Platform competition in online knowledge search markets

This dissertation investigates to what degree the characteristics of online content influences platform competition in two online knowledge-search markets: one is a market for knowledge-sharing platforms and the other is a market for search engines with a knowledge-sharing service.;The first chapter explores the dynamic competition between two knowledge-sharing platforms. In the past decade, several knowledge-sharing platforms, such as Yahoo Answers, Microsoft Live QnA, and Naver Knowledge In, have emerged. These websites offer members platforms to share their knowledge by asking and answering questions. As a result of knowledge sharing between members, knowledge-sharing websites have an accumulated knowledge database of answered questions that attracts people who have questions. This research asks how long the knowledge-sharing website, with the largest knowledge database and membership, can sustain a leading market position. To answer this question from a two-sided market perspective, this study investigates the competitive effect of accumulated knowledge, the complexity of questions, and the price structure of advertising dynamics on the long-term viability of a latecomer to an online knowledge-sharing market; we numerically solve an open-loop differential game in which two knowledge-sharing platforms compete. We show the longer the life span of accumulated knowledge, the higher the chance of survival of the latecomer. In addition, we show how the general complexity of questions asked and the price structure of advertising dynamics impact optimal membership fees.;The second chapter examines the static competition between an inferior search engine with a knowledge-sharing service and a superior search engine without a knowledge-sharing service. This research is motivated by the contrasting impacts of a knowledge-sharing service on search engine competition in the US and South-Korea: Google has been steadily increasing its market share in the US even though Yahoo, the main competitor to Google, began developing a successful knowledge-sharing service in 2005; on the other hand, the leading South-Korean portal Naver developed a successful knowledge-sharing service in 2002 and has been strongly dominant since then. This study investigates (1) under what conditions an inferior search engine with a knowledge-sharing service can outperform a pure superior search engine and (2) whether the inferior search engine should close its database of answers or not to increase its market share. We find that when online content is limited and searchers are patient enough to wait for answers from other members, the inferior search engine can outperform the pure superior search engine. The answer to the latter question is that the inferior search engine is well advised to keep its database of answers open only when its search technology is below a certain threshold, which increases in the amount of information on the Internet.;The third chapter considers the same issue of the second chapter in a dynamic setting in which searchers are myopic about their choice of search engines and the amount of searchable information changes over time. This research specifically asks when and to what degree an inferior search engine can benefit from introducing an online knowledge-sharing service. By numerically solving a differential game between these two search engines, we show that the inferior search engine generally benefits from its knowledge-sharing service, and the degree to which the knowledge-sharing service helps the inferior search engine to enlarge its market share increases as the complexity of searchers' questions increases and the amount of online content decreases. Our most important finding is that the inferior search engine may attract more searchers by opening its database of answers, but it tends to increase its market share by closing its database of answers unless its search technology is much lower than that of the superior search engine. This result, to some degree, accords with that derived from a static model of the second chapter.