Design the API and data model considerations for quick commerce replenishment under scale, incentive, and regulatory constraints

Problem Statement Description

Zepto wants to improve replenishment for busy households who repeatedly buy essentials such as milk, bread, eggs, fruits, snacks, baby care, cleaning supplies, and personal care items. The challenge is to design the API and data model considerations that enable reliable, scalable, and compliant replenishment experiences while supporting basket completion across hyperlocal dark stores.

This is a Technical PM problem. You should focus on how customer demand signals, inventory availability, substitutions, delivery promise, pricing or incentives, and regulatory constraints flow through systems. The experience may include reminders, recurring carts, auto-add suggestions, subscriptions, or replenishment recommendations, but the core task is not to design the UI—it is to define the technical product requirements, APIs, data entities, and system trade-offs needed to make replenishment work at quick-commerce scale.

Assume Zepto operates across many cities and dark stores, with high SKU churn, freshness constraints, stockouts, picker errors, variable delivery capacity, and strong customer expectations around speed and trust. The system must handle real-time inventory, user preferences, consent, promotions, and compliance requirements without creating operational overload or misleading customers.

The experience should consider:

- Core users and workflows: busy households, pickers, dark-store operators, category teams, growth teams, and customer support.

- API requirements for replenishment intent, cart creation, inventory checks, substitutions, delivery slot or promise validation, pricing, incentives, and order confirmation.

- Data model considerations for users, households, SKUs, inventory, dark stores, replenishment frequency, preferences, consent, substitutions, freshness, and promotion eligibility.

- Reliability and scale needs, including latency, concurrency, idempotency, stale inventory handling, fallback states, and failure recovery.

- Privacy, security, and regulatory constraints around purchase history, profiling, consent, sensitive categories, auditability, and data retention.

- Incentive and pricing trade-offs, including promotion abuse, margin impact, fairness, and separation of recommendation logic from commercial rules.

- Observability needs such as event logging, API health, inventory accuracy, recommendation quality, basket completion, fulfillment failures, and customer complaint signals.

- Rollout considerations, including experimentation, phased launch by city or category, operational readiness, guardrails, and rollback triggers.

Your goal is to frame a technically sound product architecture that lets Zepto support replenishment in a trustworthy, scalable, and regulation-aware way while improving basket completion and minimizing customer disappointment from unavailable, incorrect, or poorly timed items.