I’ve long suggested that modularity is a key element in the success of open source as well as a key element of success in Web 2.0. In a talk a couple of days ago at Eric von Hippel‘s Innovation Lab meeting held at the Googleplex, Ely Dahan of UCLA reported on some work with prediction markets (which he prefers to call preference markets [pdf]). It turns out that modularity is important there too. In one experiment, with various designs of laptop bags, the prediction market failed, but worked very well when they focused on individual features that distinguished the designs. It was interesting to see this same theme emerge in a very different context. Modularity breaks down a problem to the level where it can be grasped by a single individual or a small group. It also allows people to respond to features that matter to them rather than to the entire package.