Pothole Detection Data Sources and Freshness Targets for a Navigation Product
Company: Google
Role: Product Manager
Category: System Design
Difficulty: medium
Interview Round: Other
You are a product manager at Google. To improve the driving experience, the team wants to know where potholes are on the road. Explain how you would source data about potholes, and what requirements, if any, you would set for how fresh that data must be.
```hint Let the use decide the freshness
Pothole data could be used in more than one way while people drive, and each use tolerates a different delay. Potholes also appear and get repaired over time, so think about keeping the data current in both directions.
```
### Clarifying Questions
- Which product will use the data: navigation in Google Maps, Waze, Android Auto, or an aggregated feed for road authorities?
- What should a driver experience: a warning as they approach a pothole, routes that avoid badly damaged roads, or both?
- May we collect passive sensor data from phones during navigation, with user consent, or only reports that drivers submit themselves?
- Which roads matter first: busy urban streets and highways, or also rural roads where few devices pass each day?
### What a Strong Answer Covers
- The driver problem the data solves and the specific use case, which in turn sets coverage, accuracy and freshness needs.
- Several candidate sources compared on coverage, accuracy, cost, latency and privacy, rather than a single source.
- How noisy raw signals become confirmed pothole locations with a confidence level, including deduplication, filtering look-alike events and removing repaired potholes.
- Freshness requirements stated per use case, with the trade-off between publishing quickly and publishing accurately.
- Success metrics and guardrails such as alert precision, detection delay, driver distraction and privacy.
### Follow-up Questions
- On a rural road only a handful of phones pass each day. How would you keep coverage and freshness acceptable there?
- Drivers complain that warnings appear for potholes that were filled weeks ago. What would you change?
- How would you test whether pothole warnings actually make driving better or safer?
Overview: A Google product manager technical question on how to gather data about potholes to improve the driving experience and how fresh that data must be. It tests comparing data sources, turning noisy signals into confirmed hazards, privacy trade-offs and setting freshness targets for each use case.