Vetting a Vendor's Competitor Rideshare Data: Legality, Accuracy and Worth
Company: Uber
Role: Product Manager
Category: Analytics & Experimentation
Difficulty: medium
Interview Round: Other
You are a product manager at Uber. The company plans to purchase a third-party dataset containing ride data on its competitors in the United States. Explain how you would evaluate this dataset, including what you would check before trusting it and how you would reach a recommendation on whether it is worth buying.
```hint Find something to check it against
A vendor's claims are cheap to make. Before trusting any number in the dataset, think about which facts you already know for certain and could use to measure how close the vendor gets.
```
### Clarifying Questions
- Is the purchase already decided, so the job is to vet and use the data, or should the evaluation decide whether to buy at all?
- Which competitors, cities and time period does the data cover, and is it individual trips or aggregated totals?
- How did the vendor collect the data, and what does it cost, as a one-time snapshot or an ongoing feed?
- Which business decisions is the team hoping the data will inform?
### What a Strong Answer Covers
- Defining the business questions and decisions the data must serve before judging the data itself.
- Provenance, legal and privacy clearance, including how the competitor data was gathered.
- Concrete quality checks: coverage, granularity, freshness, completeness, bias, and accuracy validated against known ground truth.
- Weighing value against cost and alternatives, ending in a clear buy, negotiate or walk-away recommendation.
- Risks after purchase, such as panel drift, over-trusting estimates, and how the data is governed internally.
### Follow-up Questions
- Where you can compare, the vendor's figures for Uber's own rides come out well below the true numbers, and the gap differs from city to city. What do you conclude, and is the dataset still usable?
- The vendor refuses to explain how the data was gathered, saying the method is proprietary. Do you go ahead?
- Six months after the purchase, how would you judge whether it was worth the money?
Overview: An analytical product question for an Uber product manager about evaluating a third-party dataset of competitor ride activity in the United States before buying it. It tests framing business questions, legal and privacy vetting, data-quality and accuracy checks, and weighing value against cost.