Report Overview
Skift Research has recently published a report on the impact of Artificial Intelligence (AI) on revenue management within the airline industry. Despite the proven potential benefits of AI, airline executives have been hesitant to adopt the technology due to the complexity of airline systems and the potential for small errors to cause significant system-wide breakdowns.
However, the post-pandemic landscape has presented airlines with a range of challenges, including supply chain disruptions, capacity constraints, and unpredictable shifts in demand. In response, airlines are reevaluating their operations and exploring opportunities to modernize their revenue management and distribution departments. By leveraging technology, airlines can potentially improve efficiencies, generate incremental revenue, and reduce operational costs.
Although the integration of modern technologies into airline systems requires careful consideration, the benefits of doing so in the current climate cannot be ignored. As such, executives are taking a measured approach to technology integration, while recognizing the potential for significant improvements in revenue management and operational efficiency.
What You'll Learn From This Report
- The shortcomings of legacy revenue management systems
- The impact of technology-based revenue management systems on the existing RM processes
- The link between modern distribution strategy and advanced pricing models
- The market opportunity that AI drives within the revenue management function
- How AI improves airline ancillary revenues
Executive Summary
Since the introduction of ChatGPT in November 2022, there has been considerable debate regarding how Artificial Intelligence (AI) can improve business operations. The focus of this discussion has been on the potential AI-driven improvements to numerous processes and operations within businesses.
For airline executives, the dawn of modern technology presents a challenge. The slim profits of the airline industry and the high investment required for technology implementation has meant that many carriers continue to rely on outdated infrastructures for their core operations, such as revenue management and distribution.
However, the potential for growth in revenues, passenger and ancillary, presented by technology has left little room for skepticism among airline executives. Consequently, many legacy airlines, including Lufthansa, United, and Delta, have been at the forefront of investing heavily in their internal processes, paving the way for others to join them.
Large technology aggregators, such as Sabre and Amadeus, as well as regional technology companies, such as Airnguru, have AI-powered product suites that are assisting airlines in embedding technology into their revenue management systems. Given the value that AI brings to airlines, carriers are eager to form partnerships with such companies.
Despite the outdated revenue management infrastructure at most airlines, the potential impact is significant, as a technology-enabled platform can help improve demand forecasting, enable advanced pricing models, better understand customer willingness to pay, and avoid disruption. Furthermore, AI will drive an increase in ancillary revenues for airlines, which opens up a multi-billion-dollar opportunity for carriers to pursue.
Introduction
Artificial Intelligence (AI) is a rapidly growing field that is transforming the world we live in. To keep up with the competition in the future, individuals and organisations alike must become familiar with AI technology. Like steam power, mechanised engines, and coal supply chains revolutionised the world in the 18th century. AI technology is changing how we work, our economies, and society.
According to McKinsey, around 60 countries have developed a national AI strategy and expect AI capabilities to rival those of humans by 2030. This could add over $4 trillion annually to the global economy.
The airline industry has been using AI to personalise offers and create a better customer experience. By learning about passengers' behaviour, airlines can create offers that make the most sense. But airline applications for AI go far beyond customer experience. Airlines can also leverage AI in operations and revenue management.
Within the airline revenue management function, the goal is to accurately predict who will show up to buy a ticket and what the passengers are willing to pay if they do show up. With AI, airlines can more accurately forecast this while automating processes around data crunching and adjustments.
Airline revenue management really couldn't be done without science or AI. Airline seats are perishable goods, and there is fixed capacity in the short term, volatile demand, and very seasonal demand. Science allows airlines to manage those factors to stay in business and offer their products to many passengers.
The airline industry was the first to introduce pricing and revenue management, but many established airlines have struggled to keep up with more recent innovations. Some critics argue that airline revenue management has not progressed much since the late 1990s. However, the truth is that airlines face challenges such as legacy systems, data silos, and other obstacles that make it increasingly difficult for them to stay ahead of the curve.
Airlines have been slow in adopting the latest revenue management technologies, which has caused the industry to face mounting pressure to upgrade its revenue management systems.
We believe that AI-related technologies will shape the future of revenue management in the airline industry. These technologies allow airlines to capture and analyse big data in real-time and automate pricing decisions.
But before we estimate the value at stake for airlines that AI can for airlines, let's understand the problems with the current state of revenue management, starting with the legacy revenue management systems that somehow stayed relevant in the age of the AI revolution.
The Problem with Legacy Revenue Management
The airline industry has a history of innovation, particularly in the 80s when they were often the first to introduce new pricing strategies. One of the pioneers in this regard was American Airlines, under the leadership of Robert Crandall, who introduced several groundbreaking innovations, including the first-ever frequent flyer program, route optimization, and a central reservation system.
However, the airline industry is now facing what is known as the Innovator's Paradox. On the one hand, airlines played a role in creating the global distribution systems that exist today, but now they find themselves struggling to break free from their control. On the other hand, airlines developed the framework for revenue management, but have been slow to adopt modern pricing models, instead relying on outdated methods. This has caused a variety of problems, including dependency on outdated technology.
While modern technologies are available and have been proven to work, airlines continue to use legacy systems, leading to issues such as incompatible architectures, inefficiencies, and security risks.
In addition to legacy infrastructure, there are other reasons why airlines have been slow to adopt advanced pricing technology. These include incompatibility with legacy systems required to use advanced pricing models, organizational designs within airlines that do not promote cross-team collaboration, and the current technology's inability to accurately understand a customer's willingness to pay.
Legacy Infrastructure
Legacy infrastructures have recently caused several problems for airlines. These include inefficiencies, inaccuracies, and multiple breakdowns, which have resulted in thousands of flight cancellations and significant disruption to their day-to-day operations.
These outdated platforms present a variety of challenges, from data migration issues and difficulty integrating with modern systems to security risks.
Data Migration and Integration can be tricky with old systems.
Legacy systems often contain years if not decades, of accumulated knowledge and data critical to airport operations. This makes the prospect of migration daunting. Additionally, older systems may be highly customised to suit specific needs, offering functionalities that might not be readily available or easily replicated in newer solutions. There's also the aspect of integration. Many legacy systems are deeply integrated with other tools and processes. Finding new solutions that can seamlessly interact with existing workflows without causing disruptions can be challenging. The compatibility of new software with existing systems and data formats is a significant concern.
Also, legacy systems cannot integrate seamlessly with newer technologies. This limitation can lead to disjointed operations and a fragmented customer experience.
During a conversation with Cole Wrightson, who is the Chief Product Officer at FLYR, the issue of outdated data architecture in airlines was discussed. Wrightson emphasized that older data systems are a major obstacle in the adoption of newer technologies. According to Javier Jimenez, the COO of Airnguru, airlines cannot implement newer pricing models due to their incompatibility with legacy GDS systems.
Javier also identified specific issues that are hindering airlines from progressing, including a lack of access to data and a shortage of technology and AI expertise.
Another significant drawback of the legacy systems is their inefficiency and high maintenance cost. A study by American Council for Technology (ACT) and Industry Advisory Council (IAC) found that legacy systems can use up to 75% of a company's IT budget just for maintenance, leaving only 25% for new initiatives. This high upkeep cost diverts funds that could be used for more revenue-generating activities.
Modern airline management requires a deep understanding of traveller preferences and behaviour. Legacy systems typically also lack advanced data analytics capabilities, limiting the ability to personalise services and create targeted marketing campaigns. Without the ability to analyse customer data effectively, airlines miss opportunities to enhance traveller experiences and increase loyalty.
Security Risks and Fear of Uncertainty
Older legacy systems often need updated security protocols, making them vulnerable to cyber threats. A data breach has financial implications that can severely damage an airline’s reputation. Modern, secure systems are essential to protecting revenue and customer trust.
Also, there is a fear of uncertainty that comes with retiring and replacing one system with something new. Questions of how any upgrades will impact current operations, team performance, and customer satisfaction are paramount. Many leaders also worry about the downtime associated with implementing new systems, the potential for data loss during the transition, and the steep learning curve for employees.
Modern tech-enabled revenue management systems are said to be a game-changer in the industry. Technology aggregators, companies, and startups alike claim that these systems can unlock immense potential for airlines. By leveraging the latest technologies and advanced data analytics, these systems can help airlines optimize pricing, manage inventory, and improve overall revenue performance. Moreover, they offer greater flexibility, scalability, and ease of use compared to legacy systems. As a result, more and more airlines are turning to modern tech-enabled revenue management systems to stay competitive and drive growth in today's dynamic market.
Legacy infrastructures are not the only problem with the current state of revenue management in airlines. The organization design at most airlines, along with the slow march towards modern distribution, has also contributed to the slow adoption of technology in revenue management.
Distribution remains at the center of many developments today and as a critical aspect of revenue management, advanced pricing models require modern internet-based distribution channels to succeed.
Organizational Design Inefficiencies at Airlines don't help either
Organizational design in the airline industry is a crucial factor contributing to the inefficiency of airline revenue management. According to Sergio Mendoza, CEO of Airnguru, “The person responsible for pricing flight tickets rarely interacts with the person responsible for pricing ancillary services”. He further added, “Despite the evidence of correlation between the pricing and ancillary departments, revenue management analysts still work in silos with limited cross-team interaction”
This represents a fundamental issue with the organizational design of most airlines. While some airlines are hiring change management consultants to implement a more collaborative workplace, it will still take time for airlines to make significant changes.
Legacy Distribution Architecture Does Not Allow Advanced Pricing Models
Many airlines worldwide have made significant progress towards implementing NDC (New Distribution Capability) channels in their indirect distribution channels, with United Airlines, American Airlines, and Lufthansa at the forefront. United and American have gone an extra mile by withdrawing some fares from legacy distribution channels.
While the airlines have not explicitly stated why they have done so, one of the reasons could be that legacy distribution channels do not allow airlines to use continuous pricing functionality. However, direct distribution channels and NDC enabled third party channels allow airlines to use continuous pricing models. Despite this advantage and some legacy airlines pushing forward with NDC, the progress towards NDC has been slow. In 2021, traditional fare distribution is still heavily relied upon by most airlines. This often leads to poor forecasts, disregard for ancillaries, and outdated technology.
According to IATA, approximately only 22% of the 300+ airlines it represents (68 airlines) have been certified to use NDC to deliver at least some of the content, and they are all in various stages of technical development and maturity. Despite the fact that NDC enables airlines to adjust fares in real-time using dynamic pricing, sell ancillary services, and reduce distribution and commission fees, the lack of collaboration at the industry level has slowed down the adoption of NDC.
How Tech-enabled Revenue Management can add more value to Airlines?
Artificial intelligence and machine learning (AI/ML) have become ubiquitous daily, although most individuals are unaware of their impact. Using machine learning algorithms has become the norm in various industries, including social media, streaming services, and e-commerce.
However, the airline industry has employed AI/ML technology for several decades, with several crucial departments utilising it extensively. Specifically, AI/ML models are widely used in airline network planning, flight scheduling, pricing and revenue management, operations and crew planning.
The airline industry has experienced significant impact from AI/ML models. However, earlier implementations had limitations due to outdated architectures and development processes. The process of sourcing information from different sources and harmonizing data into one format made it challenging to integrate other models. Considerable manual intervention was necessary to keep models up-to-date with recent data. The task of harmonizing data is challenging, and cleaning the data requires time and talent, both of which are essential for airlines.
Fortunately, the industry has evolved, and most airline executives are now prioritizing investments in technology and the removal of old legacy systems. Airlines have realized that technology must be at the forefront of all innovations, and as such, most airlines have reserved a seat for a technology executive at the executive board level. This was not always the case, as in 2012, only Singapore Airlines and Turkish Airlines had a dedicated technology executive on their boards. Fast forward to 2022, and of the ten largest airlines in the world, eight of them have a dedicated technology executive driving innovation and technology within the airline.
For most airline boards, investments in technology remains a key strategic priority. This is further demonstrated by a recent survey conducted by Amadeus where they interviewed more than 1,000 technology leaders from across the travel industry, including 100 airline leaders - representing both Full Service Carriers (FSC) and Low-Cost Carriers (LCCs)
Airline executives are confident that AI and Machine Learning have vast potential. The Amadeus survey revealed that most airline executives believe that AI and ML will have a significant impact on airline operations, including revenue management, network management, harmonizing data, and improving customer experience by reducing disruption.
According to the same survey, a majority of airline executives at both low-cost carriers (LCCs) and full-service carriers (FSCs) believe that Machine Learning (ML) will have a significant impact on the industry in the next 12 months. FSC executives are particularly optimistic about the impact of cybersecurity and Generative AI on their airlines' operations. Meanwhile, LCC executives believe that cloud computing and digitalization will have the greatest impact, which is not surprising given the emphasis on streamlined and agile operations at LCCs.
From high ancillary sale potential to improving data for advanced dynamic pricing models to reducing passenger disruption, the application of AI and technology in revenue management has many benefits.
Technology executives are playing a crucial role in driving innovation strategies at airlines. Even industry associations like IATA are offering guidance to help airlines modernize their revenue management functions while incorporating modern distribution capability. This is being done through a set of guidelines and frameworks aimed at assisting airlines in their journey towards modernity.
IATA created a framework for airlines to follow towards Dynamic Offers. Dynamic Offers is a combination of three capabilities: dynamic pricing, continuous pricing, and dynamic bundling. It leverages on airline NDC and direct sales shopping workflows and enables airlines to provide relevant offers to their customers. IATA calls this a move towards a more customer centric air retailing as it encompasses advanced dynamic pricing techniques in all direct and NDC enabled distribution channels with a focus on dynamic bundling of ancillary services.
It created the Capabilities Matrix Framework to bring alignment in the industry's understanding of Dynamic Offers. Airlines may be classified into different categories at any given time, based on the capabilities of the various distribution channels or markets. This may force the airline to support different sets of rules in different channels.
The primary objective of the matrix is to establish a common understanding of Dynamic Offers and related terminologies. It also helps airlines and vendors to define their own path to enhanced capability.
The matrix's y-axis defines the level of sophistication for product determination and the overall contents of an offer. The sophistication level increases from bottom to top, moving from static and predetermined products to the determination of optimal and relevant contents of an offer using available and derived context and request data science, algorithms, and potentially machine learning and artificial intelligence.
The x-axis defines the level of sophistication for price determination of an offer. The sophistication level goes from left to right, moving from static and predetermined price points to the determination of an offer price using science and algorithms with machine learning and artificial intelligence.
The diagonal from bottom left to top right demonstrates the overall sophistication of offer creation in terms of the use of data analytics, machine learning, and contextual data. It also shows how much control an airline has over the current offer determination and creation process.
Contextual and relevant offers, continuous price points and total offer management are several key industry benefits within the scope of Dynamic Offers.
So what do these developments means? The industry’s growing technology orientation and the emphasis to change airlines operations post pandemic coupled with the renewed focus on airline revenue management and distribution function by IATA means CHANGE!
The growing idea within the airline industry is to align with a more customer centric approach in all areas of airline operations whether its distribution strategy where airlines are heavily investing to understand more about travellers and improving the sales channels or their revenue management strategy where they are looking to to provide on an infinite number of fare options to travelers. Whether its distribution or revenue management - technology is the catalyst that will drive this innovation.
As the revenue management function moves away from the current seat centric process to a more customer centric state, the underlying processes within RM will also undergo change from modest to drastic. Lets study each of them from forecasting to segmentation.
| Revenue Management Processes | Current RM (seat centric) | Future RM (customer centric) | Examples |
|---|---|---|---|
| Forecasting | Statistical Models | ML Based | Price sensitivity |
| Inventory Control | Fare Class based and pre-distributed fares | Classless Inventory | Pricing controls |
| Competitor Control | Benchmarking tools | AI-tools for predictive Analytics | Predicts what competitors will do |
| Distribution | GDS | Direct Channels + NDC | Rich contents and ancillaries to third party |
| Customer Segmentation and Product Creation | No personalisation + Static Bundling | ML Based personalised segmentation + Dynamic bundling | More flexibility in assigning customers to different segment based on behavior |
| Pricing | Static Pricing | Continuous Pricing + AI driven + Context + Real Time | Data from A/B Testing used for more AI Simulations |
| Analysis and Reporting | Historical data centric | Granular Data + Shopping analytics | More data sources |
Forecasting
Understanding travel demand is crucial for airline network planning and designing future schedules. Two key questions that help determine travel demand are:
1. Where do people want to travel to?
2. How many people want to travel between two cities?
If we can accurately answer these questions, we can optimize travel supply by determining the number of planes needed, where they should fly to, and how they should be scheduled. This can help build systems that estimate expected revenue and profitability, fine-tune retailing and distribution strategies, plan resources, and more.
Knowing the travel demand between different cities is important for various businesses, including aircraft manufacturers, airlines, hotel operators, travel agencies, local tourism-oriented businesses, and investors. However, there are about 250,000 city pairs in the world that have air service between them, and travel demand for them is constantly changing every month. This makes manual market analysis virtually impossible.
To overcome this challenge, AI/ML-based market potential forecaster models are used. These models use a combination of historical data, analyzing segment behaviors, and forward-looking shopping data to accurately predict travel demand. Using these models can potentially increase the average prediction accuracy, making it easier for businesses to plan and optimize their operations.
Inventory Control and Yield Management
Suppose you search for a hotel room and find one for €280. A week later, you look again, and now it's €345. It's the same hotel, the same location, and perhaps even the same room. What caused the price difference? Externalities are the main reason. A significant sporting event may occur, and hotels are charging more. Or it's the holiday season, and there's high demand. If the hotel doesn't consider these factors, they may lose out on potential revenue.
To maximise profits over the year, industries selling products with a fixed supply, like hotels, concert venues, and airlines, often use a pricing strategy called yield management. However, many businesses are turning to dynamic pricing, which uses data differently. Dynamic pricing adjusts prices in real time based on customer demand at that moment rather than relying on historical data like holiday dates.
So, what should airlines focus more on? Yield Management, Dynamic Pricing or both?
According to Sergio Mendoza, CEO of Airnguru, “Airlines have figured out yield management models, but it's the pricing models that remain unsolved.” He believes that while airlines have done a considerable amount of work on yield management and deserve a lot of credit for that, there’s still a lot of work to be done on the pricing model. He added, “The pricing model contributes nearly 60% of the total benefit of Revenue Management for the airlines”. Hence, focusing on the right pricing model will be key for the airline.
Pricing
Airlines have historically used a rule based pricing model where they rely on a maximum of 26 fare buckets. This has to led to airlines be unable to tap the full potential of unrealised revenue and customers do not received optimised prices. The limited number of price pointts and the predefined fares have worked only so far for the airlines and there’s a clear limitation in this process so Lets understand this through an example
| Fare Class | Fare | Allocation |
|---|---|---|
| M | $150 | 50 |
| K | $140 | 40 |
| Q | $120 | 30 |
| L | $100 | 20 |
In the above scenario on a certain flight, lets say there are four economy class fare class filed with the lowest fare L at a fare of $100 going all the way to M class with a fare of $150. In this scenario, if the allocated seats for Q Class are filled, the fares jump to $140 (to K Class). This situation is very common and now customers are expected to $20 more for that same economy seat. Even though the Q class is closed, there’s plenty of demand between $120 and $140. If a passenger is willing to pay $125 or $130 or $135, the only fares offered to the passenger will be $140. This system is unable to accurately determine passengers’s true willingness to pay (WTP)
In the world of static pricing, airline lose out on potential in-between fare class demand and price sensitive customers. On the other hand, customers also lose out comparatively lesser number of fare point availability and the high jump between fare classes. They could have paid the $130 for the economy seat but wont necessarily pay $140.
However, in the modern world of advanced pricing model powered by technology - this problem can be solved through dynamic and continuous pricing.
Continuous pricing is a new development in dynamic pricing that allows airlines to provide an unlimited number of price points. This means that the airline's pricing becomes more detailed and can adapt to changes in supply and demand at any given moment. As a result, it maximises the aircraft's capacity, leading to greater customer satisfaction (as they have more options to choose from) and increased efficiency
It utilises the New Distribution Capability (NDC) and direct sales shopping workflows to offer relevant deals to customers. But first, let us understand traditional airline pricing and continuous pricing through an example outside the world of aviation.
Airlines have traditionally used a step-function hierarchy of fares to allocate seats. This involves specifying fare ranges, or Revenue Booking Designators (RBDs), with each range assigned a letter of the alphabet. For example, “Q” fares for a particular route may define the fares range between $79 and $99, while “S” fares may include all fares between $100 and $129. When seats for one range are sold out, the selling fare jumps to the next range. This results in a stair-step pattern of fares, with each step a difference of $15 to $100.
The traditional system limits airlines to 26 fare levels in any market. For nonstop flights in one market, there may be a dozen fare ranges to cover the lowest fare (e.g. $79) to the highest for peak flights on holidays (e.g. $599) – a total variance of over $500. The ranges are closer together ($15-$20) on the low end but can get very wide ($100 or more) on the high end.
In contrast, Uber uses a dynamic pricing algorithm that taps into machine learning to vary charges on a continuous pricing curve. This allows them to vary the price of a ride by a dollar or less based on its algorithms. Some airlines can also vary fares granularly, but only by manually filing fares with pre-determined business rules (“fare rules”) for whom and when they apply. Because of the complexity of trying to approach continuous pricing, fare rules are currently used only broadly (21-day advanced purchase, for example).
Many airlines are now exploring the concept of dynamic pricing or continuous pricing as a potential replacement for the old static pricing technology. A continuous pricing curve would replace the step-function, with the next higher fare varying by just a few dollars. This would allow airlines to meet individual customer’s willingness to pay better, driving higher fares for those who value the trip more while offering lower fares for those who are more price-sensitive within a given range.
Distribution Strategy
In order to take advantage of continuous pricing capabilities, airlines must use an NDC enabled network or their own direct distribution network. Therefore, airlines that aim to improve their revenue management systems must also focus on modernizing their distribution networks.
In the era of modern airline distribution, enhanced retailing capabilities can provide significant revenue opportunities for airlines. According to McKinsey, by giving passengers more options, customised services, and flexible pricing, airlines can earn an additional $45 billion by the decade.
This is substantial in an industry where profit margins are typically narrow. As a result, airlines are becoming increasingly conscious of the potential benefits and airline executives are gearing up for this technological transformation. Modern Retailing will bring immense revenue potential to the airline industry.
Many airline executives believe new retailing capabilities can increase revenue by as much as 30%.
But there’s a challenge with NDC implementation, too. Not all airlines are on board with the idea, and the adoption of NDC has been slow. Hence, they continue to rely on legacy GDS infrastructure for distribution. Some airlines are not even thinking of NDC—they are perfectly fine in their current distribution model.
In the world of airline revenue management, this presents a significant challenge for airlines and analysts alike, as the legacy GDS system's legacy infrastructure needs to support dynamic pricing.
Customer Segmentation
Customers nowadays have more power and expectations than ever before. They demand digital experiences that are immediate, consistent, functional, and relevant to their individual needs.
To meet these demands, many retailers, including those in the travel industry, have turned to customer segmentation. However, the most successful retailers go beyond customer segmentation and strive to understand each individual customer. They treat each customer as a unique entity and provide them with personalized offerings, price points, and experiences that align with their specific needs and expectations. By prioritizing their customers and investing in long-term relationships, these retailers build a loyal customer base.
The airline industry, being a highly complex environment, may seem challenging to apply this customer-first approach. However, the good news is that while customers dictate how airlines do business, technology drives how airlines do business.
Passenger Disruption
Flight cancellations and delays are some of airline passengers' most common disruptions.
Airline analysts can use AI/ML models to evaluate the impact of a specific disruption, such as changing to a smaller capacity plane. Analysts can decide whether the action is acceptable by analysing the solution. If the tool is used and the results are applied, passengers will automatically receive a ticket for the new itinerary.
For instance, a businesswoman with an important meeting on Monday morning may choose to cancel rather than accept an alternative itinerary that arrives on Monday afternoon. On the other hand, someone on a week-long vacation in Italy may be willing to take a half-day delay. Experienced analysts can recognise these patterns and offer accommodation options that are likely to satisfy travellers. Machine learning models can also help evaluate whether the replacement itinerary would be acceptable.
Reduced disruptions result in more frequent and uninterrupted travel, which airlines greatly appreciate. This improves passenger loyalty and satisfaction, ultimately increasing airline profitability and yields.
AI and Technology opens up a multi-billion market opportunity
“There’s a second technological revolution coming for Airlines - with Machine Learning, Continuous Pricing and NDC at the centre of it.”
Sergio Mendoza, the CEO of Airnguru, believes that the airline industry is on the cusp of a second technological revolution. The first revolution came at the start of the last decade when airlines decided to take control over their distribution channels by introducing their own direct distribution channels, which allowed them to control over 70% of total bookings before the start of the pandemic and drive sustainable revenue growth. Mendoza predicts that the second revolution will come in the form of machine learning and AI, which will revolutionize airline pricing models. With the help of NDC-enabled indirect channels, airlines will be able to take full advantage of this market opportunity, which will be largely divided into two major revenue management functions: Airline Ancillaries and Pricing Models.
AI is expected to play a crucial role in improving airline ancillaries. Ancillaries already represent a significant portion of revenues for many airlines. According to industry experts and executives from Sabre, Amadeus, and Airnguru, AI can potentially enhance airline ancillary revenue by up to 15%. This could translate into a $40 billion opportunity for airlines.
AI's decisive role will also have a huge impact on pricing dynamics. It can potentially upend airline profitability by almost 1% today, which is roughly a $30 billion revenue opportunity for airlines. As the quality of data improves and legacy systems are phased out, better AI-based simulation models will not only improve the overall accuracy of the data but also enhance profitability. This could potentially contribute up to 5% to the airline's bottom line in the next 5 years - which loosely and conservatively translates to over $100 billion revenue opportunities by 2030 given the airline industry is expected to cross a $1 trillion revenue by 2030
While role of AI in pricing models is critical, it will require many pieces to fit in before it can make a more decisive impact.
According to Airnguru, optimizing prices is the most crucial aspect of revenue management in the airline industry. In fact, over 50% of its potential benefits come from optimizing fare fences and price levels. These benefits can represent 15 to 20 percent of net revenue in a segmented market. This is a significant amount, especially considering the industry's small margins. The quality of an airline's pricing practice determines whether it makes a profit.
The second most important aspect of revenue management is the optimisation of capacity or inventory allocation or yield management. This involves determining the seats allocated to each fare product and price level. This optimisation may represent between 4 and 8 percentage points of net revenues in a segmented market. Along with forecasting and overbooking, it forms the core of what currently available revenue management systems do under the supervision of revenue management analysts.
In addition to pricing, airlines can leverage the power of AI and ML in ancillary services to drive significant revenue growth.
Over the past decade, one of the major trends in airline marketing has been the consistent and significant growth of ancillary sales for air travel. A report published by IdeaWorks in 2022 revealed that airline ancillary revenue trends since 2013 indicate an average estimated annual growth rate of over 15%. This ancillary sales growth has come from individual ancillary sales (such as a la carte options) and airline-branded fare bundles.
Ancillary revenues are expected to reach $231 billion by 2028 growing at nearly 13% annually. In contrast, revenue passengers are expected to grow by 5% during that time, which presents an incredible opportunity for airlines to expand not only its ancillary services but its understanding of passengers through technology-enabled models as the ancillary penetration per passenger revenue is also expected to grow to 28% by 2028.
What’s also particularly interesting is that during the pandemic years (2020-2022), the ancillary penetration was very high. It is explained by the all-time drop in global passenger revenue during that time, but what it also shows is how persistent ancillary revenues are for airlines and their resilience against economic fluctuations compared to passenger flight revenues and hence the need for airlines to draw a clear roadmap that allows them to capture ancillary opportunities and position ancillary as a solid part of the airline revenue stream.
Previously, airline marketing teams did not give ancillary prices much importance, and prices remained static across airline regions. However, as sales and revenues have shifted towards ancillaries, the potential for more dynamic pricing with the help of AI/ML technology has increased.
AI/ML models can assist airlines in recommending optimal prices for ancillary products or services, using both supervised learning (based on estimated price elasticities, marketing segment, type of seat, position in the cabin, etc.) and reinforcement learning methods.
Michael Reyes, Senior Vice President, Product Management and Airline Platform at Sabre believes that AI can help increase ancillary sales by double digits.
As today’s travellers increasingly expect personalised travel experiences, the ability to create personalized ancillary offer is key to remain competitive. Virgin Australia, which returned to profitability for the first time in 11 years in 2023, partnered with Sabre AI driven intelligence suite to drive ancillary revenues for the company.
Change in the aviation industry may happen slowly, but it eventually happens. If you're an airline executive who wants to grow your airline, you can learn from pioneers who have already embraced new distribution models. Lufthansa was the first airline to invest heavily in NDC technology and experiment with dynamic offers on direct and indirect channels in 2020. Other major airlines like Air France-KLM and Singapore Airlines have also partnered with Amadeus and Travelport to adopt this technology. These examples demonstrate that while the adoption towards modern technology enabled strategies in distribution and revenue management has been slow, the early adopters will eventually race ahead of their peers in no time.
Conclusion
The rapid evolution of technology has paved the way for its integration into various sectors, including the airline industry. It is evident that advanced technologies like Artificial Intelligence and Machine Learning will play a critical role in every facet of airline operations.
In the last decade, the science and the art of revenue management has continued to evolve from a static step-hierarchy process of pricing to a more advanced dynamic model giving both airlines and traveller infinite fare options to choose from. This comes on the backdrop of improving technical infrastructure at airlines - which have often been a subject of scrutiny for its characteristically slow adoption of technology on its platform, infrastructure and operating systems.
With a renewed focus on improving distribution and revenue management strategy at airlines, airlines executives in conjunction with industry partners have drawn a future plan for closer and rapid adoption of technology. The post pandemic world represents a world full of capacity, supply chain problems for airlines although reduced capacity and surge in demand has helped airlines with improving unit revenues. Recognizing the need to improve in house operations amidst the global supply chain issues, airlines are going full steam by implementing technical improvements on their operation with revenue management and distribution at the core of this development.
More and more airlines are partnering with technology aggregators to implement AI and ML on revenue management processes with a focus on ancillary revenues and pricing models. Speaking to technology executives from around the globe, this development opens a billion dollar opportunity for airlines as they move ahead in a world that increasingly relying on technology for forecasting demand, improving pricing strategies, improving customer segmentation and offers, and optiming inventory.
Although the airline industry has been slower than other industries in adopting modern technology, the next few years will see a significant shift in this trend. The impact this technological revolution will have on millions of travelers will be immense.