FLYR for Hospitality Now Available on Oracle Cloud Marketplace Arrow

Resource Hub / Airlines / Revenue Optimization / Forecasting at the Speed of Change

Forecasting at the Speed of Change

Volatility Everywhere

In today’s volatile operating environment, commercial teams find themselves having to ignore legacy forecasting systems and get creative with the limited tools and data at their disposal. From reverse engineering demand through promotional sales to see where customers are willing to travel, to trying to understand network patterns and markets by hand in convoluted excel spreadsheets that rely on highly subjective interpretation.

A competitor enters your market, a storm is headed to Florida, a pandemic rises, or borders are closed as a countermeasure. These are just examples of the endless scenarios airlines encounter more frequently than ever. But why does it take so long for airline analysts to forecast a decline in demand and adjust accordingly? The limitations of tools at their disposal result in excessive time spent to manually understand the operational environment, struggling to keep up with a constant stream of changes and new market conditions.

To make matters worse, legacy forecasting solutions only update their load factor and revenue forecasts at a limited number of points in time for any given flight or market. They are limited to information such as total capacity, bookings, and available inventory, none of which will quickly surface clear trends at the required level of granularity. Even solutions that used to be best-in-class are limiting their forecasts to approximately 24 outputs throughout the booking cycle of a flight.

Fast forward to today, deep into a global pandemic, airlines’ forecasting is more critical than ever. Our airline clients speak frequently about their top-down revenue forecasts failing them now that deployed capacity, GDP, and competitor pressure are impossibly hard to correlate to future performance. Meanwhile, automated, self-adjusting, bottoms-up forecasting in this environment remains an impossibility for legacy systems, providing inadequate reliability at the flight or market level.

To effectively manage, react to and reconfigure their networks, airline executives and commercial teams have to look beyond to-date or flown revenue metrics. They require highly reactive, real-time forecasts that update automatically when new data becomes available.

FLYR Labs’ Cirrus Business Intelligence solution solves these aforementioned challenges by giving the airline a new superpower, Forecasting at the Speed of Change.

Data Silos Will Hold You Back

Data sparsity, the problem of not having enough data to form accurate forecasts and decisions, prevents the on-time identification of trends, let alone doing so in real-time. Legacy forecasting algorithms used within existing revenue management solutions heavily rely on flight-vs-flight or market-vs-market comparisons that compare this year to last. Relying on such a narrow scope will result in only very few data points feeding into the models that are tasked to forecast the future, making them naturally slow to react to changes.

Another factor that adds to the data sparsity challenge is the data silos between different commercial teams at an airline. Key stakeholders within revenue management, network planning, revenue accounting and marketing teams use different data to inform their decisions and refresh this information at a different pace. It is essential that decision-makers at an airline have equal access to critical data and insights. Especially during pandemic-induced market volatility, when competitors change their strategy, or demand patterns change because of events.

Airlines should not limit themselves to flight-vs-flight or market-vs-market information on a year-over-year basis. Instead, they should look at the complete context of their network and derive insight from similarities across their operation to increase forecast accuracy.

With Cirrus Business Intelligence, identifying similarities across markets, origins, destinations, flight durations, departure schedules, and competitive pressure is now possible. Cirrus’ Deep Learning algorithms are able to pick up on trends before they are clearly visible in data-sparse subsets of the airline network (e.g. a single market).

market embeddings example image
Real-world, anonymized example of Cirrus’ ability to understand similarities between airports. Cirrus was never supplied with geographic coordinates while establishing such mapping.

The system directly consumes and considers a vastly greater array of data compared to legacy solutions. For example, the Deep Learning models directly consider competitor fares, schedules, inventory, and events, as well as the airline’s promotions and observed search traffic across channels. Even revenue accounting data is ingested to understand the total revenue opportunity beyond the income from the base fare.

A New Playbook

Establishing a new playbook, enabled by advances in revenue, load, and pricing strategy forecasting will transform the role of analysts within the airline’s commercial team. Not only will they have the tools at their disposal to focus their time where revenue returns are greatest, they will drive closer collaboration between other teams by serving as a centralized source of knowledge and insight.

With the right insights readily available and continuously updated, airline teams can start to answer difficult questions that used to be guesswork, in real-time.

  • Revenue Management: By escalating where day-of-departure revenue forecasts are declining, analysts can prioritize their efforts and spend time where dollar-for-dollar impact is greatest.
  • Marketing: By highlighting markets where day-of-departure load factor forecasts are weaker than desired, but where passenger yield is high, demand can be actively stimulated for stronger returns.
  • Network Planning: Determining a flight’s departure time can be derived from a deep understanding of how time-of-day will impact final revenue performance. Similarly, a more accurate understanding of revenue and load-factor forecasts for the day-of-departure enable much earlier up- and down gauging of equipment or a more accurate allocation of capacity by changing flight frequencies.
  • Revenue Reporting: No more manual exports or guesstimates. Cirrus provides accurate, reactive, bottoms-up revenue forecasts of individual flights and enables instant rolled-up views across markets or the complete network, including consideration of total revenue (not just fares).

At FLYR Labs, we envision our Cirrus solutions to empower anyone within the airline, from revenue analysts to network planners, and from commercial leaders to the CEO.

Similar stories

Early adopters of AI across industries have already realized and demonstrated its significant benefits, but the travel industry is still slow to embrace this cutting-edge technology to improve business operations, customer experiences, and revenue performance.
Why this legacy airline is transitioning to a low-cost model, and how it will continue to be agile in a changing industry.
While the aviation industry as a whole has been cautious to make the switch from legacy tools and techniques, digital-first airlines have unlocked incredible value across commercial functions by leveraging the latest in AI.