The FCC plans to assign up to $20B in subsidies for rural broadband coverage through the Rural Digital Opportunities (RDOF) program. For the RDOF Phase I Auction, it plans to use a multi-round reverse auction design based on the design used in the Connect America Fund (CAF) Phase II Auction in 2018. The CAF II auction was, in many ways, a success, assigning $1.49B in support over ten years to provide fixed broadband and voice services to over 700,000 eligible locations. However, CAF II was only the first application of this auction design, and it would be wrong to think that the design is already as good as it can be. After all, the FCC’s forward auction designs have evolved to the point of barely resembling the original designs first used over two decades ago. Each point in that evolution took advantage of the data and experience of prior auctions to make efficiency-enhancing refinements. Similarly, the CAF II auction data and experience are available to help consider and examine potential refinements to this reverse auction design. In this paper, we analyze the CAF II auction to better understand its outcomes and why they occurred. The result of this analysis is a set of proposals that are likely to improve the fairness and efficiency of the RDOF auction. 1. Use CAF II bidding data to update the CAM-driven reserve prices for RDOF in an effort to make more efficient use of the budget. 2. Freeze the price clocks for uncontested areas after the budget clears to facilitate package bidding and the revelation of cost synergies. 3. Modify the information rule to avoid switching bids that are only intended to gather information on the level of competition and hurt competition discovery. 4. Reduce location uncertainty and its negative effects on efficiency and fairness. 5. Use the CAF II performance weights to represent consumers’ preferences and the proposed high latency weight to avoid paying for areas that would be served anyway.