Search Engine Advertising: Pricing Ads to Context

Each search term put into a search engine produces a separate set ofresults. Correspondingly, each of the sets of ads displayed alongsidethe results is priced using a separate auction. We investigate how bidsfor these context-based ads depends on the difficulty of making a match.This contrasts with the existing literature that focuses on the effectof match quality. We examine advertising prices paid by lawyers for 139Google search terms in 195 locations. Other things being equal, thefewer searches there are on a term, the higher the price. To identify acausal relationship between match-difficulty and prices paid, we exploita natural experiment in 'ambulance-chaser' regulations across states.When lawyers cannot contact a client by mail and matching becomes moredifficult, the relative price per ad click is $0.93 higher. We check therobustness of this result by performing a falsification test using adifferent ambulance-chaser regulation. Our results suggest that pricesare higher for context-based ads when the difficulty of both online andoff-line matching increases. This highlights that a major reason whysearch advertising is profitable is because its use of context canmonetize the 'long tail' by reducing friction in the matching process.