The problem is that most startup books tell you what not to do, not what to do. Actually finding an underserved market that can be serviced by something a founding team can build is a bit of a black art. Anyone who finds one is busy building that startup, because the financial returns of actually serving the niche are so much better than that of telling someone else about the niche.

Before anyone mentions Lean Startup or Startup Owner's Manual - I've read them, lent them to friends, actually put them into practice, and that startup (startups, actually) still failed. And I've got a half dozen or so friends and acquaintances who have done it too, and their startups all failed too. The advantage of that methodology is that you fail fast, oftentimes before you quit your day job, but if founding a startup is something you really want to do, knowing a few dozen ideas that aren't viable businesses is only marginally useful.