Volatility tied to the war in Iran has thrown a wrench into the airline industry’s pricing mechanisms.
Jet fuel costs are fluctuating as the Strait of Hormuz stalemate causes oil prices to swing. And demand is shifting, too, as travelers reconsider trip plans due to the conflict.
Artificial intelligence (AI) pricing and fare prediction models are gaining in popularity, but it remains to be seen how well the technology weathers instability.
“Due to the uncertainty of the ceasefire and ongoing conflicts, this is really a day-to-day situation for revenue managers, and [I’m] not sure if artificial intelligence (AI) can predict the end of wars,” said Lorraine Sileo, senior analyst and founder of Phocuswright Research.
Airfare and volatility
For airlines, rising fuel costs have led to higher ticket prices and lower capacity, yielding increased passenger revenue in the short term, according to Sileo.
Fuel is one of the highest operating costs for airlines, according to Haley Berg, lead economist at Hopper Technology Solutions (HTS). Berg said fuel can account for 20% to 30% of total expenses, making it a significant airfare driver.
“When oil prices rise, particularly when increases are sudden or unexpected, airlines face materially higher per-flight costs that must be absorbed by the airline or reflected in pricing to maintain sustainable operations,” Berg said.
This isn’t the first time that oil prices have spiked, but the impact on jet fuel is “outsized,” Amex GBT said in a white paper published last week on the impact for corporate travelers. Even with the ceasefire, Amex GBT said experts predict it will be months before jet fuel supplies recover.
While the sticker shock of jet fuel makes waves among executives, the technical reality for airline pricing is more nuanced. The challenge goes beyond absorbing higher costs—it’s balancing those costs against shifting traveler sentiment.
As the war in Iran continues, not everyone sees fuel cost as the main driver behind changing air prices.
Phil Donathy, VP of flights marketplace for Skyscanner, told PhocusWire that what’s happening with pricing right now is mostly demand-driven. Demand for routes that might feel more risky is softening, and alternative flight routes are absorbing redirected traffic, Donathy said.
“Fuel costs do matter, but the relationship isn't as direct as people assume,” Donathy said.
How AI pricing models hold up
Sileo predicts AI will continue to play a larger role in airline decision making when it comes to pricing, route and capacity.
And while AI may not be able to predict the outcome of a war, it may be able to work fast enough to account for one.
When asked how the volatility has impacted AI-driven pricing models, Sam Chamberlain, chief product officer at FLYR, said, “It’s quite a mess.”
But it’s not beyond FLYR’s AI pricing technology's capabilities, he said, noting they can fine-tune pricing based on variables.
The challenge arises less with modern models than it does with traditional systems, according to Chamberlain.
“The problem is that most airlines haven’t implemented these capabilities yet and continue to rely on revenue management systems that are typically oriented toward optimizing revenue without fully considering costs,” Chamberlain said, noting fuel costs aren’t typically part of inputs.
As airlines start to prioritize offer optimization, fuel cost and other data signals could play a role in constructing offers, Chamberlain said. Fuel cost also plays a “critical role” in margin optimization, he said.
Uri Yerushalmi, co-founder and chief AI officer for Fetcherr, a PhocusWire Hot 25 Travel Startup for 2023, shared a look at how the company’s models are reacting to current dynamics.
Fetcherr is seeing that fuel price volatility increases the importance of fuel costs as an input for pricing, but its technology isn’t designed to react to hourly or daily fluctuations.
“Because our pricing recommendations are generated through a predictive, simulation-based framework, there is an inherent smoothing effect,” Yerushalmi said. “The system looks beyond short-term noise and focuses on market conditions more likely to matter for future demand and pricing.”
But right now, Yerushalmi said advanced AI-based pricing is an advantage.
“In these environments, the ability to interpret fast-changing signals objectively, without emotional bias, becomes critical,” Yerushalmi said. “That creates a real opportunity for airlines that can read the market correctly and respond with discipline.”
Chamberlain agreed and said “it's worth noting that airlines using dynamic, continuous pricing and vendors offering flexible price tuning are better positioned to perform under these conditions, as they can react to real-time variability much faster."
Live data is key
Historical data typically plays a role in price prediction—but during times of uncertainty, live data is key, according to industry stakeholders.
“As in any period of external volatility, short-term price prediction does become more challenging,” Yerushalmi said.
Chamberlain said confidence and accuracy in demand forecasting and therefore price prediction often declines during periods of short-term volatility. That’s partially because those kinds of outlooks rely on historical data, and events such as the current conflict in the Middle East are outliers.
And the importance of live data extends to the price-surfacing side, too.
During disruptions, booking windows compress. People move faster and patterns shift. That's when having live data at huge scale really earns its keep.
Phil Donathy, Skyscanner
Donathy said what’s happening now serves as a reminder of how dynamic flight pricing is. Fares are already set in response to “dozens” of simultaneous signals such as inventory, demand and competitor pricing.
Keeping a real-time read on what’s going on will matter right now. Donathy said that Skyscanner searches 100 billion prices daily—that gives the metasearch provider an up-to-date look at demand and price shifts that they then share with partners.
“During disruptions, booking windows compress,” Donathy said. “People move faster and patterns shift. That's when having live data at huge scale really earns its keep.”
Long-term impact
Outlier periods can impact pricing models’ outputs in the future.
Chamberlain said that if booking behavior rapidly changes it creates distortions that will have an influence on pricing models.
And that can play out in more ways than one: It’s not just the ticket price that might shift.
“In this case, it may not be just the core fare that is affected; airlines will likely attempt to recover fuel-related losses through other levers, such as increasing ancillary pricing closer to departure or onboard, reducing flight schedules and similar measures,” Chamberlain said.
Broader shifts like that can also affect the accuracy of price predictions, Chamberlain said.
Donathy sees disruption as a positive for pricing models’ turnout going forward.
“Disruption adds to the picture rather than breaking it,” Donathy said. “The pricing behavior you see during volatile periods all feeds back in and makes future models more resilient.”