One crucial task for e-commerce systems is to help buyers find products that not only satisfy their preferences but also reduce their search effort. Usually the amount of available products is far beyond the upper limit that any individual could process by hand; thus product search tools are employed to generate target product (s) by eliciting the buyer's preferences and then executing some kind of choice strategies. We propose in this paper an extended effort-accuracy framework for measuring the performance of various choice strategies in terms of cognitive effort, elicitation effort and decision accuracy. The performance of a variety of basic choice strategies is further studied by theoretical analysis as well as empirical simulations. It shows that the performance of a given choice strategy is a tradeoff between choice accuracy and effort required from the users. The proposed framework also suggests a new efficient method of evaluating the user interfaces of e-commerce systems by analyzing the performance of the underlying choice strategies. Copyright 2005 ACM.