At Momondo we’ve been building a user level segmentation product that enables us to adjust our marketing spending and messages on a user level in different marketing channels (typically performance marketing) and also to provide personalized user experience and recommendations on the website. It’s a huge project that we are in the middle of, with significant impact on both product, marketing and monetization by processing search queries and behavior data measured on the Momondo platform around the Globe. The model is simple yet robust, processing millions of searches every day using machine learning algorithms as well and passing the user level insights back to our marketing channels and to our frontend.
- How to derive as much information as possible from a very limited set of input and build a flexible, yet solid architecture
- How to use user insights in targeting and realtime bidding and have a direct impact on user level ROI
- How to use user insights in personalization and recommendations even on users we have never seen before