Basket completion rate dropped suddenly in quick commerce replenishment. Diagnose the root cause under scale, incentive, and regulatory constraints

Problem Statement Description

You are the PM responsible for Zepto’s quick-commerce replenishment experience for busy households—users who build baskets of repeat essentials such as milk, bread, eggs, fruits, snacks, cleaning supplies, and personal care items. A sudden drop has been observed in basket completion rate, meaning fewer users who begin assembling or replenishing a basket are successfully reaching a completed order.

The issue is occurring in a high-scale, hyperlocal operating environment where product experience, dark-store availability, picker accuracy, pricing, incentives, delivery promises, and regulatory constraints can all affect whether a household completes checkout. The RCA should account for both digital funnel behavior and offline operational realities, including stockouts, substitutions, minimum order value, delivery fees, coupons, assortment visibility, and serviceability.

Your task is to frame how you would diagnose the root cause, prioritize the investigation, separate signal from noise, and identify what evidence would confirm or reject each hypothesis. You should also consider how privacy, data quality, accessibility, cost, and operational load constrain the investigation and any immediate mitigation.

The experience should consider:

- Clear definition of basket completion rate, including numerator, denominator, time window, and whether it applies to replenishment flows, repeat baskets, subscriptions, or all carts.

- Segmentation by geography, dark store, user cohort, platform, app version, acquisition channel, basket size, category mix, payment method, delivery slot, and incentive exposure.

- Funnel breakdown from replenishment entry point to item add, stock validation, substitution, cart review, checkout, payment, and order confirmation.

- Instrumentation checks to rule out tracking changes, event delays, duplicate events, app releases, experiment contamination, or data pipeline issues.

- Operational hypotheses such as stockouts, stale inventory, picker substitution failures, delivery capacity constraints, assortment gaps, freshness concerns, or dark-store-level outages.

- Commercial and incentive hypotheses such as coupon changes, minimum order thresholds, surge fees, subscription benefits, pricing mismatches, or margin-control rules affecting checkout.

- Regulatory and trust-related factors such as restricted items, compliance-driven assortment removal, privacy limitations on personalization, payment failures, or misleading availability claims.

- Mitigation and prevention planning, including short-term containment, owner alignment across product/ops/growth/engineering, customer communication, monitoring, and recurrence prevention.

The goal is to demonstrate a structured RCA approach that can operate under ambiguity and scale: define the anomaly precisely, isolate where and for whom the drop is happening, validate hypotheses with reliable evidence, protect customer trust, and recommend a disciplined path to mitigation without jumping prematurely to a single explanation.