Build a metric tree for Search after a major redesign

Problem Statement Description

Product context: Airbnb is a travel marketplace connecting guests and hosts; its products include stays, experiences, host tools, guest booking, trust and safety, payments, and reviews.

Airbnb has recently launched a major redesign of its Search experience in a global marketplace where guests discover stays and hosts depend on qualified demand. The redesign may affect how travelers enter queries, apply filters, evaluate listings, and move from search results to booking requests or reservations. It may also change demand distribution across hosts, markets, listing types, and trip intents.

You are asked to build a metric tree for Search that helps Airbnb understand whether the redesign is improving the product experience and marketplace outcomes. The metric tree should connect user behavior in Search to downstream booking value, host outcomes, and marketplace quality, while separating true product impact from seasonality, traffic mix, inventory changes, or ranking/model changes.

Your scope should focus on the Search journey after users arrive with travel intent, including result discovery, refinement, listing engagement, booking progression, and host-side consequences. The goal is not to design the redesign itself, but to define a clear measurement structure that product, data science, design, and marketplace teams could use to evaluate performance and diagnose issues.

The experience should consider:

- A clear top-level success metric and how its numerator, denominator, and unit of analysis are defined.

- The major branches of the metric tree across search usage, result quality, engagement, conversion, booking outcomes, and host impact.

- How to instrument key events across search entry, result impressions, filters, map/list interactions, listing clicks, wishlists, booking requests, and confirmed reservations.

- Relevant cohorts such as new vs. returning guests, domestic vs. international travel, city vs. leisure destinations, flexible vs. fixed-date searches, and professional vs. individual hosts.

- Guardrail metrics for trust, cancellation risk, guest dissatisfaction, host concentration, supply fairness, latency, and customer support contacts.

- How to distinguish redesign effects from marketplace supply changes, seasonality, pricing shifts, marketing traffic, or experimentation artifacts.

- How the metric tree would support diagnosis if bookings rise but host outcomes, guest satisfaction, or marketplace quality worsen.

Your goal is to present a practical, structured metric tree that would let Airbnb evaluate the Search redesign, monitor the health of the guest-to-host marketplace, and make confident decisions about iteration, rollout, or rollback.