Virgin Atlantic

Virgin Atlantic launched as the UK’s challenger airline in 1984 covering the London to New York route. Today they fly to 40+ routes spanning 4 continents and have a global network through their airline partners. Their brand ethos of innovation, passion and positivity lays behind their openness to new ideas to help empower their employees.


Challenge

Virgin Atlantic’s Revenue Management team is responsible for which price bands (or “ticket classes”) are available for any given flight on any given day. Each route has multiple flights with 3 cabin classes with 20+ bands of ticket prices that can purchased over 330 days ahead of take-off. In addition, there are several types of plane, with different configurations and capacities. The Revenue Management team’s objective is manage demand and supply patterns to optimise revenue per flight. If they allow tickets to be sold at low prices too early in the booking window, then the plane is full with a low average price per seat. If they set price bands too high, then the plane has a high average price per ticket, but lots of empty seats. Each analyst in the Revenue Management team covers thousands of flights.


Solution

Virgin Atlantic engaged Algospark to review opportunities for applied AI in revenue management. The team were keen to develop a bespoke tool that allowed them to identify and prioritise outlier flight demand patterns. This would allow them to act early on opportunities with the largest impact to maximise revenues. Using a series of workshops and weekly readouts, Algospark worked with the Revenue Management team to co-develop a tool to help them identify and prioritise flights for intervention. Focusing on the most popular routes first, and using several applied AI techniques, Algospark developed a tool that derived a benchmark sales path for each flight across all days prior to take-off. This allowed outliers identification, prioritisation and action recommendations for all flights. This meant Revenue Management analysts could focus on the most interesting and effective cases for intervention. The tool also logged actual versus recommended intervention, so that the system had a learning feedback mechanism. This allowed the tool to improve over the time and become even more effective.


Outcomes

  • An increase in average revenues per flight.
  • An increase in the number of routes and flights that analysts can successfully manage.
  • More engaging work for analysts as time shifted from finding outliers to actively managing outliers.
  • Supporting analytics became more effective as focus could be aligned to trends with outliers.

“Algospark were instrumental in beginning data science at Virgin Atlantic. They are top notch when it comes to bringing commercial and financial acumen to technical knowledge. Where they also excelled was their ability to work incredibly closely with stakeholders in the Revenue Management team, across all levels. Algospark’s work continues to live on since project completion, and is now being implemented globally. I can not more highly recommend Algospark and their contributions to Virgin Atlantic.”

Tim Lum, Chief Data Officer, Virgin Atlantic.