Airline revenue
management has long been seen as an art, but as the industry transitions to
offer and order management, it is science that is taking the driver’s seat.
First developed as a
discipline by the likes of American Airlines in the 1970s, revenue management
has changed substantially over the years. Initially carriers introduced
discount fares based on time of booking but this soon grew into fares priced
based on time of day, day of travel as well as seasonality.
To start
with, fares were adjusted at the flight leg level and optimized in isolation.
Many network carriers have since transitioned to origin and destination
(O&D) revenue management: if there is one seat available from Atlanta to
New York, O&D revenue management would prioritize the passenger who wanted
to connect in New York onto the airline’s flight across the Atlantic ahead of
the passenger wanting to end their flight in New York as it increases the
overall revenue.
Now airlines are
using smart platforms to take a hybrid approach and it is this that requires
some scientific thinking.
PROS, which works
with carriers like Emirates, Lufthansa and Hawaiian Airlines, is investing
heavily into its science. John McBride, Lead Strategic Consultant and formerly
the company’s vice president of product management for its airline
products, has overseen much of the
company’s operations in travel revenue optimization, e-commerce and retail and
has a PhD in mathematical statistics, for example. This is the field in which
mathematical probabilities can be updated based on new evidence – perfect for a
world of setting airfares that can change hundreds of times a second based on
customer demand, the weather, the actions of competitors and hundreds of other
factors.
PROS uses Bayesian
statistics and a healthy sprinkling of artificial intelligence and other
innovations in its revenue management forecasting models.
One of the newest
innovations is its Willingness-to-Pay (WTP) science, a crucial concept behind PROS’ platform for
airlines. In a traditional airline pricing world built on Reservation Booking
Designators (RBDs) – better known as fare buckets – an airline might have filed
fare buckets priced at $100 and $150. If the airline offered the $100 fare but
the passenger was happy to pay $125, the airline loses the opportunity to get
an additional $25. If the $100 fare is not available and the airline offers the
$150 fare, the passenger walks away and the opportunity to generate $125 is
lost.
“An airline has a
massive amount of fixed cost, whether the plane flies empty or full, you still
have a crew, you still have people that have to load baggage, you still have
pilots,” McBride said.
“So it's always
tempting to just sell the next seat for nothing because you're just trying to
get something in return, but the reality is that passengers are willing to pay
different amounts.”
The company’s
analysis shows that adding WTP information to the forecast can increase average
revenues by between 1 and 3% by improving price targeting for customer segments
with higher WTP, reduced discounting and more competitive pricing.
Yet WTP is a very
hard problem to solve because of the high number of variables – hence the need
for revenue management companies like PROS to hire both smart people and invest
in technology, particularly AI.
Dynamic pricing
The move away from
traditional revenue management with fare buckets is focused on the concept of
dynamic pricing where fares are priced on a continuous curve between those
filed fares based on the demand and the willingness-to-pay of the customer.
The origins of
dynamic pricing were laid out in a seminal paper published 30 years ago by
Garrett van Ryzin, where he showed mathematically that there is an optimal way
to control demand using it.
It is not a complete
revolution. Revenue was managed in the fare-bucket era by adjusting the number
of seats available in each bucket rather than the price associated with those
buckets.
The interest in
dynamic pricing has two main drivers: the boom in low-cost carriers where fare
classes have become largely irrelevant and the need for network carriers to
offer more price points, even for loyal customers.
“It became about
product, schedule and price. Passengers are willing to forgo certain things
based on schedule and/or price,” McBride said.
The Lufthansa Group
of airlines started its dynamic pricing journey with PROS in 2014 and went live
with continuous pricing across multiple distribution channels in 2020, just as
the pandemic shut down the airline world.
Early results were favorable
and the group has since doubled down.
In 2024,
PROS launched a new product with Lufthansa Group called Request-Specific
Pricing, using deep neural networks and machine learning in conjunction with
what the company calls dynamic Bayesian generalized linear models, which has
now been implemented by Lufthansa.
“We've
seen some incredible results from this and even published a paper on what we
have learned,” McBride said.
Speaking at
the 2024 Outperform customer conference, Lufthansa Group’s head of commercial offer
methods and automation, Florian Martin, said, “The first version of a
predecessor of this system, which was in-house, Lufthansa-developed, gave us a
2.3% physically significant revenue uplift. With the new PROS solution, we
upped this to 5.2%. This is net of continuous pricing that's even coming on top.
We have high confidence that this is really worthwhile to do.”
Although it cannot be
laid entirely at the door of continuous pricing, Lufthansa Group has posted
record revenues since the pandemic.
While Lufthansa Group
has been an early mover in adopting dynamic pricing, some other airlines have
adopted a wait-and-see approach.
When Hawaiian Airlines
migrated from another vendor to PROS in 2021-22 there was a lengthy evaluation
process of the benefits.
“They had a 7%
increase in revenue for leisure markets and a 9% increase in yield using PROS
and their analysts didn't have to do so much. Now they're starting to ask the
question, ‘What is the next evolution?’” McBride said.
Despite these
impressive figures, adoption of dynamic pricing has been slower than expected.
“It's not because
airlines don't have the appetite for it,” he said. “It's just that being able
to consume continuous pricing and integrating it back into other systems is
very slow or costly, which is the largest impediment right now.”
Ancillary pricing strategies
McBride predicts that
now that airlines are moving away from their pandemic focus on operations and
with all the discussion around AI, more airlines will follow the lead of
airlines like Lufthansa and Hawaiian.
Dynamic pricing of
ancillaries also promises to allow airlines to earn higher revenues by matching
a passenger’s willingness to pay for them.
For the
past 20 years, airlines have been unbundling services, such as checked bags,
Wi-Fi and food and beverage offerings from the base fare, driven by the new
model of low-cost carriers. In 2023, the total
value of airline ancillary revenue was $117.9 billion, 275% higher than ten
years earlier.
IATA and
McKinsey say that the transition to offer and order management will generate $4.10 per passenger boarded as airlines add new ancillaries and implement
dynamic offerings.
Tackling
dynamic pricing of ancillaries required new thinking from PROS.
“Ancillaries like bag
or seat selection are not really driven by capacity and demand, as is the case
with flight prices. Passengers might not be willing to pay more for such
products closer to departure but might instead decide they need ahead of
departure. This means that the WTP for ancillaries has to be considered
separately from the flight price,” McBride said.
AirBaltic is one of
the airlines that has rolled out dynamic pricing of ancillaries through PROS.
"The
implementation and testing of new techniques took approximately six months from
the first kick-off meeting until the moment we were able to measure the results
and have data-driven conclusions, " said Iuliia Granja Velasco, AirBaltic’s
e-commerce project manager.
The
result? A nearly 6% increase in revenue in some markets, exceeding expectations
and leading to a wider rollout.
McBride believes revenue management is not dead, it is merely changing.
“The tenets of
revenue management have not changed, he said. “You're still solving a problem where you have a
limited amount of time to make the most money off of a finite set of products
and services. That hasn't changed but some systems have become obsolete.”
Airlines must now
adapt to this changing market, not least because their competitors have started
to do so.
“Not every airline will be left behind, but
certainly many will,” McBride said.
Those that get left
behind will be those that see revenue management as an art and remain skeptical
of the science.