Facebook Demand Forecast

Traditional data for airline forecasts is outdated – replacements are ready

At the heart of an airline’s revenue generating processes lies it’s demand forecast. By figuring out when and where people are going to book, and aligning that with expected willingness to pay, they can ensure that seats are not sold for a low price when a higher one could be achieved. Until recently, progress in the field relied on improving accuracy with statistical wonkery and incorporating services for which people pay extra, such as baggage and catering, into a passenger’s expected revenue value.

Following COVID much of this has flown out of the window. The forecasts previously assumed that history was a good guide to the future, so if a certain flight was busy every Friday it would be busy next Friday too. Today it is not clear whether or not certain borders will be even be open next Friday, so airlines need access to data that is both real time and forward looking, not historical. Three technologies – credit cards, social media and mobile phones – are set to offer the airlines what they need.

Plastic places

Airlines and credit cards have a long history. American Airlines and Air Transport Association created Universal Air Travel Plan (UATP), arguably the first credit card, in 1934 so passengers could buy a ticket on credit. To make things even better, 17 airlines offered 15% discounts for members.

In those days flights were expensive and passengers well-heeled so buying tickets on credit seemed natural, and it has been ever since. UATP’s pioneers would no doubt be pleased to see continued collaboration between credit cards and airlines.

In the second generation of airline-card collaboration travel concierges, airport lounges and insurance were bundled with pricey platinum cards favoured by business travellers. More recently, co-branded cards offering air miles and the promise of free flights for members have helped make plush business class seats a mass affluent consumer experience.

But it is now time for cards and carriers to enter the fourth stage of their relationship, where data about broad consumer trends is associated with how likely people in a certain city are to purchase travel. For example Christian Hylander of Mastercard, a payments processor, told me that people who go out dining are much more likely to travel than people who buy services such as Internet dating.

Airlines and their card providers can look at the mix of products and services bought by consumers in a particular city and compare this with other cities all over the world. It is not just large carriers like Lufthansa, Qatar Airways and Singapore Airlines who use this data, smaller airlines like Aeroflot and Copa apply it too.

For example, if trends in a city are becoming similar to cities where lots of corporate travel originates, that is a signal that corporate demand is likely to increase and the airline should adjust it’s demand forecast and pricing strategy accordingly.

Book of faces

But Abheer Kolhatkar of Migacore, a big data startup, told me that airlines should be cautious and new data sources will not necessarily provide all the answers, although this may come in time. Instead they provide a signal to “push” existing demand forecasts up or “pull” them down.

Migacore captures data about what people are talking about across both traditional media and social media. The idea is that if a TV programme about elephants is broadcast in Germany, airlines can expect demand to countries such as Botswana and India for people to see elephants in the wild, which they can monetise by putting up their prices.

The problem is that the impact of such a programme cannot be seen from the TV schedules alone – it needs to be combined with real time data from the social media chatter generated to show how much impact it had. Migacore collects this data and provides the signal to airlines like Swiss and vueling.

Over the long term, changes in consumer trends and preferences can be detected that influence fleet and network planning, the process an airline takes to decide which cities to fly to and how many flights there should be.

Mobile chasers

A third opportunity for airlines to learn about how consumers travel before they start shopping is to consider mobile phones. Nawal Taneja of the University of Ohio pointed out to me that when he makes a call in Ohio one day and the next day receives a call in France the phone company knows that he flew, creating data that can be shared with airlines.

Amit Nagpal of Aetha, a consultancy, told me that while a mobile phone’s movements from one mast’s service to another is not recorded, the billing data certainly is. Such data could be used to help airlines understand how individual travellers move around.

Data governance applies

These three data sources may offer the chance for airlines to forecast demand at the individual passenger level and permit personalised pricing. But data governance regulations are unlikely to allow this in practice. My advice to airlines is to focus on sharing value with the consumer by offering attractive, personalised product choices rather than pure monetisation alone.