Two snack shops sit at a school gate and operate simultaneously. You have 14 days to recommend which will deliver higher profit over the next quarter. Design a measurement and analysis plan that yields a defensible decision under real-world constraints. Specify: 1) The exact primary metric (e.g., profit per passerby per open hour) and why it is decision-aligned; include what costs and revenues are in-scope and how you will normalize for traffic and hours. 2) The minimum data you will collect (hourly foot traffic counts, transaction counts and AOV, item-level margins, labor hours, rent/utilities allocation, weather, school calendar, competitor promotions/price changes). 3) Your identification strategy—experiment (e.g., randomized flyer distribution or alternating queueing) or quasi-experiment (e.g., hour-level difference-in-differences with fixed effects and weather controls)—that remains valid under: weekday/weekend and exam-week spikes, Store B extending hours, Store A weekend discounts, and a price change by one shop on day 9. 4) The model and uncertainty: write the DID regression you would fit (define outcome, treatment, fixed effects, controls), how you’d compute a 95% CI for the profit delta, and a minimal power check showing 14 days is sufficient (order-of-magnitude inputs are fine). 5) Guardrails to detect stockouts/cannibalization and your explicit decision rule (e.g., recommend Shop X if estimated Δprofit > $Y per day and CI lower bound > $Z).