How would you measure success for a Zepto-style quick commerce replenishment launch?

Problem Statement Description

Product context: Zepto is an Indian quick-commerce company; its products include rapid grocery delivery, fresh produce, everyday essentials, local inventory, and dark-store operations.

Zepto is considering or has recently launched a replenishment experience for quick commerce customers, focused on busy households that repeatedly buy essentials such as milk, eggs, bread, fruits, vegetables, snacks, diapers, pet food, and cleaning supplies. The core promise is to help customers complete recurring baskets quickly and reliably, while fitting into Zepto’s hyperlocal operating model of dark stores, fast picking, freshness management, and last-mile delivery.

Your task is to define how success should be measured for this launch. The interviewer is looking for a metrics approach that connects customer value, business impact, and operational feasibility without jumping directly to features or solutions. You should clarify what “replenishment” means in this context, who the primary users are, what behavior change is expected, and how the launch should be evaluated over time.

The measurement framework should account for the fact that replenishment can affect many parts of the quick-commerce system: basket building, repeat purchase frequency, stock availability, substitution behavior, delivery reliability, picker workload, customer trust, and margin. It should also distinguish between short-term launch health and long-term habit formation.

The experience should consider:

- A clear success definition for replenishment, including the customer action, eligible product categories, and denominator for measurement.

- Primary and secondary metrics that reflect basket completion, repeat usage, customer retention, revenue quality, and operational performance.

- Instrumentation needs across app surfaces, recommendations, cart additions, substitutions, checkout, fulfillment, delivery, and post-order feedback.

- Relevant cohorts such as new versus repeat customers, high-frequency households, category-level replenishment users, city or dark-store clusters, and customers exposed versus not exposed to the launch.

- Guardrail metrics for stockouts, cancellations, late deliveries, poor substitutions, refund rates, picker errors, customer complaints, and cost-to-serve.

- Data quality and privacy considerations, especially when using purchase history, household patterns, reminders, or personalized replenishment questions.

- Decision usefulness: how the metrics would help determine whether to iterate, expand to more categories, scale to more locations, or pause the launch.

The goal is to demonstrate how you would build a practical, decision-oriented measurement system for a Zepto-style replenishment launch—one that captures whether customers are forming a reliable habit, whether baskets are being completed more effectively, and whether the operating model can support the experience at scale.