A Guide to Reporting on Airline Dynamic Pricing

     

A-Guide-to-Reporting-on-Airline-Dynamic-Pricing-RMS-Blog.png

Reporting has been a challenge for providers of airline dynamic pricing (“revenue management”) systems from the initial launch of the systems decades ago. Naturally, dynamic pricing systems initially focused on their proprietary models often hidden behind complex algorithms, and business reporting seemed an afterthought. Somewhat slowly, vendors supplemented their systems with reports and alerts, mostly directed toward “flags” that might trigger analyst interventions:

  • Are bookings progressing on trend?
  • Is booked load factor higher than same time last year?
  • Is the system properly closing lower fare classes when the booked load factor hits a high level?

However, suppliers soon found that every airline had a different reporting requirement, and each defines its own reporting needs and its own triggers. At airlines I’ve worked with, we ended up developing our own reporting capabilities, specifically addressing our market and competitive issues. To meet diverse customer reporting needs, the revenue management supplier must offer huge data sets (lots of options) in a highly flexible format that enables each airline to report in its own way. As data continues to get “bigger,” new reporting requirements continuously arise.

Thus, as dynamic pricing evolves and expands, reporting will need to keep up. Let’s construct a clear map of reporting needs for a dynamic pricing system.

Diagram_1-3.png

Insight Into Customer Purchase Behavior 

Increasingly, airlines are integrating booking data with other customer-related data sets to better understand customer behavior. What is actually driving changes in booking behavior? Of course, although this data is outside any airline revenue management system, reporting must easily pull in such relevant data for use in analyzing customer behavior. Many revenue management systems, for example, now integrate competitive fare tracking into booking analysis. Airline websites can now track the sales funnel:  visits versus searches versus “shopping cart” versus completed bookings. Ancillary purchases and channel-specific performance can also provide additional insight into overall customer behavior.

  • Have there been any changes in the traditional sales funnel?
  • Are customers using one channel or more?
  • What ancillary services are most popular? With which segments?
  • Is market share stable?

Transparency Into Model Assumptions/Forecasts/Optimization

Often the more sophisticated the pricing model, the less transparency is provided in what’s going on inside the model. On the other hand, reporting on how the model is working is critical for analysts to truly understand how dynamic pricing working. Too often, the model is treated as a “black box” and not well understood by analysts.

  • How accurate are the model forecasts of customer behavior and demand? Are the forecasts converging quickly on actuals?
  • How is the forecast model responding to changes in trends?
  • How is the optimization algorithm changing based on changing demand?

For airlines, systems that allow analysts to manage inventory control directly, while often much less sophisticated than a multi-variate forecast model and a multi-dimensional optimization algorithm, offer far more transparency.

Triggers for Analyst Interventions

No airline dynamic pricing model can be left on its own – they require analyst oversight and management. In fact, reporting offered by many airline dynamic pricing systems focuses on possible opportunities for analysts to intervene, that is, to override the model. Triggers include booking trends by flight, by date, and by fare class. Are bookings departing significantly from trend? If so, the model may take time to incorporate such new information in its forecasts and thus may benefit from an intervention – adjusting the model forecast. Similarly, “alerts” are established to identify flights that defy normal inventory allocations: flights that are heavily booked, but still offering low fares or weak flights that have restricted sale of low fares.

  • How are bookings performing versus historic trends?
  • Are bookings following the normal booking curve?
  • Are flights with high booked load factors properly closing off low fare availability?
  • Are flights with low booked load factors continuing to be “open”?

Transparency Into Analyst Interventions 

Reports designed to identify opportunities for analyst intervention are far more common than reports that track whether such interventions add value. Many airlines find that analysts intervene too much – “flags” drive inappropriate overrides. Because analysts may intervene inappropriately in models, monitoring interventions is critical. The system should track the frequency of interventions and require analysts to record the rationale for their adjustments. Then it should follow through and determine the value of the intervention in terms of forecast accuracy or inventory allocations.

  • How frequently did the analyst override the model? What was the rationale?
  • How did interventions impact forecasts and allocations?
  • Is intervention “sticking” or is the model migrating back to algorithmic results?

Performance Measurement

Finally, reports need to monitor overall revenue performance, including an assessment of the “value add” of both the dynamic pricing model and analyst interventions.

  • Is forecast accuracy acceptable? By flight, by class, and by O&D.
  • What is the forecast value-add for:  Model vs. “Naïve” Forecast and Analyst Intervention vs. Model?
  • What is estimated “spillage” and “spoilage”? Is performance improving?
  • Is unit revenue, including ancillary revenue, increasing?

Dynamic pricing models present unique reporting needs. There are reports for each step in the big data/analytics optimization process. Airlines, as leaders in dynamic pricing, need to consider the appropriate reporting for each step in order to fully exploit their sophisticated systems.


Explore five areas to focus on when using big data analytics to optimize to optimize seat allocation across fares and passenger itineraries. 

Comments

Subscribe Here!