Use Case Financial
Destination Recommender System enhancing Airline Revenue
In the airline industry, personalized customer engagement is crucial for boosting bookings and enhancing satisfaction. The effectiveness of these recommendations hinges on the precision of integrating and analyzing extensive user data.
The system uses advanced machine learning to parse travel history and preferences from Google Analytics and CRM, improving recommendation accuracy. Targeted web banners based on these recommendations aim to increase booking likelihood at key decision-making moments.
Challenges
The project faces challenges in Data Integration, due to difficulties in merging Google Analytics with CRM data caused by inconsistent email usage and device sharing, and in Model Accuracy, as variations in user behavior and data gaps impact recommendation precision.
Data integration
Merging Google Analytics with CRM data is complex due to varied email usage and device sharing, complicating user identification.
Model accuracy
Maintaining high recommendation accuracy is challenging, influenced by variations in user behavior and lack of data.
Solution
A dual-model ensemble has been implemented to enhance destination recommendations for airline customers. This system combines a Neural Collaborative Filtering (NCF) model with a content-based approach, each adding distinct strengths to the recommendation process.
The NCF model predicts user preferences based on past interactions using deep learning techniques, while the content-based model utilizes destination embeddings generated from descriptive content. This hybrid approach is specifically designed to address data sparsity and improve recommendation accuracy by integrating diverse data types.
Tech stack
Results
The project has generated millions in annual profit through personalized destination recommendations, significantly boosting bookings. High precision in user targeting has improved recommendation accuracy, leading to higher conversion rates.
40%
Precision at k=5
MS€
Millions Annual Profit
