Build a metric tree for Uber Eats after a major redesign at global scale

Problem Statement Description

Product context: Uber is a mobility and delivery platform; its products include rides, Uber Eats, grocery and retail delivery, freight, driver and courier tools, and marketplace pricing.

Uber Eats has completed a major global redesign of the courier-facing experience, affecting how couriers discover delivery opportunities, accept trips, navigate pickup/dropoff, manage earnings visibility, and understand marketplace signals. You are asked to build a metric tree that can help Uber evaluate whether the redesign improves the courier experience while preserving marketplace health and business outcomes across many countries, cities, vehicle types, and operating models.

This is a metrics interview question. Focus on defining a clear measurement structure rather than proposing product changes. Your metric tree should connect high-level business and marketplace outcomes to courier behaviors, operational quality, and product experience indicators, while accounting for the fact that Uber Eats operates a three-sided marketplace involving consumers, merchants, and couriers.

Because the redesign is global, your framework should be useful across heterogeneous markets with different courier supply dynamics, regulatory constraints, payment models, app versions, delivery modes, and competitive intensity. It should also distinguish launch monitoring from long-term success measurement and include ways to detect unintended harm.

The experience should consider:

- A clear top-level success metric or set of outcome metrics, with precise definitions and denominators.

- Breakdown of the metric tree into courier engagement, fulfillment reliability, earnings-related outcomes, marketplace liquidity, and product usability.

- Instrumentation needed across the courier app journey, including impressions, offers, accepts, cancellations, pickup/dropoff events, navigation events, and support contacts.

- Cohorts and cuts such as geography, courier tenure, vehicle type, time of day, delivery density, app version, and marketplace maturity.

- Guardrail metrics for consumer experience, merchant operations, courier safety, fraud/abuse, unit economics, and marketplace balance.

- How to separate redesign impact from seasonality, promotions, local operations changes, pricing changes, and supply-demand shocks.

- Decision usefulness: which metrics would inform rollout expansion, rollback, iteration, or deeper diagnosis.

The goal is to produce a rigorous, actionable metric tree that Uber leadership and cross-functional teams could use to judge whether the redesigned Uber Eats courier experience is working at global scale, where it is helping or hurting, and what evidence is needed before making rollout or iteration decisions.