Kev Quirk ran a one-question survey recently about how people discover and read his blog posts, and I think it's a pretty inspiring idea. Sure, consent-based surveying introduces biases... but what analytic technique doesn't, really?