We analyse a model of oligopolistic competition in which consumers search without priors. Consumers do not have prior beliefs about the distribution of prices charged by firms and thus try to use a robust search procedure. We show that the optimal stopping rule is stochastic and that for any distribution of search costs there is a unique market equilibrium which is characterised by price dispersion. Although listed prices approach the monopoly price as the number of firms increases, the effective price paid by consumers does not depend on the number of firms.Uncertainty lies at the heart of any optimal search problem. Most of the search literature reduces uncertainty to risk, by assuming that options are sampled from a known distribution. In the consumer search literature this risk comes from mixed pricing strategies employed by firms (Stahl, 1989;Janssen et al., 2005), or from a random component in utility (Wolinsky, 1986;Anderson and Renault, 1999;Armstrong and Zhou, 2010). In the recent literature, a more sophisticated approach is often employed: consumers are unaware of some parameters of the model and 'estimate' them in a Bayesian way using the observed prices (Dana, 1994;Janssen et al., 2011). All these models require consumers to have certain 'priors' -distributional assumptions over prices, utilities or other parameters of the model.As the consumer search literature deals with situations where consumers do not know prices, it is natural to assume that in such situations they also do not know how many firms operate in the market nor what those firms' costs and market shares might be. Therefore, in this article, we propose a robust search procedure, where the solution of the optimal stopping problem does not rely either on priors about parameters of the model, or on the market model structure itself. As consumers are agnostic about the distribution of options from which they sample, they cannot base their decisions on traditional expected utility. In our study, consumers use the optimal stopping rule which minimises the loss in comparison with an (imaginary) informed searcher (who knows the distribution of the prices) in the worst case scenario. By doing so consumers are guaranteed that the expected difference between their utility and the utility obtained under ex ante optimal behaviour is below a certain bound.In our study, we look at a market with a homogenous good in which consumers search for the best price. However, our results can be applied to a wide variety of