Build a metric tree for Uber Eats after a major redesign

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 just shipped a major redesign affecting the courier experience, such as how couriers view offers, navigate pickups and drop-offs, manage earnings visibility, and understand their status during delivery. As a PM, you need to build a metric tree that helps Uber understand whether the redesign is improving the courier experience while preserving marketplace health for consumers, merchants, and Uber’s business.

Your task is to define a structured measurement framework for the post-redesign experience. The metric tree should connect a clear top-level outcome to supporting input metrics, diagnostic metrics, and guardrails so teams can identify whether changes in performance are caused by courier behavior, marketplace matching, operational reliability, or measurement artifacts.

The answer should be specific to Uber Eats as a three-sided marketplace, with particular attention to couriers as the target segment. Consider how courier-side changes may affect delivery supply, acceptance behavior, fulfillment reliability, earnings perception, safety, and downstream consumer or merchant outcomes.

The experience should consider:

- A clear definition of the primary success metric, including numerator, denominator, time window, and why it reflects redesign success.

- The major branches of the metric tree, such as courier engagement, delivery execution, marketplace reliability, courier earnings, quality, and business impact.

- Instrumentation needed across the courier app funnel, from going online to accepting offers, pickup, drop-off, and post-trip actions.

- Relevant cohorts and cuts, including new vs. experienced couriers, geography, mode of transport, time of day, order type, and app version.

- Guardrail metrics for consumer experience, merchant operations, courier safety, cancellations, lateness, defects, and unit economics.

- How to distinguish true product impact from seasonality, market supply-demand shifts, experiment exposure, or logging changes.

- How the metric tree would be used by product, operations, design, data science, and engineering teams to make decisions after launch.

The goal is to create a practical measurement framework that can guide post-redesign evaluation, surface trade-offs early, and help Uber decide whether to iterate, expand, pause, or roll back parts of the redesigned Uber Eats courier experience.