Airbnb — Pricing — A/B Test & Causal Inference Questions

Role context: Data Scientist, Pricing Data Science · Est. study time: 55 min · 4 questions

How Airbnb measures pricing

Pricing is where Airbnb's experimentation is hardest, because a price change is a marketplace intervention:

  • Listings in a market compete, so a treated listing's lower price steals bookings from control listings — a SUTVA violation that biases a naive A/B test. Airbnb's own research studies this and the designs that fix it (cluster randomization by market, listing-vs-guest-side, shadow-price corrections).
  • Many pricing launches can't be cleanly randomized, so you fall back to observational causal inference (diff-in-diff across markets, staggered rollouts, matching).
  • Pricing effects are heterogeneous (elasticity varies by market, season, lead time) and read on the platform (ERF) with guardrails.

For the fundamentals — p-values, power, error types, distributions — see the Probability & Statistics section.

Each answer is a coaching walkthrough: a Sample answer (clarify → approach → a simulated back-and-forth → a clear call), then a Deep dive with illustrative example, then a Grading rubric.

Questions (4)

A price cut fills one calendar by emptying a neighbor's — is your A/B biased?Experiment designHardYou can't cleanly randomize a pricing launch — how do you measure it?Causal inferenceHardOccupancy is up but host earnings are down — do you ship?Decide from resultsMediumOne price elasticity for all markets — why is that wrong?Heterogeneous effectsHard