I won't start a new project without a good problem

I won't start my next project without defining the problem very well. Everything starts with a valid problem for which people (or even better if companies) really want to pay. Now my problem is to find a problem. Any ideas on how to open my eyes to problems?