I examine the relationship between customer income and firm markups using rich data on household transactions and wholesale costs. Over the observed purchases, high-income households pay 15pp higher retail markups than low-income households. Half of the markup gap is due to differences in markups paid at the same store. Conditional on income, markups paid by a household also increase when a household shops in high-income areas, shops at retail chains with locations in other high-income areas, or purchases products with a high-income customer base. A model in which household search intensity depends on opportunity cost of time can account for these facts. Consistent with the model’s predictions, I document that retail markups across cities rise with both per-capita income and inequality. Through the lens of the model, changes in the income distribution since 1950 account for a 10–14pp rise in retail markups, with 30 percent of the increase since 1980 due to growing income dispersion. This rise in markups consists of within-firm markup increases as well as a reallocation of sales to high-markup firms, which occurs without any change to the nature of firm production or competition.