Here's part of an interview from the FailCon archives way back in 2009.
Diane:What companies come to mind when you think of failure?
Chris:I worked for a company called Vidoop (which you might remember). To me they represent the epitome of failure, and of failing badly. I think Bear Sterns and Lehman Brothers are also up there. More locally, a community-support agriculture startup called MyFarmSF just pulled the plug... after taking subscriptions from a bunch of people and then being unable to pay them!
Diane:What are common mistakes you've seen at (pre or post-revenue) companies? And how could they have corrected those mistakes?
Chris:I think failure to focus is a big one. I wrote about Vidoops' failure to focus (http://factoryjoe.com/blog/2009/06/05/the-fall-of-vidoop/).
I also think that trying to grow too quickly or focusing on "shallow growth" can be killer... Of course you need to get momentum, but it's kind of like those schemes that offer you 10,000 Twitter followers... it's like, if those 10K Twitter followers are all bots, does that help you succeed? Probably not! So, I think building what Umair Haque calls "thick value" is important because it requires you to connect with your customer base.
I also think that you need to have external validation of what you're doing — I see a lot of engineers who have a specific problem and therefore believe that everyone has the same problem and that's it's as salient to the next guy as it is to them. This is often NOT the case (to be fair, I fall into this trap sometimes). On top of this, timing is also very important — many technologies are simply too early and fail because the market isn't ready yet. Knowing how to time your solution is probably about as important as location if you have a physical business like a coffee shop.