The other night, at LA Ruby, I ended up telling the same story to two different people about two different industries. Joe O'Brien of Edge Case asked me about my interest in acting, and we ended up discussing how Hollywood now basically runs on a small number of web apps which everybody uses and which would not impress any Ruby geek with their Ajax, their clean HTML, their fast response times, or their beautiful design. Later I talked with Pat Maddox about my interest in Internet marketing, and told him how that industry to some degree runs on a small number of web apps which everybody uses and which would not impress any Ruby geek with their Ajax, their clean HTML, their fast response times, or their beautiful design. Or, to be perfectly honest, with any other characteristic either.
I hate to write a blog post that seems to endorse the "it's Facebook, but for dogs!" mentality of so many silly entrepreneurs, but I could name a lot of totally crappy web apps raking in tons of money by providing trivial features, badly, to very specific audiences. The money lies in those audiences. These apps capture the network by providing exactly what the network needs, and then they make a lot of money. The winning part of "it's Facebook, but for dogs!" isn't the "Facebook" part, it's the "for dogs!" part, and really it's the "for" part, more than any other part, that brings home the bacon. It's probably better to do exactly what your customer needs, but in an incompetent way, than it is to do something which nobody needs, better than anyone would have ever thought possible.
The only problem is that we geeks fucking love doing something which nobody needs, better than anyone would have ever thought possible. I'm not going to hammer on these kinds of projects. It's difficult for me to criticize when the number one thing I'm known for is very close to science fiction. But let me just say that this is why the Agile Manifesto values customer collaboration; more than anything else, you need to know what your users want.
It makes me wonder about GitHub, actually. GitHub's a terrific site running on terrific code, and I'd always assumed its success came from this, but now I think the success comes from giving a niche market exactly what it needs, and the fact that it's a terrific site running terrific code is just a nice side effect of who the target market happens to be.