[JohnHammersley]

One good route is to solve a problem you're experiencing yourself (or even better, one that you and some friends / colleagues are all experiencing).

It's how our startup came about -- we built something which did what we needed, and it turned out that a whole load of other people had the same problem (unsurprising with hindsight, but the point is we set out to solve a problem, not to "create a startup")

Happy to share more thoughts if you're interested.

[jamesbritt]

Counterdote: One company I was with created a product to solve our own problem (time-tracking for projects) and while we loved it we did not find anywhere near enough other people to buy it.

If you're going to be building something anyway, even if there's no market for it, that's fine. If you simply assume that a given solution to a problem that happens to please a handful of geeks is going to please many, many more, you may be in for disappointment.

In retrospect we screwed up by not first validating the market for what we were so sure people would want.

A corollary here is that, for all I know, people would have loved the product if we pitched it to them correctly. Being unable to sell/market a good product is as bad as having a dud offering.

In any event it is far from a given that scratching your own itch means you have market validation.

I've become increasingly convinced of sales+marketing first, serious product building after validation of the idea and of the ability to sell it.

[akg_67]

I second this approach. My current side project, a data analytics web service, born out of similar process.

1. I was downloading data on regular basis and manually analyzing data using Excel.

2. My wife wanted to analyze the same data but she didn't like my manual approach. And, asked me to automate the process, so I did and setup a system at home.

3. Then, she wanted to be able to remotely access the system at home. So I setup a VPN to access home system.

4. Then, some of her colleagues and friends wanted to use same system. So I setup a website with same system and user authentication.

5. That's how my web service evolved and currently has 1,000+ registered users.