We explore the effects of television advertising in the setting of the National Football League’s Super Bowl telecast. The Super Bowl is well suited for evaluating advertising because viewers pay attention to the ads, more than 40% of households watch the game, and variation in ad exposures is exogenous because a brand cannot choose how many impressions it receives in each market. Viewership is primarily determined based on local preferences for watching the two competing teams. We combine Super Bowl ratings data with weekly sales data in the beer and soda categories to document three primary findings about advertising. First, the relationship between Super Bowl viewership and sales in the week leading up to the game reveals the brands customers buy to consume during the game. We find some brands are consumed while watching the game while others are not, but this is unrelated to whether a brand ever advertised during the Super Bowl or advertises in a specific year. This rejects the theory that advertising works by serving as a complement to brand consumption. Second, we find that post–Super Bowl sales effects of ad viewership are concentrated in weeks with subsequent sporting events. This suggests Super Bowl advertising builds a complementarity between the brand and sports viewership more broadly. Finally, we collect data on National Collegiate Athletic Association basketball tournament viewership to test this theory and find that the complementarity between a brand’s sales and viewership of the tournament is enhanced by Super Bowl ad viewership. Together, these findings identify advertising as a determinant of why some brands outperform others for particular consumption occasions such as sports viewership.