I'll elaborate to answer what I think are the questions behind the question, since I answered them a hundred times in the last three days at the Business of Software conference and one more time won't kill me:

a) It makes bingo cards for elementary schoolteachers, which is one very specific problem a large group of people know they have, is close to trivial technically to solve, and for which almost all solutions suck. This makes it very achievable for one guy to tackle as a hobby project. (It eventually got to the point where full-time employment was my poor paying quirky hobby project.)

b) The business, as distinct from the application, works because over the course of four years I became very good at organic SEO and metrics-based optimization. These are broadly applicable skills in software (and marketing generally). Surprisingly few people can do them well.

c) No, I do not see myself being the bingo guy for forever. If you wanted to address a similarly small niche yourself, I would say it is a wonderful learning opportunity and has been lifechanging for me. I wouldn't recommend starting with bingo cards for elementary schoolteachers, because if you are capable of beating me at it your skillset is worth a heck of a lot more money than can ever be extracted from that market for that problem, but virtually any developer can find something similar and learn enough to make it work. The amount you learn from having your own experimental laboratory is better than grad school, and you get paid to study. I wouldn't particularly recommend that route if you take money from investors. Two models, vastly different lifestyles, pick your favorite.