I'd like to think that most successful founders fail at their first few not so noteworthy "side-projects" before finding success and sticking to it. Just building a crappy version of something not-so-cool can teach you a lot about tech, business and design, which then helps you build better products in the future.