READ THE CLASSIC ARTICLES. Alex at Marginal Revolution asks, "In IO theory we teach our students that price discrimination requires monopoly power and monopoly power allows the firm to make above-normal profits. So why don't the industries that practice a lot of price discrimination seem especially profitable? Airlines, movie theatres, universities - all classic examples of users of price discrimination yet none seem especially profitable." Arnold at Econ Log addresses part of the puzzle: "My first reaction is that it is difficult for me to come up with an in industry where there is not price discrimination, which means the attempt to charge higher prices to customers willing to pay them. For example, software pricing is all about price discrimination. Microsoft tries to charge less to educational institutions and to home users than it charges business users. Drug companies try to charge less to groups that can afford to pay less." Quite so. Any coupon, or volume discount, or reward for loyalty such as a frequent flyer club, is an attempt to segment markets by providing an incentive for the low-marginal-valuation buyers to self-identify.
Turning to the question of profitability, let me offer two explanations. The first draws on a lot of ancient work on public utility pricing. In the presence of large sunk costs, marginal cost pricing does not provide sufficient revenue to cover the opportunity costs, including the sunk costs, of the firm. A two-part tariff, which is isomorphic to a coupon or a volume discount (too bad Alchian and Allen's University Economics is long out of print, if you have a copy, go and look it up) is one way of obtaining sufficient revenue. Under competitive conditions with differentiated products, there is no theorem that rules out a free-entry equilibrium with marginal cost pricing to marginal buyers and higher prices to buyers who don't qualify for the discount. (Monopolistic competition and spatial competition models with one price are characterization results.) Thus, movie theaters.
Second, and again drawing on principles of public utility pricing, often the two part tariff is supervised by a regulatory commission. Used to be true of airlines (where there might be network scale economies or irreversibilities leading to equilibrium existence problems that I'm not qualified to discuss at length) and is true of universities. Under regulation, another phenomenon, which transport historian George Hilton called "The Basic Behavior of Regulatory Commissions," kicks in. (American Economic Review Proceedings, May 1972.) What is that basic behavior? It is to generate cartel profits somewhere in the industry and then dissipate them elsewhere. Under the regulation of air fares, the high air fares on busy corridors went to support non-stop flights from numerous smaller cities to coastal destinations (I will admit to taking some pride in learning about such things as a boy in Milwaukee) or to provide better food in first class. The universities are nests of expense-preference behavior, but that's a subsequent rant.