were reported as important included seller reputation, past deals with the seller, and seller reliability.

The findings of this study have implications for businesses selling products through the Internet. Essentially, sellers can compete by offering the lowest price and/or they can compete using a strategy by which they differentiate themselves from other sellers because they are less risky (more trusted, better reputation) in the market. Competing based on reducing consumer risk, when the seller/buyer relationship is supported electronically, can be described as an electronic virtual partnership. Consumers and sellers become defacto partners where each side benefits from a long term trusting relationship.

This is interesting because it describes how, over time, sub-markets may form within the overall Internet market because consumer knowledge is limited and there is still a cost to gather information about new sellers. Over time, there are lower incentives for consumers to search the entire Internet market for sellers of a product that has been purchased in the past. Unless a seller's price is significantly lower than prices from a trusted seller, the potential risk inhibits the consumer from buying from the unknown seller.