Every lead vendor says their data works. We publish ours. Each month we freeze our top-scored homes, wait 60 days, then check them against recorded deeds. The results go here, good or bad.
North Carolina, coverage-matched back-test (July 2026): new-build homes in our high-appreciation ZIP set resold at 13.83% over 2024-2026 versus 10.31% for the same kind of homes outside that set. Same vintage, same counties, investor-owned homes excluded, restricted to counties with at least 80% deed-history coverage so the comparison is fair.
What this means honestly: homes our signals favor sell at a meaningfully higher rate. It does not mean any single home will sell. Predictive prospects, not guaranteed sales.
2,198,504 scored homes frozen on 2026-07-12. Grades against recorded deeds on September 10, 2026. The numbers publish here whatever they turn out to be.
2,198,671 scored homes frozen on 2026-07-20. Grades against recorded deeds on September 18, 2026. The numbers publish here whatever they turn out to be.
1. On snapshot day we freeze the full list of homes our engine scores 3-star and up across NC and FL.
2. Nothing gets edited after the freeze. The list is what it is.
3. 60 days later we match every frozen home against recorded sales, using stable parcel identity.
4. The per-tier hit rate publishes here, whatever it says. Sales are verified against public deed records
with plausibility checks; a recorded sale is the only thing we count.
Updated July 26, 2026 · Generated automatically from the same nightly engine that scores the leads · sellersignal.io · Methodology