Comment: Why big data is key to airline recovery

Months later the predicted V‐ or U‐shaped recovery has not materialised, and it looks increasingly likely that airlines can expect this downturn to continue for a lot longer yet. Even the most optimistic analysts at this point are predicting the demand curve to look like a recumbent L at best.

IATA expects that as a result of the pandemic up to 25 million people, linked either directly or indirectly to the aviation industry, are at risk of losing their jobs with a contraction of about 30 per cent likely in the short term. Longer-term forecasts suggest that global passenger traffic will not return to 2019 levels until around 2024. But for those who know what to look for, there are signs that the industry’s lowest point may already be behind us.

There are two key factors driving the current decline in air travel. The first is the closure of borders and restrictions placed by governments that aim to curtail the international spread of the virus. The SARS epidemic of 2002 and the Ebola outbreak in 2013‐2014 were important case studies in just how quickly and easily pathogens can be spread by the greater connectivity air travel brings, and countries that have responded rapidly to this pandemic have no doubt limited transmission to some degree.

The second is the reduced demand for air travel among consumers, driven by both short‐term economic uncertainty and concerns about the safety of travel generally and air travel specifically, considering the highly transmissible nature of the virus. The good news there is that there are very few instances to date of suspected in‐flight spread of the virus, but without a compelling reason to get on an aeroplane people are choosing to remain grounded.

As a result, the aviation industry is expected to lose around $84 billion this year alone, based on IATA’s latest forecast, and while revenues should improve in 2021, they are still expected to be well below 2019 levels. A number of carriers have already closed or opted for bankruptcy protection, while the rest are desperately trying to limit losses by putting aircraft in long‐term storage and furloughing staff to reduce fixed costs.

But while this is a sound strategy at present, it could cause a big problem as and when a recovery begins to take shape.

VIEW ALL

Airlines are notoriously slow at reacting to changing market conditions, with their network planning departments usually relying on historical data to forecast future demand. But the cost of being reactive rather than proactive can be high in the face of a rapidly changing market, as redeploying aircraft and staff is expensive and missing out on new opportunities can be even more so.

As the inevitable (yes, inevitable) recovery from the pandemic gains altitude, airlines need to be ready to capitalise on the increasing demand, and decisions and strategies based on empirical data are likely to have the best outcomes. And this is where Big Data and science come in. With social media now a ubiquitous part of our lives, there is a wealth of data available online that can be used by analysts to determine all types of human behaviours and preferences. By aggregating the data and looking at the trends that emerge, airline network planning departments will be able to quickly respond to any uptick in demand before it happens.

This was the idea behind a recently published paper by a pair of researchers in Europe. Using supply and demand data for airline seats for the remainder of 2020, they were able to predict what a recovery might look like on a global level. The results come with some caveats, but an understanding of the methodology they used would be valuable to many airline network planning departments around the world.

When the researchers looked at the supply of airline capacity, the number of seats available based on published airline schedules, they found a dramatic increase in the second half of the year. Across the globe, the data indicated that airlines were hopeful of soon being able to fill seats at levels similar to those seen in 2019, but this statistic alone can be a bit misleading.

An increase in supply months down the track does not always materialise as services are cancelled at the last minute, and the rules regarding borders and the allocation of slots are in flux right now. Changes to schedules are inevitable. Now, well into the third quarter of the year, the anticipated surge in supply has not materialised to the extent that the data was suggesting a few months ago.

A more reliable indicator of future travel patterns is demand, and the researchers looked at two key metrics: the desire to travel, based on internet searches, and the intention to travel, based on the selection of a particular flight. After the rapid drop in demand earlier in the year, there were signs of slow growth from about July through to the end of 2020, with a few regions potentially returning to 2019 levels around December. There was clearly a loss in desire to travel in the short term, but this was less notable in longer-term forecasts.

Zooming in on more localised data revealed some interesting differences within regions. For example, air travel in northern Europe is anticipated to recover more strongly than in southern Europe, as is the Middle East compared with south-east Asia, while demand in North and South America is likely to increase at roughly the same rate. But in all cases, the L‐shaped recoveries are only slowly trending upwards, and that suggests that the overall economic impact will be devastating for many who work in or rely on the aviation industry.

The researchers point out that the data they used to arrive at these conclusions was not an estimate or an extrapolation but was real. If the global situation changes significantly, such as might be the case if there was to be a second wave of the pandemic, re‐evaluating the data would immediately produce a different result without the need to go back and revise old theoretical models. Similarly, if an effective vaccine becomes available, analysts would expect the data to immediately show a rapid increase in demand.

This is the point that airline network planning departments need to take on board. If you need the most accurate and up‐to‐date information in a rapidly changing situation, using big data is the best way to get it. And, for airlines, confidently being able to predict the number of crew and aircraft they will need weeks or months in advance will save unnecessary redundancies and reduce the costs of storing and reactivating aeroplanes and retraining pilots.

But just how many airlines will abandon their old methods and embrace this new way of thinking? That might not be so difficult to determine as we watch to see how successfully they navigate the biggest challenge the industry has ever faced.

Daniel Lambeth is an Australian Cathay Pacific pilot who has flown A330, A340 and A350 aircraft. He previously flew the 737NG for Virgin Blue.