We analyze a toy dynamic value model for auctions, where the buyer’s valuation depends on the elapsed time since his last successful bid. That the current value depends on the outcome of the previous auctions is an overlooked aspect of repeated auctions in digital advertising. This dynamical nature stems from three facts: (a) ad slots are auctioned in real time, (b) an advertiser might repeatedly show banners to the same user, (c) the marginal value of winning an auction drops momentarily after a banner has been shown to the user. Facing dynamic values, the buyer must balance the immediate benefit of a current win against the potential decrease in future values. The toy model we introduce addresses a gap in the understanding of the implication of such dynamics. We demonstrate that traditional bidding strategies are suboptimal in this model and present an algorithmic method to derive the optimal bidding policy.