Theory predicts a positive relationship between market concentration and profitability in most scenarios. In empirical work, however, this relation is frequently not found or only a weak connection is observed. We compare the performance of concentration and market share variables, which are generated on the basis of the official industry classification, with information collected directly from firms. Information from companies on the number of competitors, their relative size and the intensity of price competition is highly significant in explaining profit levels, while none of the concentration indices performs well. Hence, the poor quality of industry data is responsible for the loose connection that is usually found between concentration and profitability.