More talks in the program:
How does a search engine approach recommendations and rankings? From “generalization” – recommending popular, best-seller-type listings – to “personalization” – deciding when there is enough user info to customize what a particular individual sees. Unlike many e-commerce sites, which try to keep people online and clicking, reading, watching and consuming, a booking site for vacation rentals tries to get people back offline as swiftly as possible. As a global metasearch engine for vacation rentals, we aim to make the trip-planning phase easy, so people can start getting excited about upcoming travel and not be bogged down by logistics. Once a “sale” is completed (in our case, after someone books a vacation rental), he or she is unlikely to book another at the same time, or even in the near future. The user journey is complete; there is no need to keep suggesting that because they booked one type of rental, they may consider looking at a similar bunch.As user tastes vary to a large degree, we need a very efficient, precise approach to be able to react to user behavior in real-time. As such, we’ve built our own recommendation systems.