This is interesting in a fake world with flat demand. In the world we actually live in, the world is <i>very</i> lumpy, and information about the disposition of lumps is not exactly trivially available, but information about the disposition of competitors is.<p>There are boring real-estate specific reasons why having two coffee shops on the same intersection is very unlike having one very big coffee shop at the same intersection, by the way, chiefly with regards to traffic flow. Traffic flow is a key consideration for "non-destination" commercial stores, like Starbucks. You don't get in your car to go to Starbucks, you get in your car to go to work, home, or your "third place" for the day -- perhaps that swanky restaurant you're visiting with the missus. (You will note that work/home/third place are statistically unlikely to be on corners, principally because there is a price premium for corners, given that they are commercially much more valuable for non-destination commerce than mid-street.)<p>Any Starbucks which catches your fancy while you're in your car can catch your business, but they're overwhelmingly more likely to be a) directly on your path between A and B and b) require a minimum of complicated driving to deviate from your path to their cut to the street (i.e. you statistically prefer a right turn straight into a store on your right versus doing a U-turn onto the other direction of a busy street, etc).<p>Given that you want simple turns into your place of business, it's trivially obvious why one would prefer having two streets for ingress/egress versus one street, which is only available on a corner. (Fun driving game: try to find any retail establishment on a corner which <i>is</i> a location that the average customer would specifically get into their car to go there. Fun variant: attempt to convince them to sell their property to you, because you're already 50% of the way to flipping it to a bank/fast food restaurant/WalGreens profitably.)<p>There's another wrinkle: all corners are more valuable than non-corners, but all corners are <i>very much</i> not equally valuable. How valuable a given corner is for a particular business is a <i>very hard problem</i> to solve for. Some firms and individuals are very good at solving for it. A <i>very effective</i> way to piggyback off of their expertise is to simply grab the corner opposite of them: while the lumpy world of geographically distributed demand for coffee is very non-trivial to access, the simple rule "If you see a Starbucks, to a first approximation, siting our next coffee shop across the street from it <i>will</i> be profitable" generally works.