Targeting Advertising Spending and Price on the Hotelling Line

Are competing firms better off with more targeting? We examine this question in the context of targeting advertising spending and price in the Hotelling model. By “more targeting,” we mean expanding the scope of targeting from one of these variables to both. When the incremental cost of such expansion is small, the question boils down to whether the firms are better off targeting both advertising spending and price (“full targeting”) or just one of them (“partial targeting”). We show that it is individually rational for the firms to expand the scope of targeting in such circumstances. However, their profits may or may not be higher in the resulting full targeting equilibrium. The paper identifies the conditions under which each type of targeting is collectively optimal. Targeting advertising spending only is optimal when the products are highly differentiated; however, when product differentiation is small, full targeting is optimal when the cost of advertising is small, and targeting price only is optimal when the cost of advertising is large. These results provide managerial guidance on how firms should respond to any restrictions on the scope of targeting imposed by regulators or online retail platforms. On the theoretical front, our results show that expanding the scope of targeting has benefits and costs for the firms. The benefits come from being able to align local advertising levels to local prices—what Dorfman and Steiner’s theorem says firms should do. The costs come in the form of more intensified price competition, which takes different forms depending on which type of targeting is added. When price targeting is added to advertising spending targeting, it has the effect of replacing product differentiation by informational differentiation; when advertising spending targeting is added to price targeting, it has the effect of delegating control of the local price equilibrium to the weaker firm. Comparing our results with previous results suggests that targeting on preferences may be fundamentally different from targeting on loyalty and switching behaviors .