Uber Eats · Metrics · Mid · 60 to 90 minutes

Order Drop Diagnosis for One Large City

Uber Eats is a marketplace app that lets people order food and other local goods for delivery or pickup from nearby merchants. Answer this product management case and get it graded line by line against a weighted rubric.

Metric diagnosis
Root cause analysis
User segmentation
Prioritisation
Experiment design

The situation

Your director for Core Consumer and the city General Manager have asked for a rapid diagnosis of a sharp decline in weekly Uber Eats orders in one large city. Assume the following numbers are illustrative. Orders in that city fell 18 percent in week 1 and 27 percent in week 2 versus the prior four week average, while nearby and peer cities stayed within plus or minus 3 percent. Revenue is down, but the finance team also sees higher courier incentives and refund costs in the city. Customer support reports more complaints about late arrivals and missing items, while merchant account managers say several large restaurant groups reduced menu availability at dinner. No major national pricing, policy, or marketing change happened in the same period. Leadership needs an answer this week because they must decide whether to spend more on promotions, courier supply, merchant recovery, or app fixes before the next monthly business review. The hard part is separating a local demand problem from a merchant or fulfillment problem, then proposing the smallest set of actions that can stabilize orders quickly.

What to produce

A 1 to 2 page diagnostic memo.

Your answer must

Build a clear metric tree for weekly orders in this city, including demand, merchant, and fulfillment drivers.
List at least 5 plausible hypotheses for the order decline, and rank them by likelihood and impact.
Specify the key cuts of data you would review first, including at least 4 dimensions such as daypart, neighborhood, new versus returning users, top merchants, app version, or payment method.
Recommend the top 1 to 3 actions to take in the next four weeks, with explicit trade offs and why lower priority options should wait.
Define success metrics and guardrail metrics for the recommended actions, with a target direction for each.

Constraints

You have 48 hours to prepare the recommendation, with support from one analyst.
Use existing dashboards, logs, merchant reports, and support data only, no new instrumentation can be shipped before the decision.
Any immediate intervention must fit within a local four week budget of 250000 dollars.
Recommendations must work within current courier pay, safety, and marketplace fairness policies.

Format

Start with the order metric tree and your highest priority hypotheses. Then show the analysis plan, the likely root causes, and the top actions with owners, timing, and metrics.