Fermi Problem: Sizing the Pothole Count on Mainland American Roads
Company: Google
Role: Product Manager
Category: Product / Decision Making
Difficulty: medium
Interview Round: Other
Estimate how many potholes there are in the continental United States.
```hint Anchor on what you can size
Start from a quantity you can estimate with some confidence, then work out how the pothole count relates to it and where that relationship changes a lot.
```
### Clarifying Questions
- Does "continental" mean the 48 contiguous states, or should Alaska be included?
- Do you want the number of potholes that exist at one moment, or how many form over a year?
- Should I count only public paved roads, or also unpaved roads, parking lots and private driveways?
- How large does a break in the surface need to be to count as a pothole?
### What a Strong Answer Covers
- A defined scope covering geography, the surfaces counted, what qualifies as a pothole and a point-in-time count versus a yearly flow
- A clear equation with each input labeled as an assumption and sized with round numbers
- Segmentation where pothole density differs by a large factor
- Correct arithmetic to a single figure with a stated range
- A sanity check from an independent angle, and the assumptions that move the answer most
### Follow-up Questions
- How would your number differ between late winter and late summer, and why?
- If a maps product wanted to measure potholes instead of estimating them, what signals could it use and how would you validate them?
- A city has a fixed repair budget. How would you decide which potholes to fix first?
Overview: A Fermi estimation exercise asking how many potholes exist in the continental United States. It tests defining what counts and over what time window, anchoring on a quantity that can be sized with confidence, segmenting where density differs, keeping the arithmetic clean and sanity-checking the result.