Morbi vitae purus dictum, ultrices tellus in, gravida lectus.

In his case study presentation, Sémi Gabteni, Former Head of Products and Science in the airline industry explains:

  • How to position use-cases to obtain stakeholder engagement
  • How to overcome behavioural barriers when driving change
  • How to align IT and business strategy for successful transformation

Semi has significant experience in the airline industry, and argued that pilots have a lot to teach us about developing a data-driven culture in our organisations. 

Fuel efficiency is one of the biggest issues for any airline, with not only financial but also environmental implications.  Data can be key in reducing these costs, with best practice adoption from pilots growing from 19% to 79%. 

Pilots and operations tend to have very different perspectives.  For instance, pilots are focused on the amount of fuel they will carry with them, when their engines can be turned on and off, and whether the plane can idle on its descent.  Similarly, operations controls have insight into delays and cancellations, and air traffic control considerations. 

Overcoming resistance 

As is common with any change management, there was inevitably resistance to overcome.  This resistance could revolve around standard practices in the business, or from employees who are reluctant to take on new tasks.  Similarly, in industries such as aviation, there is such a strong safety culture that this can inhibit change, especially where pilot unions become involved. 

With costs becoming such a factor, especially during COVID, fuel efficiency has become increasingly important.  Numerous best practices have emerged over time to ensure fuel efficiency at the various stages of a journey.  By using machine learning and big data, the aim was to reduce fuel usage by up to 5%, which would cut overall costs for airlines of up to 1.25%. 

Ukraine International Airlines began their journey by creating a series of fuel saving best practices, which were created via industry guidance and statistics from each journey.  They thought, however, that they were applying these best practices roughly twice as often as they actually were. 

Improving what you measure 

One of the use cases where machine learning was deployed was the arrival procedure at Dubai Airport.  Typically flights would arrive from the north west, which takes them to the south east before they swing around and land.  This is a convoluted route that is largely done to help air traffic control.  Short cuts are possible, but it usually takes an experienced pilot to spot when this is feasible. 

The machine learning model worked to identify when such a short cut could be safely taken, and when doing so would produce the biggest savings in terms of time and fuel.  The use of machine learning was crucial because there was such a huge volume of data that doing it manually would have been impossible.  The model was updated on a daily basis so that pilots had the latest information in their flight briefing. 

The model is fed with data from the black box in the plane itself.  The device is recording hundreds of parameters per second, while also containing flight plans and airline operational data. 

Resistance from pilots is common, not least because they benefit so little from the savings that are made, and many regard the savings as irrelevant.  So it was vital to try and develop engagement with the pilots, and this was done by using behavioral economics approaches to highlight how much fuel was saved, and importantly how they performed alongside their peers. 

Key takeaways 

  • The project illustrates how high resistance can be, especially if you’re working in a mission critical context. 
  • Data can overcome this resistance through by helping to build engagement, even in a community as safety conscious as pilots. 

Three short-term actions to take: 

  1. Do you know where you are at the moment, and the root cause of any problems you encounter?
  2. Do you understand what the best practices are, and how this can help you to identify low hanging fruit for quick wins?
  3. Can you develop a proof of concept to help test solutions and secure alignment from stakeholders?

Three long-term actions to take: 

  1. Make sure that your actions are aligned with the corporate mission and vision.
  2. Always ensure that you think of people and process alongside IT.
  3. You can then invest in the data infrastructure to support this change.

Group discussion 

The delegates then joined their colleagues to discuss a couple of key questions related to Semi’s presentation. 

  • How would your organisation be able to use data as an enabler for driving change? 
  • What does a successful business and IT collaboration look like and how do you achieve it? 

A common theme among delegates was that a business sponsor was vital in using data to drive change.  This is vital not only in helping secure the buy-in for the change, but also to ensure that the change project is driving real business change. 

By defining the business problem that is being solved, this helps to ensure the project stays focused on the real end goal, while also helping to ensure stakeholders adopt the right roles and responsibilities.  As identified in previous sessions, the ability to translate between IT and business domains was also crucial in securing successful outcomes.

Data Leaders members can access the full case study and further discuss it with their peers here. Interested in becoming a member? Get in touch.

THE AUTHOR

Richard Pope

The global Peer-advisory community for chief data officers and their teams

Members Get More

If you value exclusive tailored content, then you’ll love our community

Learn about memberships