An interesting idea from an old suggestion is [running our own usability studies](http://meta.ux.stackexchange.com/questions/355/should-we-run-our-own-ux-studies). It's not something I'm set up to do myself, but if we could make it work it would be *awesome* to have blog posts sharing original research, especially answering questions from the site that don't already have research-backed answers.

I don't want to sign us up for anything too huge, but it'd be easy to conduct some simple A/B tests, perhaps using Usara or other online services if no one wants to donate their lab.