Diagnose Sunday Miami same‑day outages

Instacart provides both same-day and scheduled delivery, and it turns off same-day when there aren't enough shoppers available. The data indicate that same-day was unavailable on two of the past three Sunday afternoons in Miami. a) Specify exact supply–demand balance metrics: the live shopper-to-order ratio, SLA fill rate, distribution of queueing delays, shopper acceptance rate, and effective capacity after accounting for shopping time and travel constraints. Which one leading metric would you choose to trigger the disablement, and why? b) Build a root-cause tree listing the precise data you would collect (minute-level traffic, batch creation rate, online/active shoppers, substitution complexity, store closures, pay rates, weather, events), and explain how you would causally distinguish demand spikes from supply drops (for example, using exogenous weather as an instrument or difference-in-differences against other markets). c) Suggest three interventions (supply incentives, shifting scheduled orders to same-day, and demand shaping through ETAs or fees), and for each one design an experiment specifying eligibility, randomization unit, guardrails (cancellations, NPS), and success metrics; estimate the expected impact and call out the main risks (cannibalization, fairness, marketplace instability).

Overview: This question tests a data scientist's skills in marketplace analytics, causal inference, metric engineering, and experiment design by requiring precise supply–demand balance metrics, a root-cause analysis, and randomized interventions for repeated same-day disablements in one regional market.