If you've built a gazillion stuff like me, you'll notice a trend of functionality that you find yourself building over and over again. Perhaps it's time to abstract that away into a separate product or service that many of your projects can end up using. Improve it and you'll improve all your projects in one swing.
So here's a personal anecdote of mine. I've gotten pretty good at crawling the web mainly because alot of the stuff I build deals with it. I love doing it because I'm attracted to working with lots of data but what I don't love about it is having to re-build my crawlers for every new project I work on! So I've recently side aside time to make it into a hosted service. And hey, maybe it can be even be a product.
But take everything with a grain of salt. It's easy to get into the trap of prematurely doing this for everything. There are just some things not worth abstracting until you've grasped a really good handle on it. Only do this once you've become a subject matter expert on it.
I also think this is a really good way of coming up with startup ideas. Good ideas come from immersing yourself within an industry and experiencing a problem first-hand, especially if you're trying to do something in the B2B market. Bad ideas usually come from assuming a market has a problem that needs solving. Do your market research like encountering first-hand what needs to be fixed and you'll increase your chances of success.