paulgraham.com, November 2012
The way to get startup ideas is not to try to think of startup ideas. It's to look for problems, preferably problems you have yourself.
The very best startup ideas tend to have three things in common: they're something the founders themselves want, that they themselves can build, and that few others realize are worth doing. Microsoft, Apple, Yahoo, Google, and Facebook all began this way.
When you try to think of startup ideas, you tend to produce ideas that sound plausible but are actually bad. I call these "sittin' around trying to think of startup ideas" ideas. They sound reasonable, but they're not ones that founders arrived at organically.
The problem with these ideas is that they're not grounded in a real need. They're solutions in search of a problem. And that is almost always fatal.
The best ideas almost always seem like bad ideas at first. That's what makes them available. If they seemed obviously good, someone would already be doing them. So you need an idea that most people think is bad, but that you know is good. This is a surprisingly narrow intersection.
The best way to generate startup ideas is to live at the frontier of some rapidly changing field and notice things that are missing. Live in the future, then build what's missing.
When something is annoying or broken, and you think "surely someone has built this"—and they haven't—that's a startup idea. When you find yourself writing a script or building a tool to scratch your own itch, that's a startup idea.
The reason this works is that the future arrives unevenly. The people at the frontier experience problems that the mainstream won't hit for years. If you solve those problems now, you'll have the solution ready when everyone else catches up.
There's a particular type of great idea that's systematically overlooked: the kind that involves a lot of tedious work. Most founders have a "schlep filter"—an unconscious tendency to avoid ideas that seem like a grind.
Stripe is the canonical example. Processing payments online was a massive headache. Every developer knew it. But nobody wanted to deal with the banks, the regulations, the compliance requirements. The Collisons were willing to slog through it. The schlep was the moat.
The best startup ideas are the ones hiding behind effort that others are unwilling to put in.