Well that depends on the market and the product, doesn't it? The investment required to produce or consume a Coke, Cherry Coke, Vanilla Coke, or Diet Coke is extremely small and the potential market extremely large. Neither of those apply to RPG campaign settings. Coca-Cola can produce gallons of these products in seconds, I can consume them for well under a dollar and in a few minutes time and there are hundreds of millions of people like me. Yet would you believe that until Coca-Cola actually invested substantial research and then test marketed this, they were actually worried they would split their market? It's true.
WoW and Diablo cost substantially more to product and consume, the markets aren't as large as for Coke, yet they too aren't all that difficult for their users to play as consumers. Yet, even so, I'm sure their relative costs have an impact since they are also more expensive. Chances are there actually are some people out there willing to pay for one but not both and that's more noticeable than with all of the choices available in sweetened drinks.
Now let's look at RPG settings - MUCH more expensive to produce (relatively speaking) and by smaller companies and directed at a market orders of magnitude smaller. They take hours to read, prepare adventures in, and play. Not just a few minutes as in drinking a soda and not with such easy prep and cleanup as a computer game (though those too may take hours to fully play). And that investment generally has to come from multiple people at the play end, compounding the time I'm spending as a consumer. Those factors are probably going to lead the analysis in different directions than for Coke. Moreover, that analysis was done, as I pointed out, when WotC analyzed TSR's problems as part of their due diligence in deciding to buy the company and save D&D from years of bankruptcy limbo. It may be that WotC could credibly support more than one setting (though I really doubt they could support the same number as TSR did) but they have apparently decided the business case for doing so isn't strong enough for them to do it... at least not yet as far as we know.
Ultimately, you can't just look at strongly different products, assume they don't split the market, and then assume the same applies to the RPG industry - even the biggest dog in the industry. And you really can't just brush off the analysis they conducted, using the data they were able to collect, by doing so.