The economic concept of rationality seems inappropriate in the context of creative innovation, because of its assumption that the tastes and preferences of agents are fixed. The concept of copying, of imitating the behaviour of others, has equal claim to the description 'rational' in an innovative context. Models of 'binary choices with externalities' are predicated on copying and potentially show us not only why most innovations fail, but also why big social changes do not necessarily require big causes. In the 'Long-tail' world of a huge range of choice, however, many choices are not 'binary, either-or'. In the long-tail world, popular choices tend to become more popular, but not forever, as innovation drives a constant turnover in the popularity rankings. A very simple model of 'neutral' copying with occasional originality of choice can explain realworld patterns of long-tail distributions under continual turnover.