“We can give airlines some cushion to defend against tougher times and increase their ability to cope with changing conditions— something machine learning is perfectly suited for,” says FLYR CTO and co-founder Alexander Mans.
FLYR, a travel tech company focused on commercial aviation revenue management and dynamic pricing analytics, has raised more than $10 million in a second round funding to promote R&D, hiring and product updates as it takes on more global airline clients as an enterprise-facing business, Kambr Media has learned from several sources outside the company.
The funding, which brings the company’s total raised to over $25 million, comes amid elevated investment activity and intensifying competition in the race to enhance airlines’ RM systems. Just last week, aviation booking startup Volantio followed up its February $2.6 million Series B with an additional investment from Amadeus Ventures.
FLYR previously raised $15.3 million between 2016 and 2017 in two seed rounds and a Series A. The intervening years have seen a significant change in direction of the travel tech space – and in airlines’ recognition that their RM systems need to catch up to consumers’ demand for greater personalization and automation in transacting with carriers.
The way airlines market their seats and ancillaries has also added more complexity in the pricing of services, notes FLYR CTO and co-founder Alexander Mans in an interview with Kambr Media.
In a sense, the evolution of FLYR’s business model mirrors the progression of airline-oriented travel tech. When it first raised its series A from such notable backers as PayPal founder and early Facebook investor Peter Thiel, JetBlue Technology Ventures, Streamlined Ventures, and Western Technology Investment (WTI) along with an early, small investment from RM services provider Amadeus (which did not participate in FLYR’s Series A or B rounds,), the idea of finding ways to empower consumers real-time booking intelligence seemed to be a highly lucrative business model.
At the time of its initial funding, FLYR was a peer of consumer-facing apps like Hopper and travel booking engines like Kayak. But as Mans says, that area of travel tech is extremely costly to implement and scale. Plus, the risks tend to outweigh the potential rewards. For example, that may be a reason that some observers may have been surprised when Lufthansa Group made a “multi-million dollar” investment in Hopper earlier this year to pursue collaborative research on machine learning. Clearly, Hopper recognized that serving consumers solely was limiting, if not a dead end.
As FLYR shifted its approach over the last few years, it found more demand for its dynamic forecasting capabilities by airline partners. Sensing greater opportunities, a new direction was set by the company’s founders.
“We got greater insight into how airlines actually operated their revenue management teams on the inside,” Mans says. “The core practice of revenue management is not necessarily to handle pricing; it is to manage inventory against a set of bucketed price points and booking classes. So, we thought, ‘It doesn't make any sense. Why would you manage pricing as an inventory optimization problem?’ So we said, ‘Okay, let's look into how we can change it for the better.’ Over time, we got more and more airlines interested in working with us on developing and deploying a new paradigm in pricing.”
FLYR, which has 85 employees spread across Silicon Valley and Krakow, Poland, currently has five carrier clients, including “major LCCs, and full-service carriers in the U.S., Southeast Asia, the Middle East, and Australia.” The company expects to serve 10 clients by 2020.
The move to an enterprise model and expectations of an expanded client roster culminated in the release of FLYR’s FusionRM in 2018.
FusionRM encompasses five components: A central datastore of all the airline's data, machine learning based willingness-to-pay and demand forecasting, network pricing optimization, and performance reporting and analytics.
“We deploy the exact same product with every airline; whether you're a low cost or full service carrier, it doesn't matter,” Mans says. Of course, we do optimize our machine learning for the airline to the network the customer base data, etc.”
“One, those solutions are still very much on premises,” he says. “Two, legacy vendors run probably dozens of different versions between all their customers, which makes maintenance very costly – and makes it very difficult for airlines to transition to new systems.
“Instead, what we do is we run on any of the three major clouds; Google, Amazon, or Microsoft,” Mans continued. “We run every airline deployment against the same version of the system. We roll out upgrades and deploy new features to everybody at the same time – if they want them, of course. It increases maintainability and reduces costs, allowing us to make the solution more competitive and reduce the time to market new capabilities for any of our customers.”
From his conversations with carriers, Mans says that airlines are now at a point where they are “fed up” with the inability of existing vendors to either move fast or to give them access to the latest technologies.
“You're seeing some of these vendors offer their own cloud solutions, but, really, it's no different than taking the very same thing that's already deployed on the airline’s servers and putting it in a cloud environment,” Mans says. “I would almost describe our approach as working toward a future of ‘pricing as a service.’ That's a really big difference from what you see from other vendors.”
The FusionRM platform takes in airlines’ search data, booking data, scheduling and inventory information, catalogs, competitor pricing, fare filings – “everything you can imagine,” Mans says, adding that the system is powered by close to 30 specific data sources.
Once that information has been gathered, FLYR organizes it into a standard data format, which is common across its airline customers, “from the data ingestion and transformation pipelines.”
“We give the airline the opportunity to remove the silos in their organization that segregate their data, concentrating it in one place,” Mans says.
The second step in the FusionRM services process is forecasting. Instead of primarily attempting to predict demand, FusionRM’s larger aim is to determine how many bookings an airline will get and then factor in any arbitrary price point an airline might want to sell at.
“For every day up until departure, we basically establish how many bookings we expect until departure, at up to 200 different price points. These price points can be determined by the airline’s fare filings or any other set of arbitrary price points.”
FusionRM’s optimization engine then takes in all FLYR’s forecasts and delivers what the system deems to be the best price points across the airline’s network.
“The beauty is simply coming up with the optimal price,” Mans says. “It is very easy to convert the optimal price to the equivalent booking class, or seat availability. But it's nearly impossible to do it the other way around. It allows us to deploy our technology and show the ROI, which goes anywhere from 2-to-6 percent revenue growth, depending on the carrier. From there, we build a relationship with the airline and phase out their existing systems over time.
“We did all this together with our first customers,” Mans adds. “I cannot speak about all our customers but at this point, we work with everyone from low cost carriers that run on systems such as Navitaire and airRM, all the way up to full service carriers in the Middle East and all through Asia that have home-grown solutions or make use of systems such as PROS.”
Typically, in order to complete an airline booking, you need two, perhaps three components, Mans says. You need a filed fare for a booking class. Then you need the availability of a booking class. After that, an airline needs to calculate taxes and various surcharges.
Those components are all filtered through the GDS, where an airline can then can discount fare overrides to specific travel agents or via selected booking channels.
An airline can apply whatever discount it wants on their website, since there they control that platform and channel, but offering the right discount to the right traveler at the right time is another matter. Many countries prevent airlines from directly targeting fare discounts to specific customers, so it is difficult to deliver anonymized discounts without violating local laws.
FLYR will be pressed to prove it can strike the right balance.
In one scenario according to Mans, a customer books a flight through a specific airline’s website or app. FLYR data analytics then deploys the options to create a discount in real-time designed to inspire a sale at that moment within the airline’s targeted range. The real-time adjusted models FLYR presents consider several factors: customer segment, whether they’re part of a loyalty/rewards program, and whatever else may be in that customer’s current shopping cart.
“One thing you're starting to see is that airlines have begun to treat their experience more as a shopping cart experience as opposed to a traditional branded fare,” Mans says. “Between seeing the shopping cart in real-time and knowing who you are, we have the ability to make an override, or adjustment, on top of the base fare that would have been shown if all else were equal. That’s how we return a personalized offer.”
The level of personalization FLYR can apply varies from airline to airline.
“If I had to classify it, all the way on the left of the spectrum you have the low-cost carriers that are already reliant on up-selling and cross-sales,” Mans says.
“On the very right end you have the full service carriers: the Lufthansa's, the Emirates', the British Airways, United's,” Mans continues. “They value brand consistency, loyalty, et cetera, so they will never allow us to sell the same products to two people sitting across from the same table at the same time at a different price.”
There are ways FLYR can manage that as well, Mans says.
For instance, a traveler from San Francisco or New York is seeking a seat in basic economy. FLYR’s system might present a $400 fare offer from its airline client. If another prospective traveler does the same search 20 minutes later, FLYR can simply return a cached version of the price that was shown to the previous airline shopper.
“There are still ways you can personalize, but it's kind of a first come, first serve prioritization basis to provide consistency in the marketplace,” Mans says.
For Mans, his interest in solving the problems associated with air travel reflects an early fascination with flying and being an entrepreneur. Born and reared in The Netherlands, Mans started his first business at age 11 and went on create a network security company at age 15. By 2013, he moved to California and founded FLYR with CEO Jean Tripier and CMO Cyril Guiraud to start the travel data science company now known as FLYR. Alex is a licensed pilot and avid aviation enthusiast.
“One of the first words I could speak was the Dutch equivalent of ‘airplane,’” he says I grew up around airplanes, stay up-to-date on all news in the industry, and like bringing new technology to legacy environments (airlines being a prime example).
When asked about the last years’ investor interest and funding activity in the aviation travel tech space, Mans considered the trajectory of FLYR and the changing demands of airlines and venture capitalists.
“Here's the interesting thing: you have consumer travel on one side, such as Hopper. And then you have enterprise SaaS, enterprise IT on the other, with FLYR, Airnguru, and others,” Mans says. “Consumer travel tech is heavily influenced by the public’s sentiments and general growth of the industry. But it is low margin, high risk, and very capital intensive. You won't find many startups like Hopper or Kiwi.
“Most consumer-facing travel startups that get seed or first round funding never survive because there are too many companies in that space,” he adds. “We used to operate in that space, but we decided not to go after that challenge and to focus instead on the enterprise side.”
The enterprise-focused side has enjoyed wide support from investors over the past two years, Mans notes. They’re working on everything from revenue management to airline operations to cargo services and customer support.
“One of the big drivers of that enterprise growth is that airlines have gone through an investment cycle over the last two years,” Mans says. “They’ve become more aware of cloud computing, machine learning and all the things they don't have today but that everybody else does. They look at Amazon and Google and they're like, "Why don't we have this?’
“That awareness creates some interest in driving investment,” he adds. “Secondly, airlines have had a couple of really good years, on average. There've been big returns; they're building up some cash, and they're building more shareholder value. Raising capital on the market has become easier over this period, so they now have money to spend.
“Thirdly, I think airlines are getting more exposed to startups and they're realizing that it might be a way for them to get things done faster,” Mans says. “They can’t get new technology from their existing vendors — or can't it get fast enough. There’s also the realization that startups actually get things done faster. This all makes it easier to get airlines as customers, but also improves investor sentiment in that space because of the growing and developing customer base.”
Still, Mans warns that the perception of investing is always 6 to 12 months behind what appears to be the current trend. He believes that emerging markets that have driven the rise in global travel, such as those in Southeast Asia, the Middle East, Africa, Latin America, are becoming more defensive as they're preparing for less profitable times.
“I believe that cycle on the enterprise SaaS space in travel is going to wind down a bit in the next 12 to 18 months, because airlines are becoming more defensive,” Mans says. “We are lucky enough now to be past the initial hump. That allows us to get the traction to keep going based on performance, and on the revenue growth we can drive for our airline customers.”
Naturally, it all depends on how one views the condition of the ever-uncertain global economy that will determine where, and how much, airlines and investors spend. For FLYR’s clients, the company’s message is one of an urgent need to develop better retailing strategies to protect sales amid fears of global economic declines.
“We can give airlines some cushion to defend against tougher times and increase their ability to cope with changing conditions, something machine learning is perfectly suited for,” Mans says. “They're getting more defensive and protecting themselves for the future. You can see this already in airlines pulling out of markets, etc. —airlines going bankrupt like Primera Air, Air Berlin. There's even a large number of 'near failures’, especially in the low-cost European and Southeast Asian markets. We’re prepared for that defensive position airlines increasingly want to take.”