How does geography influence financial services provision? Using a new data set linking life insurers to agents across 19 US states, I document 4 stylized facts: (1) insurers operating in relatively high income and high mortality counties set higher prices; (2) there is positive sorting between local income and insurer size; (3) agents are important for generating premiums, with agent shares explaining 20% of within-state market shares and 37% of within-insurer market shares; and (4) the number of markets an insurer operates in is increasing in their size. Motivated by these facts, I build a stylized model in which insurers set a uniform price to optimally serve multiple locations. Prices can be decomposed into two components: a markup component and a cost component. Consistent with the data, stronger sorting toward high income locations generates higher markups, but can lower the cost component when income and mortality rates are negatively correlated. In future work, I intend to explore how the rise of digitization and online platforms can mitigate the importance of geography and what the implications are for the spatial allocation of financial services.