How a computer-science side project became a bootstrapped survey empire - no venture money, millions of users, and a stubborn belief that research should be cheap enough for anyone to run.
