Design a Recommendation Engine for Airbnb Experiences

To design an effective recommendation engine for Airbnb Experiences, I would prioritize timing and relevance over volume. First, I'd define success by measuring the conversion rate from exposure to booking while monitoring cancellation rates to ensure we aren't causing decision fatigue. My strategy focuses on the post-booking phase, specifically the 'Itinerary' view in the mobile app. This is when a user has committed to the stay but hasn't yet planned their local activities. I would segment users based on their listing type; for example, guests booking entire homes in historic cities should see curated walking tours, while those in family resorts get cooking classes. The placement would be a subtle, non-intrusive section titled 'Local Favorites' appearing only after the reservation is confirmed, utilizing collaborative filtering to match past browsing behavior with the new location. We must also implement a feedback loop to learn from ignored suggestions. By delaying the pitch until the user is excited about their trip but still planning, we increase relevance and reduce annoyance, aligning with Airbnb's mission to help people belong anywhere.