Cargo airlines enjoyed a period of high revenue—driven by scarce capacity—during the pandemic. But after the boom of the past three years, yields are gradually falling from the 2021 peak. Belly cargo capacity is recovering, and demand is softening, leading to uncertainty as cargo airlines brace for the risk of a “back to normal” scenario.
This raises the issue of how cargo airlines can make sure that the “back to normal” is not a “hard landing”. In this environment, a new approach to revenue management could be the key that allows airlines to adjust their commercial strategies and continue to benefit from opportunities in the market. Transport By Boat
Over the past three years, the cargo market has been capacity driven and airlines with significant capacity pulled ahead of competitors. Recently, there seems to be a transition back to a demand-driven market: yields have declined, demand has slowed, and belly capacity continues to recover (Exhibit 1). Moving forward, rates are expected to decline further, although will likely remain above 2019 levels.1 McKinsey analysis based on IATA and WorldACD data.
What this means is that new ways of working may be required for individual cargo airlines to remain competitive in this changing market. As belly capacity returns, the market will likely become increasingly competitive, and airlines that don’t have a robust commercial and revenue-management strategy in place might lose out and see their yields diminish faster than the average.
At the same time, many cargo airlines have invested considerably in their digital strategies since the pandemic began. In particular, online sales have boomed, and consequently, cargo airlines have access to much more data than was possible three years ago. A recent Freightos WebCargo report found that digitized air capacity across the industry reached 57 percent in Q1 2023, compared to 38 percent in Q1 2022, and only 3 percent in Q1 2019.2 “WebCargo digital air cargo (DAC) monthly,” WebCargo, February 2023.
Taken together, the turning point in the market and the rise of digitization in the industry point to today being a crucial time to formulate next-generation revenue management for air cargo.
This article details three areas where cargo airlines can focus their efforts to re-think revenue management, specifically by relying on accurate forecasting to form actionable insights; using real-time monitoring for fast decision making; and taking a customer-centric approach.
Forecasting demand and supply is the starting point for a cargo pricing and revenue-management strategy. However, cargo demand is extremely challenging to forecast, for several reasons.
First, booking tends to be a last-minute process and late bookings are a consistent feature in this environment. Typically, two weeks before departure, less than 40 percent of an airline’s capacity has been booked. Second, the market is volatile. Air freight is often used by shippers as a last-minute restocking option, which depends on many economic factors, so the need for air freight can change almost overnight. Third, the air freight market is composed of dozens of industries, and thousands of commodities, each with different drivers that make demand difficult to predict.
But, airlines can leverage technological advances to improve demand forecasts and deal with volatility. The availability of more granular data sources, and the advance of Machine Learning (ML) algorithms, make it possible for cargo airlines to pursue better demand forecasting solutions to gain deeper insights—and ultimately make more nimble revenue decisions.
For instance, due to the increase in online sales, cargo airlines have more data available about their customers’ behavior. This is particularly the case for airlines that have their own sales portals. Through digitalization, the air cargo industry has an opportunity to build a 360-degree view of demand across the entire customer journey which includes data that is above the sales funnel, such as which flights customers search for, lead times, how the cargo request was made, how long it took to fulfill, and if there was a cancelation or modification. Airlines can also look at step-based conversion rates showing how the airline performs at each stage of the sales funnel (discovery, flight selection, product selection, price offer, etcetera). Having all of this data in one place means that cargo airlines can improve their customer experience: better understand what customers want, and when they are likely to want it. This is the type of insight that companies in B2C industries, such as passenger airlines or hotels, typically have access to and cargo airlines could consider using a similar approach and leaning into the e-commerce aspect of sales.
It’s clear that Artificial Intelligence (AI) and ML are transforming sectors and industries across the world—and cargo airlines could harness the power of AI to better predict demand. A McKinsey Global Institute study identified that the travel, transport, and logistics sector has the most potential for incremental value from AI, amounting to $1.8 trillion in value. Within this sector, roughly half of this value is likely to come from commercial applications such as customer service and pricing.3 “Notes from the AI frontier: Insights from hundreds of use cases,” McKinsey Global Institute, April 2018.
Cargo airlines are well positioned to increase forecasting accuracy through AI. For example, AI could make sense of the thousand or more commodities, as well as their inter-dependencies, within the supply chain. For instance, AI could determine how trends in raw materials and semi-manufactured products in one country could lead to a growth or decline in specific finished products in another—and how this would influence cargo demand.
There are a few pointers airlines could keep in mind when using AI for demand forecasting. It is important to select the right data as input, as it needs to be sufficiently granular. And using a blend of internal and external data can lead to greater forecasting accuracy as early as two weeks out, despite very few bookings being made at that time. Internal historical data is very important for improving forecasting quality, which tends to be overlooked.
Considering that the accuracy of ML algorithms increases with the amount of quality data being used, airlines will probably find that AI-enabled forecasts get more accurate over time. One cargo airline managed to improve its ability to predict demand significantly through the use of AI. Initially, the AI tool reduced the airline’s forecasting error from around 20 percent to 14 percent, and once it went live it continued to improve in accuracy.
The airline found that the AI model was much better at predicting seasonality patterns through multi-layered algorithms than traditional models. This allowed it to predict volume patterns to a high degree of accuracy from one to four weeks before departure. Furthermore, incorporating data on trends such as booking cancellations improved final volume predictions.
There are other untapped opportunities to leverage internal data, such as by predicting no-show rates for bookings by lane and by customer. Another airline followed this approach which led to better capacity management and, ultimately, improved profitability. Predicting cancellations allowed the airline to increase “overbooking” while still controlling for the risk of penalties (Exhibit 2). This, together with other specific use cases, helped to uplift load factors by around 8 percent after a 12-week pilot. Based on this success, the airline was able to identify potential network-wide savings worth tens of millions of dollars.
Finally, airlines need to make forecasting actionable. There is often a trade-off between pursuing the perfect forecast and moving ahead and making the right decision based on the information at hand. A fact of life is that pricing teams need to be prepared to be wrong, some portion of the time. The key is to make better-informed decisions at the right time.
Booking curves have changed post-pandemic and these changes are likely to accelerate, given that the market is at a turning point. This is why it is now more important than ever for airlines to continuously monitor capacity booking, and be even more proactive when it comes to simulating demand—and do it more frequently.
For example, a cargo airline’s booking curve changed significantly over the course of the pandemic. In 2021 and 2022, the pace of booking accelerated and exceeded 2019 levels at the two-week before departure point. At one week before departure, the share of weight booked was considerably higher compared to pre-pandemic levels, resulting in less urgency to fill the flight at the last minute (Exhibit 3). In this particular example, any last-minute price reductions would probably have been less effective than in previous years, as most of the airline’s customers were already booking closer to the flight date. This example also illustrates the last-minute nature of the industry, where more than half of a flight’s capacity is often booked in the last week.
Given this operating environment, revenue-management decisions are highly sensitive to changes in demand and supply. Therefore, in the current uncertain market, airlines could improve revenue management by evaluating supply and demand in real time, at a granular level.
Airlines could, for instance, use dashboards to monitor how flights are being filled, and where the biggest opportunities lie, based on what capacity has been booked and what is expected, at any point in time. Forecasts, or estimates of what is expected, could be based on lagging indicators (historical data) and also on the leading indicators made possible by the trove of new data that comes with digitizing capacity.
Airlines could also be more proactive when it comes to stimulating demand. For example, an airline could evaluate how consistent customers are in providing their cargo and compare that to real-time data. If a customer has consistently provided cargo on the same lane 12 days before departure, but the revenue-management team notices on day 10 that the customer has not provided the cargo—it is time to act. Sales teams could contact the customer and ask what they need, if their needs have changed, or how the airline could tailor its products. Processes need to be in place to alert the sales team if a flight is not filling up as expected or to signal that they need to review pricing guidelines if a flight is filling up too quickly.
There is significant value at stake from acting at the right time. In an industry where critical decisions are made in the last week before a flight, airlines need to be agile and well-informed. Typically, yield volatility is at its highest in the last week—the spread is relatively stable until five days before departure when it starts to widen (Exhibit 4). Often, the final cargo shipments that are booked to fill a flight have the highest yield. This means that pricing and revenue-management decisions made in this timeframe have the most potential to either be very profitable or leave value on the table.
McKinsey analysis suggests that if airlines can focus on the yields that are below average in the last week before departure, and improve this by 30 percent, then overall revenue can increase by between 7 and 8 percent.
In today’s market, airlines need to be extremely fast to respond to shifts in supply and demand, and make the most of sudden opportunities, but decisions need to be carefully calculated. Airlines would benefit from having the data and analytics in place to enhance decision making, and also by having internal processes and a decision framework in place to coordinate among sales, network, and revenue-management teams. These measures would help to break down silos and ensure swifter response to market conditions.
Digital can be a great lever in achieving this, but only if airlines are equipped to successfully leverage the data that is available. For instance, if a flight departs on Thursday but most customers are searching for a Friday departure, that is critical information that needs to be passed on to the network team. Feedback loops are essential between teams for understanding the customer journey and gathering data on each aspect of it.
Building on the insights gathered about customers, their needs, and stages of the customer journey, airlines can form a comprehensive revenue strategy that puts the customer front and center.
Typically, cargo airlines base revenue-management decisions on flight profitability. They set price entry conditions based on expected demand and operating cost—as a basic principle.
However, airlines may struggle with managing revenue at an account level. For instance, decisions around how to price a large account’s cargo on a high-demand flight—while the same customer also provides volumes on other low-demand flights—are not so clear cut.
This is something that passenger airlines have figured out within their corporate sales programs, essentially looking at the longer-term customer relationship and the incentives offered, and overlaying that with the predictions generated by their pricing and revenue-management models for any individual flight.
To maximize effectiveness, cargo airlines could keep the following three lenses in mind when making revenue-management decisions: flight profitability, customer value, and product offering (Exhibit 5).
At its most basic level, revenue management is often based on destination origin or flight profitability. This approach relies on models, fed by historical data, that show demand and supply levels—essentially, how fast the flight is filling up. This is mostly valid for lower-demand flights and/or flights that have few large accounts.
However, for flights with large accounts and/or high demand where many shipments may compete for the same capacity, airlines could consider paying attention to the customer lens. This entails looking at the network-wide contribution of any key account. By doing so, airlines could prioritize customers with whom they have a strategic fit, for instance there may be a significant overlap between the customer’s cargo and the airline’s network in that they represent growth in markets that the airline may find attractive.
The airline could run simulations to prioritize high-value customers, based on profit contribution by origin destination across the network, augmented by a view on projected growth by account. In this way, the airline could make revenue-management decisions that would prioritize clients that are likely to grow in line with the airline’s network.
Products and deal structures could then be purpose-built and customized to clients’ requests. Again, for this to be effective, revenue-management and sales teams would need to work together and have a feedback loop in place. Sales staff could be more proactive in identifying customer potential, beyond traditional revenue-management guidelines. What may prove most effective is to pair up sales and revenue-management teams for a specific large account, to run the necessary simulations and bring fresh perspectives to bear.
In terms of product profitability, airlines could take into account products and services they provide, such as special handling verticals, which have their own value drivers and cost drivers. Other ancillaries could also be pursued—such as purchase of sustainable aviation fuel or flexible capacity options. Thanks to booking portals, cargo airlines now have a better understanding of what their customers want, and may be willing to pay for. Perhaps the industry can take inspiration from how passenger airlines are able to use passenger data to anticipate needs and preferences. Cargo, by definition, is B2B but an approach that leans toward B2C in terms of understanding customer profiles and personas can help cargo airlines to re-think their offerings and pricing structure in the digital age.
Here, AI/ML can be helpful in managing large amounts of data as inputs that can segment customers’ potential, not only based on historical revenue but also on a wide range of variables including booking patterns, route portfolio, cargo density, and predictability. For instance, a medium-sized customer with a diverse portfolio of routes, products, and verticals might have greater potential than a large customer operating in a small number of markets and with high share of “no shows”. This type of segmentation—that allows airlines to prioritize high-potential customers—is extremely important as customers are likely to review their booking needs and capacity-acquisition strategies in the changing market.
In a volatile and uncertain market, where yields are likely to decline, each cargo airline will have to act wisely to protect its position. This requires airlines to be agile, make decisions quickly, and to implement new ways of working. Cargo airlines looking to re-think revenue management could consider the following three priorities to guide them on their journey:
Naturally, the recipe for optimizing revenue management might be different for each airline. A good starting point could be to identify a long list of commercial use cases grouped according to three objectives: to improve capacity utilization, increase yields, and enhance customer experience. Each use case could be trialed and tested to prove its value on a small scale. But airlines shouldn’t stop there: what comes next is key. Choosing the right use cases to scale up is critical, and to do so while continuing to build digital capabilities and work in an integrated manner across the business.
Despite the challenges that cargo airlines could be facing in the coming months, there are opportunities to improve revenue management to remain competitive and profitable. This depends on a new approach to digital, and on ensuring that the customer remains center stage when making revenue-management decisions.
Soufiane Daher is a consultant in McKinsey’s Amsterdam office, Ludwig Hausmann is a partner in the Munich office, and Mark Williams is an associate partner in the Atlanta office.
Clearance Agent The authors wish to thank FLYR Labs for the examples on which this article is based.