Keyword prices are set by auction, Google can’t make these prices any higher, and therefore we shouldn’t really worry about Google’s market share because they can’t hurt advertisers. This syllogism, which seems to be put forward with some frequency these days, was behind Randall Stross’s defense of Google’s deal with Yahoo. Here’s why it’s wrong. Search engine providers can engage in what economists call “perfect price discrimination” with advertisers. In markets without price discrimination, including ones held by a monopolist, price is determined by the marginal consumer who would cease buying if the price were higher (that’s given by the marginal revenue equals marginal cost condition for profit maximization). Consumers who buy the product generally value it more than the price set by the marginal consumer (economists call these infra-marginal consumers). With perfect price discrimination, on the other hand, every consumer is charged the most they are willing to pay. Rather than setting a single price (where marginal revenue equals marginal cost)the firm sets an individual price for every consumer (where price equals the most the consumer is willing to pay). Profits are usually a lot higher, obviously, with perfect price discrimination than when a firm charges a single price.

That more or less describes the advertiser side of search because the keywords are auctioned off (and in reality the price discrimination isn’t perfect but it is still probably pretty good). If the story ended there we would have a market in which any provider can make tons of money.

But there’s another side to this market and that is the search user. In these kinds of two-sided markets what typically happens is the profits on one side get competed away on the other side. American Express, for example, also engages in close to perfect price discrimination against merchants. But its merchant fees get competed away through competition for cardholders—think all those rewards. In search engines we would expect that the providers would compete for users to get those advertiser-side profits—and so they have: Microsoft and Yahoo both have programs to pay for search users (check out FreeCause, for example, whose business is based in part on helping Yahoo do that.)

If we end up with no competition in search then this competition for users would probably cease, or decline quite a bit, and the monopoly profits on the advertiser side wouldn’t get competed away. There may be other reasons why we should let competition just take its natural course, and if Google ends up on top so be it, but it is quite wrong to suggest that it doesn’t matter. It sure does—to consumers.