Look at user behavior
Fred Destin writes about Research In Motion, makers of the Blackberry, in the blog post Getting it completely wrong. Very good points.
"So how wrong can you be ? It turns out that a company controlling the device and the server could actually do a few things well, such as deliver on the fundamental requirement of its user base (absolutely reliable email access on the move) and getting the rest right over time (such as the phone).
My lesson from this is that what may seem to be sound strategic thinking from a VC standpoint (absolute focus on one core area of expertise, search for extreme scalability, natural preference for software and "atom-free plays") can lead you to overlook the best companies, those that go all out to crack the user's need in an imperfect but perfectly appropriate way. And yes, we passed on TomTom too...
Another useful reminder that scrutinising user behaviour and acceptance is the best yardstick for determining which companies will succeed, rather than one's grand opinions about long term market evolution or deep theories about the future of tech !"