I have a spreadsheet with hundreds of startup ideas. I've also tried to start several companies that were premised on ideas that, in retrospect, were bad. Now I have a simple framework – two tests that an idea needs to be pass before I consider it more seriously.