Economists distinguish two routes by which a firm can charge prices above marginal cost. The first relies on differentiation—genuine quality or features rivals cannot match. The second, less intuitive, relies on "switching costs": the time, money, or risk a customer must bear to move from an incumbent's product to an equivalent rival's. Where switching costs are high—relearning an interface, re-entering data, breaking integrations—an incumbent can price a notch above a rival's offering and still retain the customer, because the rival's lower price does not cover the customer's cost of switching. A subtle implication follows. The incumbent's pricing power is bounded not by what the product is worth in the abstract but by the size of the switching cost itself: it can extract roughly the switching cost, no more, before the customer defects. This bound has a counterintuitive consequence for new entrants. To win a locked-in customer, a rival must not merely match the incumbent on price; it must subsidize the switch, offering a discount at least equal to the switching cost, which is why challengers in such markets often post early losses by design. Yet the moat is not permanent. Switching costs decay when standards become interoperable or when data becomes portable by regulation, and an incumbent that has grown accustomed to pricing against a wide moat may be slow to recognize that the moat has narrowed—continuing to price as though customers were locked in after the lock has quietly weakened.