Root cause a sudden decline in retention among couriers using Uber Eats

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 observed a sudden decline in retention among couriers. In this RCA interview, you are asked to investigate what may be driving the drop, how you would structure the diagnosis, and what evidence you would seek before recommending any corrective action.

Focus on couriers as the primary user segment: people using the Uber Eats courier app to accept, complete, and get paid for deliveries. Their retention may be influenced by earnings, order availability, app reliability, incentives, safety, onboarding quality, customer or merchant interactions, competing delivery platforms, and local market conditions.

Assume this is a marketplace environment where courier supply, consumer demand, merchant readiness, pricing, incentives, and dispatch algorithms are interconnected. The issue may be global or localized, product-driven or operations-driven, and could reflect a true behavioral change or a measurement/instrumentation problem.

The experience should consider:

- How to define “retention” for couriers, including the time window, denominator, and whether the metric tracks app opens, online hours, accepted trips, or completed deliveries

- How to confirm the anomaly is real by checking data pipelines, event definitions, cohort logic, seasonality, and recent instrumentation changes

- How to segment the decline by geography, courier tenure, vehicle type, acquisition channel, earnings tier, delivery frequency, and marketplace density

- How to examine product and operational changes such as dispatch logic, incentive changes, pay transparency, batching, app crashes, navigation issues, onboarding flows, or deactivation policies

- How to evaluate marketplace factors such as lower order volume, longer wait times at restaurants, reduced courier earnings, higher cancellation rates, or increased competition from other platforms

- How to form and prioritize hypotheses using evidence, expected impact size, reversibility, and whether the pattern matches the timing of the retention drop

- How to distinguish short-term mitigation from long-term prevention, including monitoring, alerting, and operational ownership

The goal is to demonstrate a structured RCA approach that moves from metric validation to segmentation, hypothesis generation, evidence gathering, and mitigation planning, while accounting for the complexity of Uber Eats’ courier marketplace and the need to protect reliability, courier trust, and marketplace liquidity.