Homepage after-hours tours stat: the measured 41%
38% with an unbacked “vs team-only coverage” line becomes 41% booked by Clara, nights and weekends (Oct 8 meeting). Results row and Leasing row.
11 Oct 2026 · branch copy/home-stats-measured · captured against propflow-stage · at 58ede7cda54+local
Data note. Every name and balance below comes from propflow-stage, which is anonymized by construction (ADR-0097/0110 — phones in the impossible +1000 NPA, emails @example.test). The identical capture against prod would carry real tenants' names and balances and is not safe to attach anywhere, which is why the generator refuses to run against one.
1 · Before (main): Results row, desktop
Today's homepage, captured from main's parent commit (the homepage files are identical on main; the capture tool treats the tip of main itself as unusable for a before). The third card reads 38% with vs team-only coverage under it. Nothing backs that comparison: only 8 tours in prod were ever booked by a person.
Asserted in the captured DOM — Tours booked after hours · vs team-only coverage · 38%
Sections 2–6 and notes
2 · After: Results row, desktop
The third card now reads 41%, the measured share of tours Clara booked outside Mon to Fri 9am to 6pm local time (43 of 106, Apr 11 to Oct 10, two live customers; the Oct 7 review had 41 of 99). The sub-line says who booked them and when, and makes no comparison. The other three cards are unchanged on purpose: the reply time and the 92% wait on a decision (see the PR), and the NOI figure was left as it is in the Oct 8 meeting.
Asserted in the captured DOM — 41% · Tours booked after hours · Booked by Clara, nights and weekends · 47s · 92% · 15–21%. The new figure and line are on screen, and the three cards this PR leaves alone are still there.
3 · After: Results row, phone (390px)
One column on a phone. The longer sub-line wraps inside its card and nothing runs off the side.
Asserted in the captured DOM — 41% · Booked by Clara, nights and weekends
4 · Before (main): Leasing row, desktop
The Leasing row shows the same record, so it carried the same 38% and comparison line.
Asserted in the captured DOM — From first inquiry to signed lease · 38% · vs team-only coverage
5 · After: Leasing row, desktop
It changes with the Results row because both read one record in claims.ts, so the two can't drift apart. The Solutions section still prints one Projected outcomes; measured customer results coming soon. line under all four rows (not in this crop). For this now-measured figure that undersells rather than oversells it. Switching to a measured wording is Sean's call.
Asserted in the captured DOM — From first inquiry to signed lease · 41% · Booked by Clara, nights and weekends
6 · After: Leasing row, phone (390px)
The same row on a phone.
Asserted in the captured DOM — 41% · Booked by Clara, nights and weekends
What this PR leaves alone, and why
Reply time (47s). The meeting agreed on “median reply under 30 seconds, 24/7”. Re-measured on 2026-10-11, the median first reply is 21s from April to October and 25s over 90 days, but 36s over the last 30 days and 34s over the last 7. That holds for both customers. So “under 30 seconds” is not true of the product today, and the figure waits for a decision.
92% qualified on first touch and 15–21% NOI lift have no source. In the Oct 8 meeting the team kept both slots for now, until Sean picks replacements.