Metrics question

How would you measure the satisfaction of hosts on Airbnb?

Practice this question out loud. An AI interviewer asks it, follows up like a real interviewer would, and scores your answer. Type or speak.

What this question tests

Metrics design for the underserved supply side of a marketplace: can you measure host satisfaction specifically, not guest satisfaction relabeled.

State why this matters: hosts are the supply side of Airbnb's marketplace, and host churn directly reduces available inventory and guest choice.
Identify host specific pain points: booking predictability, payout reliability and speed, and dealing with problem guests or property damage disputes.
Propose a primary metric: host satisfaction survey score collected after key moments, like after a payout or after resolving a dispute, not just an annual survey.
Add behavioral proxies: host retention rate (percent still listing after 12 months) and relisting or repeat hosting rate after a negative experience like a cancellation.
Add an operational guardrail: average time to resolve a host support ticket or dispute, since slow resolution is a common driver of host dissatisfaction.
Define overall success as improving trend in both survey sentiment and host retention together, since either alone can be misleading.

Focuses specifically on host pain points (payouts, disputes) rather than generic satisfaction language that could apply to guests too.
Combines survey sentiment with behavioral proxies like retention, since stated satisfaction and actual behavior can diverge.
Ties support resolution time directly to a known driver of host dissatisfaction.

Measuring guest satisfaction and treating it as a proxy for host satisfaction.
Relying only on an annual survey instead of moment based feedback tied to actual host experiences.

How would you weigh a host's survey score against their actual retention behavior if they disagree?
What would you do about hosts who are dissatisfied but have not yet delisted?