Client
A globally operating aviation group with a total of more than 500 subsidiaries and affiliated companies. With a diverse company portfolio that consists of network airlines, point-to-point airlines and aviation service companies, the client is a major part of the first truly global airline alliance, and has won many awards at the global level as well.
Challenge
One of the requirements for this project was to deliver dynamic booking curve estimation as well as buy-down adjusted fares, for which the modelling was done on the basis of a PhD experiment. The core algorithm for this was created on a machine with 1 TB RAM, and the algorithm had to be operationalized and rolled out in a phased manner. Moreover, there were many complexities around operational compatibility using the approved technology set. The primary concerns were around multinency - to be used across the group - and performance (in particular nonfunctioning requirements or NFR).
Solution
Mindtree designed and implemented an iterative plan for the revenue management system. This covered all phases of implementation, including rollout and non-invasive integration. The next step involved identifying options to productionize, as well as redesigning and rewriting the core algorithm without sacrificing accuracy. Once this was done, the price elasticity curve was generated, and buy-down fares were served to downstream systems. As a result of this, multitenancy was factored into all the layers, including data stores, services, as well as UI with brand and self-service capabilities.
Benefits
As a result of design the revenue management system, Mindtree was able to deliver the following benefits:
- New platform with new capabilities that couldn’t be performed with the old platform - including the option to accept/reject buy-down fares, as well as UI for final manual tuning
- 10x increase in revenue over what was the cutoff to make the program commercially viable
- A platform that was scalable across the three scalability axes – region, layer and horizontal
