Certainly you need to consider both supply and demand, and how fast both are growing. Ignoring any of these variables turns your prediction into luck. You need

Certainly you need to consider both supply and demand, and how fast both are growing. Ignoring any of these variables turns your prediction into luck. You need a way to assess their relative size. Looking at how successful others in the market currently are is actually a pretty good way to assess this. Except if their high margins may be based in large part on factors you can't benefit from, like network and monopoly (e.g. IP) advantages. That's kind of the case for all this kind of big-picture advice. There's some truth but some missing variables that may be just as important. "Bread crumbs" was a good way to describe it, I think.