Context is everything: decoding post-pandemic demand through new data points
Airlines need to sell the right product to the right customer at the right time and at the right price. Any additional information that can be utilized to make that happen provides much-needed context to those pricing decisions.
In traditional revenue management systems, observations can be made about demand leading up to a flight departure, the capacity of a flight or route per day, and the impact of holidays or regional and global events. However, the legacy science, systems, and processes that used to drive commercial decisions for airlines are unable to keep up with the new, volatile environment that the industry has found itself in.
Even before the pandemic, adaptability was key for airline strategy. Anything from extreme weather conditions to a marketing initiative from a competitor can impact demand and make forecasts redundant. Being able to apply context to pricing puts airlines in the best possible position to gain revenue uplift.
Adaptability was at the core of how Cirrus was designed and enabling airline analysts to understand and interpret context means they have the confidence to use our insights to make better commercial decisions.
How does Cirrus provide extra context to pricing decisions?
FLYR’s Cirrus technology maps every piece of commercial data or input to all others, even in the noisiest or most data-sparse environments. This includes bookings, schedules, competitors’ schedules, pricing or capability, ancillary transactions, and third-party demand signals.
By mapping and understanding this context within our airline customers’ networks, we can accurately forecast future performance and set appropriate pricing strategies, even for markets that have not been flown before, or which have seen dramatic disruption in their environment.
Key areas we provide context on are:
- Forecast – Airline forecasting is more critical than ever. However, legacy systems are limited by data scarcity resulting from the use of a narrow scope of metrics heavily reliant on flight-vs-flight and market-vs-market information, none of which will quickly surface clear trends at the required level of granularity. Cirrus solves this challenge by giving airline executives and commercial teams a new superpower, Forecasting at the Speed of Change. By identifying similarities across markets, origins, destinations, flight durations, departure schedules, and competitive pressure, Cirrus’ Deep Learning algorithms can identify trends before they are visible in data-sparse subsets of the airline network. With the right insights readily available and continuously updated, airline teams can, in real-time, start to resolve complex questions that used to be answered with guesswork.
- Geography – Cirrus tracks the length of each flight segment and which airports are operated into and out of. Similar airports are mapped via embeddings, for example, the South Florida markets or the LA 5, to better predict pricing. This analysis can be done with far greater insight and automation.
- Seasonality – Cirrus models seasonality with any persistent shifts in demand reflected. So, in the post-pandemic world, we can assess what the new seasonality may look like, as we cannot rely on previous years’ data. Using legacy systems can work can take months, as airlines have access to this information but are unable to utilize it efficiently. Cirrus uses machine learning to absorb this data and turn it into a richer set of insights than is possible using legacy ruled-based systems.
- Itinerary – Cirrus can forecast Willingness to pay (WTP) for a given market and differentiate WTP between customers looking to fly non-stop, on convenient single connection routes, or on more complex journeys.
- Speed and Granularity – In traditional revenue management systems, if system demand falls short of forecast, the airline might take more than a week to observe and then react through broad system-wide changes in parameters. Instead, every day Cirrus learns about demand on an individual flight level, observing bookings continuously, forecasting those changes daily, and amending prices accordingly.
- Competitor capacity and activity – Cirrus utilizes real-time competitor pricing when making revenue optimal decisions and when displaying market trends. The platform also knows when an airline’s strategy is optimal compared to a competitor’s actions. Cirrus’s competitive strategy is complex and highly depends on context, for example, the type of competitor and how many competitors there are for a certain route. For example, if you are only up against one non-stop competitor on a route, the best action may be to not match on price. However, if there’s another competitor in the market, it might be suboptimal to completely ignore competitive pricing. Airlines can price closer depending on the competitive landscape, in a game-theory-style strategy.
How can analysts manage these context data points?
Cirrus brings analysts the ability to combine all these context points within a single interface. Previously, analysts would hold historical and current performance data in separate places, often managed through spreadsheets, and then use another program to influence decisions. In Cirrus, analysts have access to all the above context data points and can see forecasts, revenue, load factor, and competitor positioning to review how pricing compares.
However, Cirrus is not 100% self-driven. If an analyst needs to influence a price, like when looking at competitor activity, they can choose to meet a competitor’s pricing by only 20% or 50%, rather than 100%.
The data afforded through Cirrus is incredibly rich. Analysts have access to 150 metrics within the user interface, and if they want to explore any specific metric, they can click on it and see a trend graph. The ability to overlay this amount of rich, granular details together is rare in legacy revenue management software.
How does context affect the baseline expectation and resulting revenue?
For every flight or market in an airline’s network, Cirrus generates a ‘baseline expectation’ for revenue and load factor. This baseline is a much more accurate alternative to year-over-year metrics or historical averages, which cannot be relied upon due to COVID-19. Once established, the baseline can objectively compare actual revenue and load factors as it builds over time.
We can characterize how similar or dissimilar airports, routes, departure times, and competitor presences are across the network in the form of a ‘vector’. Such vectors help produce market – and flight-specific – forecasts and strategies that perform even under the most volatile conditions.
FLYR runs true A/B tests to compare flights that remain underpricing control of a legacy revenue management solution to those managed by FLYR. Using this comparison, we can measure revenue and load factor uplift. Typically, we see a drastic incremental revenue lift of 5-7% amongst our customers.
For more information about FLYR and how we use advanced and intuitive technology to understand context and help airlines achieve their ultimate revenue potential, contact us today.