Airbnb Marketplace Health: Picking One North-Star Measure for Guests and Hosts
Company: Zillow
Role: Product Manager
Category: Analytics & Experimentation
Difficulty: medium
Interview Round: Other
If Airbnb could watch only one number to judge whether it is creating value, which metric should that be? Propose it and defend the choice.
```hint Check both sides of the trade
Ask what has to happen for a guest and a host to both come away better off, and test each candidate metric against a transaction where one of them did not.
```
### Clarifying Questions
- Is the metric for Airbnb as a whole, or only for the stays business rather than experiences?
- Should it weigh guests and hosts equally, or is one side of the marketplace the focus?
- Is it meant for company-level strategy over quarters and years, or as something product teams move week to week?
- Is the company prioritizing growth right now, or profitability and quality?
### What a Strong Answer Covers
- A clear statement of the value Airbnb creates for guests and for hosts, and how the chosen metric reflects both.
- A comparison of plausible candidates (bookings, nights, gross booking value, revenue, active users) with the reason each wins or loses.
- A precise definition: the counting rule, time window, and treatment of cancellations, refunds and long stays.
- Input metrics that teams can move, and guardrail metrics that stop the north star from being gamed.
- How the choice would be validated, for example by checking that it predicts retention and long-term revenue.
### Follow-up Questions
- Your metric rises while the average guest review score falls. What do you conclude, and what do you do?
- How would your answer change if Airbnb made stays of a month or longer a strategic priority?
- Which input metric would you give the search and ranking team, and how does it connect to your north star?
Overview: A product analytics question asking for a single north star metric for Airbnb's two-sided marketplace of guests and hosts. It tests how a candidate defines delivered value, compares candidate metrics, sets counting rules for cancellations and long stays, and pairs the choice with input and guardrail metrics.