Here's a simple way to validate your startup idea: take a piece of paper, and write out the names of 5 people that need your product.
Just 5 people.
This is an effective test, because it makes it real. If you built this thing, who would buy it?
Can't think of 5 people? That's a red flag. Are you targeting a product at a group of people you don't yet have a connection with?
"Having a connection with people is really important. A lot of people build tools without thinking about the people that are using them." - Alex Hillman
Eventually, when you're bigger and more successful, you might have anonymous customers - people that find you on the web and click the buy button without any social interaction.
But at the beginning, when you're deciding what to build, knowing a few potential customers by name proves something. It shows that you're involved in their community; that you've been listening to their problems. When you can name your first customers, the people you know need a solution, you can build your product with confidence.
(When I say "know people by name", this includes the online personas of folks you've met through forums, Twitter and other online communities.)
Recently, Nathan Barry gave me some good writing advice that also applies to products:
"To write good content that people care about, write to one person. A specific person. My book Designing Web Applications was written to my brother-in-law, Philip. I wrote what I knew would help him. If your writing is truly valuable to that one person, your ideas will be valuable to many."
Make sure your idea connects with a genuine human need. Before you get too excited about building something, think about who you're building it for.