Seasonal travel patterns have become more unpredictable, making it difficult for airlines to forecast revenue. To navigate the future of revenue optimization, the airline industry needs to establish a new baseline. FLYR, a revenue optimization company focused on the airline industry, recently introduced The Revenue Operating System, a cloud-based software solution that leverages deep learning technology to contextually analyze data for more optimal pricing and revenue strategies and more accurate forecasting.
SkiftX spoke with Cole Wrightson, Chief Product Officer at FLYR, about how airlines and other travel verticals can move beyond traditional thinking around revenue management, embrace a new set of technology building blocks, and plot their revenue optimization strategies for an uncertain future.
SkiftX: How have changes in consumer buying behavior and the need for flexibility altered travel and transportation companies’ outlooks on revenue management strategies?
Cole Wrightson: The slow pace of innovation has been frustrating airlines for well over a decade. Covid and the associated changes in consumer demand patterns and products offered by airlines have turned industry frustration with legacy revenue management (RM) systems from a steady drip into an open faucet of discontent. The need for adaptive, more modern systems has become so great that some airlines have completely turned off their legacy systems and opted to manage everything manually, just to survive. The inconsistency between year-over-year demand patterns and the increased frequency of schedule and capacity changes invalidate many of the outdated assumptions upon which legacy RM systems are built.
SkiftX: Can you explain a bit about The Revenue Operating System? How does it differ from the traditional understanding of revenue management?
Wrightson: Responding to the need for greater flexibility and automation, FLYR built The Revenue Operating System with the ability to offer new distribution capability (NDC), total revenue optimization, and dynamic, continuous pricing. We’re experts in applied artificial intelligence (AI), so we made sure to build an AI-first revenue management system. This directly addresses industry needs like total revenue optimization and incorporates new data sources and market context not used in other systems. The Revenue Operating System natively enables continuous price outputs that support NDC offers and also supports backwards compatible integrations with any traditional reservation system or distribution channel.
SkiftX: How would you describe FLYR’s “AI-first” approach to a more effective revenue management system?
Wrightson: FLYR leverages the true benefits of AI by allowing our purpose-built deep learning algorithms to directly correlate data inputs, pricing decisions, and revenue performance. Others in the industry are attempting to wedge simple models into small portions of traditional RM solutions, but as you would expect with that approach, minor actions lead to minimal improvements.
It’s important to point out that while “AI-first” means The Revenue Operating System is much more automated than traditional RM solutions, it is not “AI-only.” There is always a need for the airline analyst, not the algorithm, to maintain control. FLYR has made significant investments in product development and user research to ensure that outputs are explainable and can easily be adjusted when the airline desires a different strategy, such as defending market share rather than revenue maximization.
SkiftX: What technologies are embedded in the revenue operation system of the future?
Wrightson: FLYR has built AI-first, cloud-native products from our beginning, so we have nearly a decade of experience using modern architectures such as serverless data processing and distributed model training using tensor processing units (TPUs) that are 30 times faster than graphics processing units (GPUs) and a thousand times faster than the central processing unit (CPU) in your laptop at home.
FLYR has also invested heavily in bringing cutting-edge AI research algorithms into production-worthy software for airlines. This has required us to build AI infrastructure and support for optimized deep learning algorithms that are not available off the shelf. We are constantly investing in our AI infrastructure and model architecture to improve our forecasts and optimal pricing models, as well as to tackle new revenue use cases for airlines.
SkiftX: How can these technologies provide a better understanding of context to make decisions in a way that wasn’t possible before — for example, across network planning or pricing?
Wrightson: The legacy revenue management problem is often considered in isolation from an airline’s other revenue decisions. Little to no information is shared between legacy RM systems and ancillary pricing, fare filing, marketing, or network planning systems. The Revenue Operating System is built to change that, first by utilizing a common data platform that spans all relevant commercial data from internal and external sources. Second, from the forecasts to the pricing decisions, the outputs are designed to be utilized by each application in the system. For example, the ancillary revenue forecasts are inputs to the RM system. The continuous price outputs from the RM system are inputs into the filed fare optimization app and the marketing app. The forecasts, in particular, are being leveraged by many different departments among our current customers.
SkiftX: How does The Revenue Operating System play a role in marketing to the customer at the right time in the right place?
Wrightson: Low prices drive demand, but strict compliance restrictions and legacy systems often make it difficult to promote the best prices and offers available. FLYR’s always-on platform continuously updates an airline’s digital advertising with its lowest compliant price, allowing marketers to advertise compelling prices year round for all markets, not just during heavily managed sale periods. On top of that, direct connectivity with the RM system allows for the advertised markets to be selected based on revenue impact to avoid marketing destinations that were already going to drive high loads and yields.
SkiftX: How does it help revenue reporting and analysis?
Wrightson: Improved analysis and reporting is a major benefit of The Revenue Operating System. Superior automation certainly improves user efficiency, but the real gains are from having enhanced access to data and metrics. FLYR’s system accesses hundreds of metrics to answer nearly any question about performance an analyst can think of, all while putting the results in the context of forecasts they can trust and decisions they can understand. When the analysis results in a need to change revenue strategies, minimal clicks are required within the same analytics screen to shift course.
SkiftX: How does FLYR plan on expanding its revenue operating system to other travel and transportation verticals?
Wrightson: Very thoughtfully. We’ve seen firsthand how our unique platform — a combination of data, AI, forecasting, and analytics — can scale outside of traditional revenue management for airlines by enabling new, connected applications on top of an airline’s legacy platform. We had to study and understand the use cases and legacy pain points to make sure our AI-first solutions hit the mark.
We know that there are a lot of similar pain points and use cases in adjacent travel and transportation industries. According to McKinsey Global Institute, there is estimated to be half a trillion dollars in available opportunities for AI-related technology improvements within the travel industry, 90 percent of which will come from marketing- and sales-related technologies. We’re thoughtfully entering into the right partnerships and opportunities, but we know we are barely scratching the surface when it comes to travel and transportation use cases that can actually benefit from our AI.
For more information about FLYR’s revenue optimization solution, visit FLYRLabs.com.
This content was created collaboratively by FLYR and Skift’s branded content studio, SkiftX.