Last weekend, I went to Lean Startup Machine DC</a>, which is a weekend-long workshop on applying the principles of Eric Ries' book The Lean Startup</a>. The weekend is incredibly intensive, and we made a lot of progress validating and invalidating some of the assumptions we had for our local peer-to-peer lending platform, Clovest</a>.
For me the biggest take-away is the importance of talking to your customers early and often. Instead of writing code and building the website, we focused completely on customer interviews, market research, and developing a trial ad campaign to see how interested the community is in this project.
Far too often, technical people (myself included) would rather spend a few hours building something in Rails, or whatever is the current coolest technology, instead of talking with end users. Building software has instant gratification, it feels productive, and the risk is minimal (besides wasting time and getting carpal tunnel syndrome). Talking to people is difficult and uncomfortable, and forces you to consider the possibility that they will reject your idea. However, wouldn't you rather have your idea rejected while its just a sketch on a napkin and not 20,000 lines of code that took 200 man-hours to write? I'd definitely prefer the former, and that's probably the biggest takeaway from the whole lean startup methodology.