// The scoreboard · measured, not estimated
We score our own store the same way we score yours.
Every night Bilby walks our labelled demo store, Fernwick & Co., with the exact scan customers get — and the score lands here whether it is good or not. A red cell stays red until the next walk clears it.
| Our store · Fernwick & Co. (demo store) | 65 / 100 | 1 P1 | 0 P2 | walked 22 d ago | the report → |
|---|---|---|---|---|---|
| The corpus · real stores | 1739 walked | 61% had an issue | 83 avg / 100 | last walk 10 h ago | the data report → |
| This build's quality gates | 0 unsourced claims / 18 pages | 0 readability / 48 checks | 0 nav defects / 88 pages | 0 drift | as of 2026-09-03 |
How to read it: the demo store row is the same nightly walk every monitored store gets — score out of 100, findings counted by severity, report linked in full. The corpus row is every completed walk of a real store (the demo store is excluded from those numbers). The gates row ships with each build: unsourced digit claims, readability failures, and navigation defects found by the lints that gate deploys — the source files are in the open repo.