Production surface — everything below is what Circuit would show. This fold is for the presenter.

Talk track. This is the checkout page's store-to-store stat, expanded to the whole district. Two different answers live here: Store 8 is the district's fastest, and Woods Cross is drifting from its own normal. A store can be the slowest in the district and still be healthy — Store 12 runs 36s every week, and it stays green. Response time is the second checkout discipline: Woods Cross drifts on both, Store 21 on serving only — two different conversations for the same red week. The lunch window is the evidence under the Store 8 callout: the flat store absorbs the rush, the spiking store makes lunch customers pay for it. The unit of value is the operating practice worth copying, not the ranking. Their words: store to store, then district to district — then it becomes competitive.

Claims. Numbers are illustrative and match the checkout page (Woods Cross 39s, district median 33s, Store 8 best at 28s; the 41s response matches its median response). Serving time per customer is serving seconds ÷ customers served, presence-measured. Response time is the median from a customer waiting at the counter to staff present. The lunch window is the same serving metric cut to 11–1 — a daypart comparison, not queue depth (queue depth is not built; the checkout page's queue-by-daypart and response histogram were retired, and this page is where their signals moved). Color comes from each store's own 4-week normal, never from the district ranking. The camera sees the lane state at Store 8; it does not see why a store is slow — that is the district manager's visit. No dollar estimate on the gap.

Build readiness. Serving and waiting states are a shipped primitive; per-store adaptive baselines are a shipped primitive elsewhere in the product. Store comparison, response time, and the daypart cut are derivable. Not built: the district rollup and this page.