Operations stream: Lounge visitor forecasts
The client, one of the largest and best-rated airlines in the world which operates lounges in airports globally, wanted to increase their business bottom-line by limiting resource wastage.
The operations team aimed to reduce resource wastage in lounges by ensuring that optimal number of staffs were deployed, while optimal food was prepared depending on the footfall.
MathCo developed a comprehensive forecasting framework to receive daily and hourly forecasts on the footfall using an exhaustive dataset of customer, lounge, membership/loyalty, delays, airport, weather, flight/destination related information.
- Daily lounge visitor forecasts displayed a 92%-95% accuracy and hourly forecasts displayed an ~85% accuracy.
- Accurate peak time forecasts helped reduce resource wastage and potentially saved an estimated $5 million annually.
Commercial stream: Smart trip recommendations
The client, one of the largest and best-rated airlines in the world, wanted to accurately predict the most likely location of customer trip bookings on its website.
Millions of trips with different layovers and routes were analyzed to develop an aggregated dataset of customer attributes, trip details, external influencers & more.
MathCo developed a deep learning-based model for the airline to effectively identify the next best destination for each customer and thereby improve conversions.
- Recommendations provided for over 6 million customers on their next best destination/trip with an accuracy of over 85%.
- An automated solution improved effectiveness of promotions by sending customized mail notifications to customers, with greater conversion ratio and potential annual savings estimated at $0.5 million.