Preview with sample data (July 10). Weights per Mitchell's priorities: Stock 45% (preventable outs as % of active sellers, 5% = a truly bad day, + how stocked the store is) · Checklists 35% (completion of scanned sheets; scan coverage is Mitchell's job, shown as info but never scored) · Return 15% — shown as actual people ("88 of 412 back"), scored on the repeat rate's slope across weekly snapshots (improvement, not level; sample data shows "↑1.1pp over 4wk") · Punctuality 5% (informational; 1 minute late isn't make-or-break). The breakdown line under the ring shows each area's 0–100 score so "why this number?" is always answerable; all weights and targets live in one config block.