Share
Transavia is a leading Dutch airline, providing flights to over 100 destinations. The airline is a digital leader, with a customer-first ethos that’s reflected in its online booking platform.
We hosted Transavia’s Lead Data Scientist, Liveris Avgerinidis, at a Xebia event for data, technology and product leaders in London. Liveris presented an exclusive session explaining how his team has built a next-gen recommendation engine that’s powering better booking experience for the airline’s customers. You can watch the event recording here:
00:00 - Introduction
01:45 - Importance of data strategy
08:25 - Creating an actionable roadmap for AI solutions
12:15 - Introduction to Transavia airline
14:38 - Airline customer journey
15:23 - Applying personalization across the customer journey
21:37 - Proving the business impact of the AI flight recommender
27:24 - New possibilities
Our Managing Director Data, Giovanni Lanzani, introduced the recommendation engine topic, explaining how a fundamental data strategy is necessary to achieve success with any complex data initiative, including a recommendation engine build. Giovanni’s main takeaway was that a data strategy is so much more than a document – it should be an actionable rallying point for complete cultural change toward a data-driven culture.
Giovanni then expanded on how to build a data strategy, providing several useful guiding points.
Read more about how to build a real data strategy here.
After sharing these insights Giovanni handed over to Transavia’s Liveris Avgerinidis. Liveris’ data team serves Transavia across its entire operation, providing advanced data solutions for everything from operations to commerce.
The recommendation engine solution was ideated from a need to make the end-to-end customer journey more relevant for passengers. Personalisation throughout the booking journey increases the likelihood of successful bookings and happy passengers. Advancing this by recommending relevant flights would be another step forward in the personalised experience.
The need to build a recommendation engine that provides relevant and timely options to customers is challenging due to the complexities of data science and machine learning techniques. Transavia’s data team built a recommendation model with a three-step process that would provide a shortlisted group of flight options from a pool of millions. The model was trained on real-time flight booking characteristics, and it applied real-time smart filtering options to provide customers with recommendations based on their inputs.
Transavia ran many tests to measure the impact of its recommendation engine. They observed some critical insights that helped Liveris’ team make incremental adjustments to the engine that improved metrics over time. Ultimately the engine has had a measurable impact on bookings and customer satisfaction at Transavia, and Liveris’ team has an exciting future roadmap that will enhance Transavia passengers’ experiences even more.