The electronic market hypothesis (EMH) has found little support in research practice. Alternative hypotheses have been formulated and tested with more satisfactory results. However, today’s Internet successes (e.g. Amazon bookshop) seem to support the original EMH. This implies that the EMH may be correct in forecasting electronic markets, but wrong in determining the conditions under which these markets would arise. This article re-investigates the EMH conditions, and proposes a framework, based on marketing and transaction cost theory. The framework is used to score and compare some 50 electronic market sites on the Internet (books, cds and mortgages).