The study of the economics of information was launched in the seminal work of Stigler [17] in 1961. In this paper, Stigler cites several examples of observed price dispersion, which he attributes to the costly search procedures that consumers face. Consequently, he notes the utility of trade journals and organizations that specialize in the collection and dissemination of product information, such as Consumer Reports and, of course, shopbots. Stigler reminds us that in medieval times, localized marketplaces thrived in spite of heavy taxes that were levied on merchants, demonstrating how worthwhile it was for sellers to participate in localized markets rather than search for buyers individually. Similarly, shopbots today serve as local marketplaces in the global information superhighway, and accordingly, we find sellers sponsoring shopbots and paying commissions on sales, as they essentially pay for the right to participate in the shopbot marketplace. This strategy encapsulates the current business model of Internet sites such as DealPilot.com that design shopbots.
Since Stigler, many economists have developed and analyzed formal
models that attempt to explain the phenomenon of price
dispersion. 
We have
already mentioned the work of Varian [19] and Burdett and
Judd [2]. The former study is a special case of our
model of shopbot economics in the case of two types of buyers, namely
type 1 and type S, existing in fixed proportions. The latter
authors, on the other hand, consider independent buyer decisions among
a set of search rules of fixed sample size. In this paper, we
integrate these two approaches by specifying a fixed minimal
proportion of buyers of type 1 while allowing all other buyers to
choose their sample sizes. Moreover, we extend these ideas with the
notion that shopbots themselves may well price for the information
services they provide, acting as economic agents in their own right;
this creates an additional strategic variable, namely the cost of
search, as determined by shopbots attempting to maximize their
profitability.
In the absence of shopbots and strategic search costs, a model of sellers' price adjustment was studied in Diamond [5], in which a somewhat paradoxical outcome arises: for any positive search costs, no consumers search and all sellers charge the monopolistic price v. More recent work on the dynamics of price-setting includes the evolutionary approach of Hopkins and Seymour [11], where it is argued that the mixed strategy game-theoretic equilibrium is dynamically unstable. This raises legitimate concerns about the validity of our assumption that f(p) describes sellers' behavior. On the other hand, since we also assume that sellers are represented by software agents, it is in fact plausible to suppose that these agents would price in this manner, as the computation of f(p) is a simple and well-specified algorithm. Another recent work of relevance which considers the potential impact of reduced buyer search costs on the electronic marketplace is Bakos [1]. This model is in some sense more general than ours in that it allows for product differentiation, but it does not allow for varying types among buyers. It remains to incorporate features of product differentiation with the buyer search strategies of the present model.