OUTPUTS > PILOT / USE CASE 4

Ancillary pricing for airline revenue management

SIMULATION AND GENERATION OF OPTIMAL PRICES FOR ANCILLARY SERVICES FOR AIRLINES

In the airline industry ancillary services is everything in addition to the actual air transportation such as extra seats, baggage, hotel bookings and many more.

Ancillaries are increasingly becoming a large source of revenue for airlines.

Finding optimal pricing for each customer and ancillary in every situation is a key issue, with not much research conducted so far.

Summary

The computation of optimal prices for ancillary services in the airline industry leads to a major improvement for airlines and customers.

An increase of the airline’s revenues will be accomplished when shifting from a pre-determined pricing to an advanced method offering customers a more passenger-centric air traffic experience, with the most useful offers at the precise moment.

Key aspects of future ancillary pricing are:

  • Merchandising any service item.
  • Controlling the entire offer.
  • Optimizing prices according to the customer’s interest.

Challenges

Traditional revenue management fails to address the complexity of ancillary pricing for various reasons:

  1. Data volume: there are millions of different transactions and price decisions every day.
  2. Complexity: numerous factors are influencing the willingness to pay of airline customers.
  3. Speed: a pricing response is required in real-time, i.e. milli seconds.

Objectives

The goal of the Use Case: Ancillary Pricing for Revenue Management is to develop a set of dynamic pricing solutions for airline ancillaries. 

The approach includes 3 steps: 

  1. An analysis of current ancillary sales and hidden sales pattern.
  2. A machine learning algorithm in order to generate price quotes for selectable ancillaries.
  3. Simulation of potential revenue increase through dynamic instead of static pricing.

Methodology

Techniques and procedures

The main activities and operations implemented are:

  • Data collection.
  • AI methods.
  • Colletion processes.
  • Interface.

Resources

The software architecture is divided into:

  1. Batch layer.
  2. Service layer.

The following software was used for this use case:

  • Cloudify.
  • Docker.
  • Python.
  • Java.
  • R.
  • h2o.

Results

For the field of activity (business environment)

Main achievements are:

  • The implementation of an ancillary data storage which is capable of storing the incoming ancillary data in a way that allows easy exploitation. This comprises an architecture supported by toolstack ecosystem as well as a prototype which will showcase the transformation of historical data.
  • The development of an optimization method that includes algorithms using state-of-the-art methods of artificial intelligence such as:
    • Machine Learning.
    • Random Forest.