Conclusion

This article considers a paid placement strategy for search engines. On the one hand, paid placement appears to be a financial necessity, embraced by most major Web search engines. On the other, paid placement can hurt the search engine's market share and its potential for revenues brought by users. We have developed a mathematical model for optimal design of a paid placement strategy, examined this tradeoff and analyzed sensitivity of the placement strategy to users' perceived disutility, the service quality of the gatekeeper, and the advertising rate. Our preliminary results are as follows. We show that the negative impact of paid placement on users causes the search engine to set paid placements at a below-ideal level. However, when disutility for paid placement is quite low (though not zero), the search engine can maintain its ideal placement revenues. We find that an increase in the search engine's quality of service allows it to improve its utilization of paid placement, moving it closer to the ideal; this also increases surplus for all players. However, an increase in the advertising rate motivates the search engine to increase market share by reducing further its reliance on paid placement and fraction of paying providers. As consumers get a better understanding of the factors underlying paid placement, the search engine would likely need to spend heavily on marketing campaigns in order to minimize users' perceived disutility for paid placement. While this research is set in the context of Internet search engines, our model and results apply more generally to many other contexts that share similar characteristics as search engines. This broader category is often called information gatekeepers, that intermediate between a set of users (or buyers, or consumers) and a set of products (or content providers, or vendors). Baye & Morgan [#!Baye-Morgan-2001!#] argue that modern markets for information tend to be dominated by ``information gatekeepers'' that specialize in collating, aggregating, and searching massive amounts of information available on the Web - and can often charge consumers, advertisers, and information providers, for their ability to acquire and transmit information. Wise & Morrison [#!Wise-Morrison-2001!#] emphasize the increasing role of information gatekeepers in today's economy, noting that in business-to-business markets, ``value has shifted from the product itself to information about the product.'' Specific categories of information gatekeepers to which our work applies include recommender systems (e.g., at Amazon.com), comparison shopping services (e.g., mySimon.com), e-marketplaces and exchanges (e.g., FreeMarkets), and more traditional information gatekeepers such as investment advisors and television networks. Like search engines, many information gatekeepers generate user-based revenues, but also seek to obtain revenues from their provider-base by offering some form of preferential placement. For example, some Internet booksellers are influenced by advertising fees in determining their bestseller lists. Similarly, certain Internet exchanges provide preferential service (such as real time notification or favorable recommendation to buyers) to some clients in return for higher fees.

We are pursuing extensions of this work, including a formal derivation of the optimal bias, generalization of demand assumptions, and elimination of free placement by the gatekeeper. Our models can be extended to examine conditions under which the information gatekeeper will begin to charge users, and specifically the case where the gatekeeper differentiates between users by offering two versions: a fee-based premium service with no bias in the query results, and a free basic version with paid placement bias. The fee-based premium version will bring additional user revenues to the search engine, however it may reduce placement revenues because paid placement becomes less attractive to content providers. In addition, the search engine's market coverage and placement fee may change as well, and the models can be used to determine if it is optimal for the gatekeeper to offer differentiated service. Similar models can be developed to examine the impact of differentiation based on advertising. Some search engines have already began to offer fee-based premium search services that contain no advertising. If this is the trend, it may eventually change people's view of Internet search engines as a free resource for fair information.