Building a Recommendation Engine for Largest Timeshare Vacation Exchange Network in the World
THE BACKGROUND
One of the largest timeshare vacation exchange networks in the world has revamped and reintroduced its recommendation engine to help customers find the best units. The newly deployed solution followed a blended approach to recommend resorts based both on a content-based, collaborative filtering and a popularity-based algorithm. However, the recommendation engine failed to meet the desired performance. The recommendations were not as diverse as expected and were biased toward popular resorts. In addition, the extremely sparse user-resort matrix made it hard for the algorithm collaborative filtering to discover similarities across users with confidence. The company engaged with Soothsayer to assess and improve current implementation of its recommendation engine