One data point is not generalizable. I was very successful building enterprise products, and I never figured out how to build anything for the masses.
I disagree strongly with your conclusion. If anything, save yourself some agony by building something for a market that you know well, and are interested in.
This is what I hear from everyone besides Ed that I've ever heard on the subject: It's easier to make money with a product that appeals to a few rich people than with one that depends on large numbers of average people buying in.
Usually what I've seen is that successful products are often ones that let average people become richer. ViaWeb, DropBox, AirBnB, AdSense, AdWords, EBay, Etsy, Kickstarter, Google Apps - these are all B2B2C products that let ordinary people make some small amount of money doing things that are quite within their grasp.
It's easier, but what I've noticed with many companies that follow that model is that a handful of clients(maybe even just one) end up controlling most of the revenue. That sort of buyer power is not good for any company's long-term prospects.
I disagree strongly with your conclusion. If anything, save yourself some agony by building something for a market that you know well, and are interested in.