Pricing & Revenue Management Blog

Is Data Science the future for airline revenue management?

May 2, 2016 9:52:52 AM
| |

Is Data Science the future for airline revenue management?

Artificial intelligence and data science isn’t new in the airline industry. However, its development has only recently begun to make a leap forward for two reasons:

  • Machines started learning and filling gaps for themselves based on their experience, improving precision from what used to be exclusively rational critical data analysis and interpretation.
  • Airlines began looking beyond historical trends analysis and envisioned projective data and behavioral analytics.


In October 2015, the University of Oxford opened a new Data Science Lab in collaboration with Emirates, to use cutting-edge analysis to help the airline “make its services more efficient and customer-focused”. From the University, a select group of mathematicians, scientists, engineers and social scientists joined carrier’s experts in the industry ad data, to research how to apply optimization techniques to Emirates' data, and develop machine learning to interrogate complex data sets.

The industry has been behind in using these pools of big data analysis tools, despite the latest trends in sectors such as the retail business, who has been able to personalize their value proposal with demand segmentation in e-commerce channels.  Some airlines have taken this trend a step forward, as they realized that, since data science expertise was transversal to all units of activity, it needed to be at the top of the management funnel rather than just at a service-level position.

In 2015, EasyJet appointed his first head of data science, Alberto Rey-Villaverde, to accelerate the use of artificial intelligence (AI) to improve efficiency, reduce cost and increase revenue and customer satisfaction. His profile sums up greatly how has this discipline also risen up the ranks in the airline industry.

An MSc in Data Mining from the University of Westminster – a program  that will later be upgraded to Business Intelligence and Analytics-, Mr. Rey-Villaverde began his careerin 2006 at EasyJet as a Yield Developer, and was subsequently promoted to Network Pricing Manager and Yield Strategy and Development Manager, before becoming head of data science.

“Technology and data is not something that falls either on the customer’s side or on the business side. We see it as something integrated that goes across the whole business”, he stated at the BTiQ 2015 conference. “The airline seat is a perishable asset. Once the plane departs, if we have a flight taking off from Luton to Malaga at 6 PM, if my team didn’t manage to sell that seat, the seat is lost. All the effort that was put in place to bring it to market is wasted”. “What we do is look at the past performance, we learn a lot. After 20 years flying we earn a lot of knowledge on how the customers behave, what are the patterns, how much people is willing to pay at different times of day.”, he also said. “We need a lot of data. Not only historical data but that of what is going on, on a day-to-day level. Those 95% of price adjustments that we make every day, they’re coming from what is happening every time every minute.” 

 

 
The data scientist speaks about the applications of this for airline business intelligence and yield management: “Data supports changes in the current market conditions, and we started to use machine learning as traders weren’t making smart decisions and the algorithms weren’t picking up everything,” he revealed. “To date, analytics has been about diagnostic capability and looking backward. Now advanced AI is more focused on predictive capability so we can better understand the future and plan for it.” 

 

EasyJet has applied data science to revenue management, setting fares by responding to passenger demand, helping increase revenue per seat in nearly 20% between 2010 and 2014, according to the Financial Times.  According to the newspaper, EasyJet plans to use AI to predict demand and analyze over 1.3bn searches made on its website each year, to optimize destinations and flight times. 

Despite these trends, Legacy airlines are still some steps behind, and need further development, as their yield management is away from EasyJet's simpler logistical and revenue structure: its entire booking procedure is online, their fleet has only one type of plane and only operates short haul routes. “This combination of simplicity and scale enerates an enormous quantity of data,” Mr. Rey- Villaverde admits at EasyJet’s Spanish blog. He also addressed a growing internal fear at the BTiQ 2015 conference: machines are not replacing humans. “We are using machine learning on our team. We did not displace anybody. We didn’t have to reduce the number of positions that we have in there. On the contrary, we have more people that are capable now, just because we are using machine learning.  "Machine learning is used to automatize what we are doing.” He points out: “I think we are far away from machines that think for themselves, but they are more intelligent. Machines pick up the data, learn from it, spot patterns and get a program from that.” 

What will happen to Data Science, Predictive Analytics and Artificial Intelligence for medioum to long-haul flights and legacy sales? The escalation of AI in the entire industry is an upcoming challenge, one that benefits all. Rey Villaverde will discuss these topics as a Keynote address “How to apply data science to revenue management to maximize sales through supply, demand and competitor pricing” at the Aviation IT Show Americas, part of the Aviation Festival Americas, held on May 24th-25th 2016 in Miami. He’ll discuss the identifying of prescriptive, predictive, diagnostic, and descriptive analytic capabilities, the evolving perception of data sharing and automation, as well as talk about segmenting the market and daily price adjustments through historical and real-time data, among other issues. Airnguru will be present at the event to engage with industry representatives and continue to develop airline pricing intelligence. 

How far is your use of data science in your company? How is it looking to view predictive analytics? We look forward to discussing your own concerns with industry representatives, so we are open to your comments below. 

Request a walk through

Tags: Revenue Management
Related Articles

Airline Profits: The High Price of Doing Nothing – Infographic

Jun 28, 2017 8:00:00 AM | Sergio Mendoza, PhD
Airline Profits: The High Price of Doing Nothing – Infographic

Airlines can add up to 5% additional net revenues to the bottom line if they improve their pricing processes, access real-time data to calculate fares, and rely on cloud-based technology to uncover hidden revenue opportunities, according to "The High Cost of Doing Nothing," our recently published infographic.

Read More

Revenue Management & Pricing Experts: is AI a Threat or an Opportunity?

Aug 19, 2016 3:00:08 PM | Daniel I. Pizarro
Revenue Management & Pricing Experts: is AI a Threat or an Opportunity?

Will Artificial Intelligence rival airline revenue management and pricing analysts?

It is a fact, and surely you heard it is happening. The time when machines learn has come. In the airline industry, machines can learn about fare gaps from themselves, and airlines are looking beyond historical trends for yield management. They are running predictive analytics and interpretations. But will they replace or empower revenue management and pricing analysts?

Read More