Hotelling's Law observes that in some markets it is rational to become as much like your competition as possible. (This near-zero-sum strategy does not apply everywhere, thank God.) This means that once goods and prices are identical between two stores, then the location of the stores becomes the product.
So imagine a town with two competing burger restaurants (McDonald's and Burger King) scheduled to be built. Where should each be located in order to maximize market share?
The at-first-glance best scenario might seem to be to place both stores at opposite sides of the town, which would neatly divide the town into the McDonald's customers and the Burger King customers. But this leaves each restaurant open to attack: Burger King could set up closer to McDonalds to "eat into" McDonald's territory. In fact, Burger King could move riiigght next to McDonald's. As long as Burger King is the closest option for most of the town, they'd claim most of the town's burger market. (Of course, the reverse applies as well.)
Now, let's imagine that these two competing restaurants are moved inward from the edges of the town. Instead of sitting on the perimeter, each restaurant is embedded right in the middle of each half. This reduces the average time that customers spend traveling to their local burger joint, and we've reached what's known as "Social Optimum."
Problem solved, right? Not really. While customers may be happy, we've just replicated (at a lower magnitude) the same risk that Scenario 1 poses. Burger King can still move closer to McDonald's to claim more market share, or vice versa.
The only perfect solution is to move both restaurants to the center of the town. Each now claims exactly half of the town's burger market, and—as far as location goes—there's nothing that either restaurant can do to claim even 1 percentage point more market share.
America's two-party political system runs on Hotelling's Law. Nine months before the general election, each party entertains a mix of potential candidates. And since primary elections are a reflection of sentiment within a party, candidates may campaign on more extreme platforms in order to curry favor from their party's voters.
Eventually, however, the chosen candidate for each party must start acting like a burger restaurant. Now competing for the votes of the entire US electorate, each candidate must begin their rush toward the middle, lest the other candidate swoop in to capture swing voters looking for the most reasonable candidate.
Plenty of good businesses obey Hotelling's Law and congregate around some common area—whether that's price, location, or product. Home Depot and Lowe's, for example, often share a parking lot.
But every outstanding business avoids the Hotelling trap and obviates the competition through innovation. In the town with the two burger restaurants, the truly great business would be food delivery by drone.