Exploratory protocol chronology: An initial two-address example preceded this county audit. Before inspecting expanded county results, the analyst fixed the Franklin County prefix 39049 frame, largest positive land intersection association, identifier tie-break, usable direct B19013 estimates and MOEs, cross-ZCTA unordered pairs, strict ordering reversal, interval nonoverlap screen at both levels, and 90%/95% land-share subsets. After seeing initial county results, an explicitly post-result robustness check gave equal weight to each distinct ZCTA pair, omitting empty groups from the separated mean. No selection based on reversal magnitude. Every census tract in the 2020 relationship frame was assessed. Two unavailable income estimates were excluded. Land shares do not allocate households. Shared tract/ZCTA estimates create dependent pairs. Neither tract nor ZCTA is ground truth for commercial catchments, and the results are not business-demand or performance evidence. Source vintage: 2020 geography and 2020-2024 ACS five-year income in 2024 dollars. Calculation date: 31 October 2026.