Abstract
Consider the problem faced by a decision maker (DM) who must select the order in which to evaluate the unknown alternatives displayed by an online search engine. DMs do not know the distribution of the realizations that result from clicking on an alternative and must therefore account for the differences that may exist between the ranking provided by the engine and their subjective potential evaluations. The current paper formalizes the information retrieval incentives of DMs through a combinatorial function that incorporates the position of the results displayed by the search engine and the order in which they are evaluated. This function defines a benchmark framework measuring the ability of search engines to identify the preferences of DMs and the capacity of the latter to assimilate and evaluate the information provided by the engines. We compare the cumulative frequencies derived from the implementation of evaluation models of varying complexity with the average traffic shares and click through rates of the alternatives ranked within the first page of Google results. A seemingly paradoxical result is obtained when simulating the information acquisition behavior of DMs after performing an online search, namely, artificial agents require more complex combinatorial abilities than actual DMs to approximate better the heuristic choices of the latter in online evaluation environments.