Case Study: Big data improves operations and fuel performance at Cebu Pacific
Author: Francesc Torres, Director Operational Support, Cebu PacificSubscribe
In this article, I want to share our journey through data. There’s always a lot of material about EFBs, new systems, operations systems, maintenance systems but what is really important is how those systems are deployed in the operation, how are users able to analyze the data that is being collected. First, to set the scene, let me tell you a little about Cebu Pacific, who we are and what we do.
CEBU PACIFIC AIR
Cebu Pacific is a Low Cost Carrier (LCC) that started operation some 23 years ago in the Philippines. The Manilla based airline serves most of South East Asia, North Asia and the Philippines (figure 1).
At the time of writing, Cebu Pacific operated about 75 aircraft but there is an order in place for a further 68 aircraft and that’s important because we really have to be efficient and one way we see to ensure our efficiency going forward is analysis of data to see where we can be increasingly efficient as new aircraft are delivered and every time the operation grows.
DATA MANAGEMENT PRINCIPLES
For any business, setting out on the data journey, the most difficult thing that has to change is the data culture. It was very clear to us that we wanted to be agile, that we wanted to break down silos in the business and that we really needed a cultural change to create a culture that was data friendly.
In order to be agile, we had to empower people and let them think for themselves, as well as empowering departments by ensuring that they could access data and generate information quickly. The more access departments could have to data, the faster they could resolve the problems with which they were faced. To destroy the silos, we had first to democratize data so that everybody had access; in other words, data available to all. People can analyze the problems in their department together with the other departments: they don’t have to work in isolation or be in isolation so neither should data be available in isolation. Transparency is incompetence’s worst enemy and, in this, we really wanted to bring about a cultural change, for people to come together to discuss their experiences around data, around sharing information with other departments. This was how we could promote continuous improvement in how the tools are used and, with analysis and graphs, we could remove barriers so that people feel comfortable with change and the management of change. That was what we had in mind, where we wanted to go on this data journey.
We also created a timescale for the journey (figure 2).
In 2016, we started with our first experience with the fuel management system. We could see a lot of benefits, a lot of improvements with people starting to get used to dashboards and analytics and KPIs (Key Performance Indicators) and so we decided to move one step further on by doing Operations data and automating KPIs. Then, as increasing numbers of data sources came into play, we felt that the issue was becoming too big and that we needed to deploy on the Cloud, merge more and more data sources and ensure that we could have more visibility of all that information. Now, where we see this going is that, in the same way as other businesses see it, data is an asset which we need to start applying into the operations of the airline. We see ourselves now building a data ecosystem into which we’ll have to put strong governance. In the coming years, governance will be increasingly important, the way we work and the way we structure data and the way we adopt it across the business.
But, at the same time, when we talk about these great ideas and we talk about what we want to do at the execution level, we need to be very careful on how to make it happen. So, we have defined different levels of analytics (figure 3) and we know that we can move to the next level when we are ready; so complexity will be a big issue here.
First, we want to report what happened and then we need to empower people to be able to analyze why it happened and we need to continue the monitoring. And, of course, at the end of the path is the prediction, what might happen? We felt that a lot of times, people wanted to jump to the prediction side really quickly when there was no proper basis for that. So we’re very careful about how we move to the next step every time we undertake a project.
There is something similar for the hierarchy of needs (also figure 3). First, the data has to be available, then we need to be sure that it’s reliable; we need to clean it, detect any anomalies and transform it to make it available. We need to get the metrics out there and, the last step, to deep dive the analytics, analysis and machine learning. We do believe that we need to go step by step in order to make a change. If we jump straight to the end and the data is not reliable, then it will not be possible to progress with the whole cultural change.
However, at the same time as we were considering all that, the business’s management were asking for propositions for what value all this was bringing into the operation (figure 4) in terms of what we do; the business performance, streamlining the operation and making the business more efficient.
In the figure is a diagram of the main projects tying those projects to the requirements of improved performance and streamlined operations. Management felt that it all was great, we were moving forward, but what would it bring to the operation; how would it add value.
We had to look back to ‘how do we make this happen?’ There were metrics at the execution level, we knew how we wanted to do it, we knew the route and where we wanted to reach but our experience has been that we need to enhance skills and capabilities. We needed people to be trained and to tackle a learning curve; we needed to have monthly sessions between departments to break down silos; we needed to improve the data culture in the organization. To not be afraid of making decisions based on data and on analytics. At the same time, we see those roles in different departments in the commercial arena in terms of data and to be sharing their experiences: preferably, this push should be coming from the departments. We don’t really believe in having a Chief Data Officer role where all the data analytics comes from one department. What we see is that each department should take the data, analyze it and use that analysis to improve what they do. The data has to be close to the teams so that they can use it and analyze it.
The question has to be, ‘how did we reach here?’ What was the trigger?
SPECIFIC CASE FOR FUEL MANAGEMENT – SKYBREATHE
In 2016, Cebu Pacific adopted SkyBreathe from OpenAirlines and we thought it was fantastic… the available data, the visibility for us; we’d never had that much visibility before. Then, as we started looking at the data more and more, we realized that this was just a small part of the picture, there is more that can be done, more data sources that could be merged with fuel. But how were we going to do it? Slowly we brought all this into the picture and now we had a bigger picture of what is available and possible plus, when we’d completed that, we realized that we were still just looking at a small part of the picture. It was just fuel and a small part of the Operations whereas we have much more data out there. There’s HR data, finance data, maintenance systems… there’s a lot of data available and if we only look at a small part of it, we won’t be able to get the whole picture; won’t be able to see what is really going on.
So, we sat down and analyzed a path forward plus, obviously, there were learning points; things we could have done better or wished we could have understood better from the outset. One of those was the importance of having clean data. When we give data to the analyst that data has to be clean; we underestimated the complexity of data anthologies but working with OpenAirlines we learned how complex those anthologies can become, it’s not easy. Therefore, we needed specific expertise to manage all that. We also realized the importance of a data ecosystem. It’s not only the fuel department that analyzes fuel but it’s all the things that go on around the department that the fuel department might need to calculate; whether the savings are there, whether more savings are possible, whether there are more opportunities. So we really needed to look at the bigger picture of these complex and multiple systems, running and working together in the operation. And a further important point is the need to empower staff: to train an analyst on the tools, keep giving them training on how important this all is. The OpenAirlines system helps the end user to understand using the visualizations but with training they can better absorb the information. However, we don’t need specialized people to put that visualization together.
ENGAGEMENT OF SENIOR MANAGEMENT AND STAFF
How do we engage people? We can put everything together, we can have great analysts, great data engineers, we can have all the data sources in the company put together but the key is to engage the people in the business to go there, to use the data, analyze it to improve what they’re doing in their daily work.
For users the system deployed has to be easy to understand, it has to be accurate and complete, and easy to access. Everybody nowadays has a tablet, has a cellphone so why can’t the data be made available there? It was very important for Cebu Pacific to be able to reach where the user is, i.e. not expect the user to come to us but, instead, we would go to them: the pilot fuel App or a simple App on the staff member’s cellphone or the tablet to make it more accessible.
The way analysts do analytics has to be intuitive and it has to be easy to create content. Not everybody was born being a data analyst or educated to be able to program with Python so we needed to make it easy and accessible for most people, if we need to teach people how to update python we will provide the necessary training. The more people that are doing analytics, the bigger the data absorption in the system will be and the cultural change will be. Plus, of course, continuous training and continuous support should be provided from the organization.
And, if we look at the management, they just want to see a quick visualization of performance and they also want to see that on their cellphones and on their tablets while they’re having a coffee in the morning. We also needed to deliver information to where the manager wants to see the information and we needed to consider which kind of information. Most importantly, it simply has to work. Every time the COO opens the App, he wants to see the financial performance: it’s got to be there. It also has to be reliable information that they can trust; a single source of truth for all the data they see.
That’s what we learned through working with OpenAirlines and SkyBreathe and all these points are very important for fuel management but they have to be kept important as the business grows into a bigger data set and a bigger data culture type of management.
The whole company-wide big data project took Cebu Pacific a couple of years to complete: we had to learn lessons, identify and tackle pain points and there were some important things that were highlighted in the process that we need to take with us into the future.
Governance is very important, especially at the higher levels. Management and ‘C’ level executives want to see a high level of governance; they do not wish to see chaos in the airline because everybody is doing things differently; they need peace of mind about the way things are being done. Obviously, data quality has to be assured – sometimes the expertise might be in-house, sometimes not. In the Case of Cebu Pacific, all fuel data goes to OpenAirlines who provide it to the airline which then displays it on dashboards. We still use the tool for analytics and we still have analysts but want to display clean data for management which is what we get from OpenAirlines who know how to clean the data and know the anthologies on fuel.
The organization itself has to make a change in a similar way to when we’re talking about EFB (where pilots made the change from paper-based to electronic-based) which, culturally, was a challenge; we also have to start looking at the challenge that our staff faced when they move into a data culture based on making decisions having analyzed problems. That’s also a challenge and, from our point of view, first experiences are very important in life, so if you select a good provider, you do your first experience and it works well, it’s much easier to go to the next level and keep growing from there. So, we recommend to choose very well the first provider to have a good first experience in data because that will drive how much your management will trust and how far you can reach in this cultural change.
Francesc is Director Operations Support in Cebu Pacific Air leading the digital transformation, the introduction of big data and AI in the operations arena, improving operational efficiencies, and merging market technologies and trends with Airlines’ Operations, including Fuel Optimization, Dispatch and oversight on Technical Operations Support. With experience in several operational environments, from LCC to flag carriers, and in different regions he holds an engineering degree in Aeronautics and a Bachelor Degree in business management.
Cebu Pacific Group is a Filipino airline group that operates subsidiary low cost-carriers; Cebu Pacific and Cebgo. CEB’s 72-strong fleet is comprised of 52 Airbus (one A321neo, seven A321ceo, 36 A320 and eight A330) and 20 ATR (eight ATR 72-500 and 12 ATR 72-600) aircraft, one of the most modern aircraft fleets in the world. Between 2019 and 2022, Cebu Pacific will take delivery of 31 more Airbus A321neos, five A320neos, and four ATR 72-600.
OpenAirlines is an international software company based in Toulouse, with offices in Hong Kong, and Miami. It provides consulting and software solutions for airlines flight operations. Since 2006, OpenAirlines has been on a mission to help airlines save 2-5% of their fuel consumption with its innovative eco-flying solution SkyBreathe. The software uses Big Data algorithms and automatically analyses the large amount of available data in flight data recorders to assess flights’ efficiency.
Today, 40 airlines all over the world use OpenAirlines’ software.
In 2019, their customers saved more than 150 million USD and 590,000 tons of CO2.