Revenue management has long been a pioneer in the area of data science.
Its history is rooted in mining large amounts of data, making sense of it, and providing a recommendation to the airline for availability that maximizes revenue.
Since the COVID-19 pandemic has taken hold, AI-powered revenue management systems have become an even more imperative tool for airlines looking to recover revenue from sharp declines in passenger numbers.
This fact is compounded when you consider the latest IATA research suggesting that global passenger traffic will not reach pre-COVID-19 levels until 2024. With greatly reduced passenger demand, global airlines are even keener on ensuring revenue is maximized while considering their overall corporate strategy during the pandemic.
COVID-19 has been tumultuous for revenue management. Passenger number predictions have been completely skewed by the pandemic. As a result, revenue management systems have had to rely on adaptive learning models more than before.
These models need to learn faster, based on the new trends that are coming as a result of the pandemic and have had to be complemented with additional data sources too.
The origins of revenue management – pre-COVID
When revenue management was first launched in the mid-eighties it was done so as an overbooking module.
Overbooking is the concept of overselling a flight, given there was likely not all of the passengers would show up for the flight.
Since the outbreak of COVID-19 the situation has become uncertain for revenue management because many predictions – based on what was seen, planned and forecasted before the pandemic – no longer apply.
Justin Jander - PROS
Thus, the first systems set out to predict how many passengers would show up relative to the total number of booked passengers. Using this information, along with the costs associated with a denied boarding and the revenue opportunity associated with selling the extra seat, optimization was performed to determine how many additional seats should be sold.
This concept was the foundation of revenue management and continues to be used today. The story of revenue continued to evolve from there.
After overbooking, the math showed the best approach to maximizing revenue was to forecast the number of passengers that were expecting to fly, combined with the value those passengers bring to the airline and then perform optimization to determine the optimal number of seats at each price point that maximizes revenue.
These fundamentals of revenue management remain the same, but the options for solving that problem continue to evolve.
The research around how to forecast this demand, as well as approaches to optimization, began expanding. The approaches to forecasting require advanced data science techniques that are robust enough to calculate small numbers and produce accurate results.
This is important because the forecaster does not just estimate the number of passengers that want to fly, but also the likelihood of a passenger choosing a particular itinerary, and the willingness of the passenger to pay a certain price. This opportunity opened the door for AI-based solutions to step in.
With AI-based solutions, the demand prediction is a learning process that not only estimates the demand but also updates the next forecast with information from what happened in the last. This iterative process results in more accurate predictions.
Since the outbreak of COVID-19 the situation has become uncertain for revenue management because many predictions – based on what was seen, planned and forecasted before the pandemic – no longer apply.
Combating COVID-19 through collaboration
Despite these challenges, revenue management has not become useless, cancelled, or unnecessary during this pandemic. It can be adapted to reflect a new set of circumstances and data sets. To be successful with this type of approach, it helps to collaborate with experts around the globe, as well as using data from many sources.
As an example of collaboration in the industry, PROS has set up a COVID-19 Taskforce with 17 customer airlines, to use external data sources – like government closures and infection rates, amongst others – to better understand what shape recovery will take.
These third-party data sources are paired with the bookings data from those partnering airlines, which allow the system to identify what signals indicate a return to bookings.
For example, if there is a reduction in infection rates, how far behind do bookings lag and for which departure dates. By pooling together the data from the airlines – keeping it anonymized – there is a larger data set by which conclusions can be drawn.
Through this, better market insights are garnered and recommendations for forecast adjustments are made. By translating data relating to the recovery and applying it to different regions, each airline can customize their recovery strategy from COVID-19 and improve revenue performance.
Forecasting and optimization
After the collaboration to determine the signals for recovery, airlines can then use data analysis and forecasting tools to predict bookings more accurately.
This is done by combining automated settings within the tools, as well as revenue management analysts providing valuable manual adjustments.
Given the nature of the pandemic’s circumstances, the airline can be more aggressive with the learning and information sharing process to react faster to changing market conditions.
Once forecasts are set, fares that are expected in the marketplace are paired. Just like the forecasts, the price points represented by the fares can also be adjusted, so capturing representative fares is equally important for airlines to accurately dictate availability that maximizes revenue.
Coming back stronger
Whilst the aviation industry continues to adjust to COVID-19 in these tough-to-predict times, there are signs that green shoots are emerging.
A recent survey from Inmarsat Aviation has revealed that whilst passenger numbers have reduced, for now, a new trend for digital transformation – already a catalyst for the modernization of aviation - will be critical to sustainable and profitable recovery for the aviation industry, when passengers do return en masse.
Through smart use of AI and data, airlines can protect their bottom line to recover from the financial turbulence COVID-19 has created.