How big data could help airlines bounce back after COVID-19
Researchers are using analytics to gain new levels of insight into how people choose a seat when they fly – and how much they're willing to pay for it.
Even before COVID-19, the airline industry was doing it tough. According to the International Air Transport Association, the industry average margin was just 1.407 per cent from 2004 to 2017.
Since being hit by the pandemic, the association expects carriers around the world to post a collective loss of more than $10 billion in 2022.
While downsizing fleets is one immediate solution, another may lie in using data to get more value out of under-utilised ancillary products such as Advanced Seat Reservation, or ASR.
ASR gives customers the ability to select specific seats prior to check-in, and would allow airlines to price each seat accordingly – if they could measure seat-specific demand.
To meet that challenge, new research by Melbourne Business School Chair of Management (Econometrics) Michael Smith, with University of Munich co-authors Professor Göran Kauermann and Dr Shuai Shao, has shown a novel way forward for computing seat-specific prices for ASR and other ancillary services.
"Even prior to the pandemic, airlines had notoriously low margins, and one important source of revenue growth was 'unbundling' their products," Professor Smith says.
"One key aspect of unbundling is the separate sale of ancillary services, such as ASR. However, airlines typically struggle to price these, and our model greatly enhances their ability to do so."
'Big data territory'
The trio's paper, Whether, when and which: Modelling advanced seat reservations by airline passengers, was published in international journal Transportation Research Part A: Policy and Practice.
While attempts to predict what factors drive customer willingness to select certain seats have been made in the past, never before have researchers attempted to use such extensive and detailed data to explore the reasons behind why people choose to book seats in advance, and the extent to which passengers are willing to pay for that feature.
Exploiting big data, Professor Smith argues, is important for airlines to understand where savings can be made, and where greater value can be extracted, without impacting the overall customer experience.
"While there is a lot of expertise in airlines in scheduling and capacity management, the use of large customer level databases to model the demand for ancillary products is in its infancy," he says.
"This is big data territory, and the models we use are detailed statistical ones."
Looking at a dataset of 485,279 bookings on five intercontinental routes, extracted from the complete booking database of a major European airline, Professor Smith and the team were able to find patterns such as middle-seat avoidance and front-seat preferences, as well as other factors contributing to which seats and when people choose to book ahead of time.
"No prior published study had used such rich data, nor such a detailed model, to capture the drivers of demand for ASR," Professor Smith says.
While the study was published in the first months of the COVID-19 pandemic, Professor Smith says the model can be used just as successfully with more recent data.
"A key feature is that the model is 'interpretable', in that there were three distinct model components to capture whether, when and which seats were selected by passengers," he says.
"While the model was fit using pre-pandemic data, this interpretability means that it can continue to be employed and calibrated using post-pandemic data."
Dynamic seat pricing
The study found that factors such as the day of the week, whether the booking is for multiple passengers, the distribution channel of the ticket sale, the relative price, the number of days prior to departure and the route flown all influence the likelihood of people booking certain seats, showing that seats are valued differently by customers at different times.
These findings could give airlines a leg-up in returning to the black after COVID-19 – for those that are prepared for price seats differently, and dynamically.
"By adopting the characteristics of the seat heat-map, airlines have many possibilities to differentiate these products," explains Professor Smith.
"For example, they could bundle bookings for multiple passengers with advanced seating reservations, or they could give different prices to aisle or window seats, as well as for different plane sections.
"Moreover, the time-based variables allow for dynamic pricing of ASR that reflects the relative value to customers of reserving a seat as the flight departure nears."
Prior to the study, the airline that supplied its data to the researchers varied its prices for seats based on which cabin section they were in and ticket type.
Since adopting the new model created by the researchers, it has increased revenues from ASR sales "significantly", Professor Smith says.
To read the full research paper, visit Whether, when and which: Modelling advanced seat reservations by airline passengers on ScienceDirect.
For more analytics information and research, visit our Centre for Business Analytics page.