For Vrbo's search algorithm, should we emphasize properties with the highest guest ratings or those that generate the most revenue for the platform?

Introduction

The trade-off we're examining today is whether Vrbo's search algorithm should prioritize properties with the highest guest ratings or those generating the most revenue for the platform. This decision sits at the intersection of user experience and business performance, two critical factors for Vrbo's success in the competitive vacation rental market.

I'll approach this analysis by first asking clarifying questions, then diving into the product understanding, identifying key metrics, designing an experiment, and finally providing a recommendation with next steps.

Step 1

Clarifying Questions (3 minutes)

• Context: I'm thinking about Vrbo's current market position. Could you share any recent shifts in market share or competitive pressures that might be influencing this decision?

Why it matters: Helps frame the urgency and strategic importance of the algorithm change. Expected answer: Increased competition from Airbnb or new entrants. Impact on approach: Would influence whether we prioritize short-term gains or long-term user satisfaction.

• Business Context: Based on Vrbo's business model, I assume there's a balance between commission-based revenue and subscription fees from property owners. Can you confirm the current revenue split and if there are any strategic shifts planned?

Why it matters: Affects how we weigh the importance of per-booking revenue vs. overall platform health. Expected answer: 60% commission, 40% subscription, with plans to grow the subscription base. Impact on approach: Would influence whether we lean towards immediate revenue or building a larger, more satisfied property owner base.

• User Impact: I'm curious about our user segments. Do we have data on the booking behaviors of first-time vs. repeat users, and how they interact with ratings?

Why it matters: Different user groups may have varying sensitivities to ratings vs. price. Expected answer: Repeat users are more rating-sensitive, first-time users more price-sensitive. Impact on approach: Would help us tailor the algorithm to optimize for user retention and acquisition separately.

• Technical: Regarding our current algorithm, what level of personalization are we currently employing in search results?

Why it matters: Determines the complexity of implementing and testing changes. Expected answer: Basic personalization based on search history and preferences. Impact on approach: Would influence whether we consider a more nuanced, personalized approach to this trade-off.

• Timeline: Is there a specific event or deadline driving the timing of this decision?

Why it matters: Affects the scope and depth of testing we can perform. Expected answer: Aiming for implementation before the summer travel season. Impact on approach: Would determine how extensive our experimentation can be and how quickly we need to move to a decision.