In many ways, the commercial aviation industry has been indelibly changed. This has been especially felt within the practice of forecasting flight demand.
Historical data used to be the go-to source of data, but during these unprecedented times, it’s essentially been rendered useless.
So, what information and tactics can airlines turn to instead? We consulted with several industry sources to find out.
Uncertain Forecasts Lead to a Need for New Approaches
The revenue manager’s job difficulty has been cranked up to level 10 because the data and tools that were once the backbone of operations are no longer producing accurate results.
“Using historical data is normally the key to any Revenue Management System (RMS) but given the current situation, it is now mostly irrelevant. Airline teams must find a way to add value without this data and with the uncertainty of future customer demand,” said Benjamin Cany, Director Offer Optimization, Airlines, Amadeus.
“Carriers need a forecasting tool that can use only a few months of live sales data to give revenue management analysts an accurate idea on passenger demand evolution.”
3Victors CEO Rick Seaney paints the picture another way, “we need to find a closer with a 100mph fastball, as real-time is the key that translates to velocity.”
Any way you put it, speed and agility are the name of the game for airlines constantly adjusting and readjusting to ever-changing market conditions.
“We need to find a closer with a 100mph fastball, as real-time is the key that translates to velocity.”
“The COVID-19 pandemic represents a unique challenge for RMS. Numerous months of cancelled flights have left holes in the historic database, customer’s behaviors have significantly changed compared to pre-pandemic, and finally, there is a significant uncertainty about future customer demand during and after the recovery phase,” said Cany.
With so much change happening so quickly and so often, many revenue managers find themselves only having time to try to assess what the new market looks like rather than responding to the market and finding permanent solutions. It’s akin to the old cartoon where the character is left plugging leaking holes until he’s out of fingers.
"It has been more important to have the right set of thermometers to rapidly detect changes and act than trying to get the perfect medicine,” said Juan Afeltra, founder, TARMAC. "Not because there might not be one, but because the context changes so fast that you find yourself constantly researching for new alternatives to overcome each new challenge.”
Airlines Flock to New Data Signals
How can airlines put away panic and patchwork tactics in favor of more sustainable approaches? The answer lies within alternative data sources, which we’ll simply define for now as data different than the traditionally used historical data. Let’s break down a few of those data sources.
Current News & Events
Airlines have been no different than anyone else, as we all seek out the latest news to understand the current climate.
PROS cites a 43% average increase in bookings of a specific O&D once a country or region lifts or reduces its restrictions.
"Infection rates and government closures are two key data points affecting airline bookings that must be considered when building forecasts. When news information is released on border openings we see an influx in traffic to those locations," said Aditi Mehta, Solutions Strategy Director, PROS.
As a matter of fact, PROS cites a 43% average increase in bookings of a specific O&D (origin and destination) once a country or region lifts or reduces its restrictions.
Correlated Booking Data
PROS has also cited the need for airlines to leverage correlated booking data on flights yet to depart (as opposed to historic data) to derive accurate forecasts.
"Cluster analysis and classification schemas are not the sexiest parts of revenue management, but they are vital parts of optimizing RM performance," said Steve Hendrickson, Director of Advisory Services, Kambr.
"And the vagaries of this current pandemic environment will only exacerbate the value of doing them well and timely."
Read Hendrickson’s article on how to best conduct cluster analysis
Not only must airlines try and forecast their own demand and pricing, but they still need to have an understanding on the rest of the market.
“You must be aware of the competitive environment and be able to anticipate or react to competitor actions by forecasting their pricing policies or by modeling the alternative schedules available on the market as well as dependencies across them,” said Cany.
One of the most discussed types of data since the inception of COVID-19 has been customer intent data, data that signals customers’ interest in specific flights and/or destinations.
“Pre-pandemic I was intrigued with the promise of several alternative data sets with "demand signal recency" potential (as our stated goal is to accurately predict the "look to book" ratio in real-time), said Seaney.
"Customer data is an extremely important demand signal."
Justin Jander, Director of Product Management, PROS, echoes a similar sentiment.
"Customer data is an extremely important demand signal. Things such as search data (where are customers looking to fly to?) and point of sale information (O&D, date of sale, date of departure, etc.)."
However, Seaney warns to proceed with caution because of data privacy concerns and technical hurdles.
“One might have thought the pandemic recovery was complicated enough in this endeavor, but not so fast, as big tech is on a privacy spree whereby cookies and GPS data are likely to become very scarce making big data wisdom of the crowd that much more important.”
One only must look to Apple’s newest iOS update, which includes a new App Tracking Transparency feature to understand Seaney’s point.
Activating Data into Offers
It’s one thing to compile various data points, but it’s another to understand and activate that data.
As Jander puts it, "Given the varying number of signals, airlines must be able to answer, how strong should I react to the changing environment?"
One of the best ways to understand signals is by ensuring your data is accurate and understandable.
“We need a platform to ingest, cleanse, scrub, change-detect and categorizes copious amounts of un-cached real-time priced itinerary data from as many channels as possible and be willing to augment current modeling/forecasting and distribution slowly but surely,” said Seaney.
“Customer segmentation should be done in real time based on anonymized criteria (e.g.: number of passengers, travel characteristics), which takes into account a traveler’s willingness to pay to create a win-win value proposition.”
Once the data is verified and understood, it still must be activated in the form of sending relevant offers to prospective travelers.
Price has always been a major component when it comes to what flights travelers book. One of the best ways to construct an optimal offer is through customer segmentation.
“Customer segmentation should be done in real time based on anonymized criteria (e.g.: number of passengers, travel characteristics), which takes into account a traveler’s willingness to pay to create a win-win value proposition,” said Cany.
“We define accurate pricing as the product of a two-layer analysis that contrasts a behavioral network optimization with a tactical detection process that simultaneously seeks out opportunities and risks that usually last for less than a day.”
“This is based on ‘Customer Choice Modeling’: understanding how people are making buying decisions when they are in front of the airline website, how they value the different product elements, how do they value a stopover connection compared to a direct flight for instance.”
Afeltra has his own approach when it comes achieving accurate pricing.
“We define accurate pricing as the product of a two-layer analysis that contrasts a behavioral network optimization with a tactical detection process that simultaneously seeks out opportunities and risks that usually last for less than a day,” said Afeltra.
“Behavioral optimization allows airlines to rapidly load and change RASM (revenue per available seat mile) flight objectives, triggered by load factor and yield conditions at different levels of aggregation and translate them into a dynamic availability strategy.”
Read more about Afeltra’s views on forecasting
However, it’s not just about the price when it comes to selecting what airline to fly, especially now when reassurance and security can go a long way towards making consumers feel comfortable enough to return to the skies.
"In addition to adjusting price to match market conditions, airlines need to focus on the customer experience, for instance, by being transparent in what is offered and offering flexible refund policies," said Mehta.
Although it’s been an unprecedented difficult stretch for commercial aviation, a possible silver lining is the fact that all the upheaval has been a catalyst for a lot of needed change in processes and data applications which will have long-lasting positive effects far beyond the time of COVID-19.