Does the future of consumer data collection lie in “hyper-personalization” or “anonymous personalization?”
How can companies make their competitors’ data work to their advantage?
Those were key questions discussed at ATPCO’s recent Elevate 2022 conference during the keynote panel, “Machine learning & AI: What the next 10 years of innovation will mean for the industry’s bottom line.”
Subscribe to our newsletter below
Confronted with data privacy laws and consumers’ own privacy concerns, airlines would be wise to adopt “anonymous personalization,” says Richard Ratliff, executive scientist and research fellow at Sabre Labs.
The approach classifies customers according to preferences - for example, those who respond well to certain ancillaries or are sensitive to price versus schedule, according to Ratliff.
“I think we can make really sophisticated systems that operate well at that anonymous level [and] respect data privacy,” he says.
Says Michael Wu, chief AI strategist at PROS: “Consumers, especially the millennials, they don't have any problem giving out the data if they know what value they are receiving.”
When data is hyper-personalized, it allows airlines to make tailored offers, according to Wu.
“I think there are ways that we can actually make these very personalized offers attractive … by making sure that value proposition’s upfront.”
Wu contends that hyper-personalization is also easier for the company because “there's really no need to do any aggregation. It’s much easier from a technology perspective.”
Sam Chamberlain, head of product, revenue management at FLYR Labs, says he hears from airlines that they don’t want that “granular level” of detail on customers.
Anonymous personalization is “the happy medium for airlines” that allows them to gather customer data on shopping and loyalty status, for example, he says.
Making use of models
Airlines can also benefit from creating a common data set across the industry, the panelists say.
During the pandemic, PROS pulled together dozens of airlines’ data to build a joint model. Airlines contribute their data and PROS feeds it through a model, and then the airline or the company has access to the output - not to the raw data.
All companies stand to gain from sharing data, even though the amount of gain may be different. PROS saw an average 8.2% improvement in forecast accuracy, according to Wu.
When it comes to comparing data across the airline industry, results need to be “interpretable and explainable,” Ratliff says.
I think whoever wins at this - whoever is successful at it - really needs to have a win-win business model that everyone’s going to benefit from.
Sam Chamberlain - FLYR Labs
“If it's well thought out, I genuinely believe that we can end up in a situation where it’s better for the airlines and better for the customer.
“Airlines are going to be experimenting with a lot of different things, and it's going to be really important to know whether those things are actually working or not. So having the competitor insights … is going to be really important.”
He cites Google’s Vertex AI as a useful platform that can automatically test and recalibrate models.
It’s incumbent on researchers to make sure that the results are “a glass box instead of a black box,” and it’s critical to avoid looking at one facet in isolation, for example, “focusing on revenue management without really taking into consideration pricing,” Ratliff says.
“If the models are more interpretable, it makes it easier for the users to understand all the different components and put that together to make sure everything is working in harmony.”
Chamberlain recently attended an aviation festival in which a presenter said: “Stop giving me data. I have enough data.”
It’s time for companies to start extracting “meaningful insights from the data that they've got and start agreeing across the industry: What are important metrics, what are important aggregations of that data? How are we going to use the data in our in a common way? And I think a lot of time and effort needs to be put into to that as well as collecting lots of different data sources.
“I think whoever wins at this - whoever is successful at it - really needs to have a win-win business model that everyone’s going to benefit from,” Chamberlain says. “And it has to provide real blanket market coverage; it has to be applicable to 90 plus percent of market.”
Data science, artificial intelligence and machine learning have formed the “foundation of everything we’re doing, because we wholeheartedly believe that’s what’s going to drive the future of the travel industry and pricing and revenue management in particular,” Chamberlain says.
“It’s become a bit of a cliche over the years to say that airlines want to be like Amazon of the world … but they are trying to become more tailored and more sophisticated merchandisers,” he says. “And I think we can look at how data science has worked in different industries and start applying learnings from that into our own.”
Machine learning will not replace analysts, the panelists say.
The analytical model paints 90% of the picture, and machine learning can predict necessary adjustments, according to Thomas Fiig, director, chief scientist at Amadeus.
“There's some randomness that is very hard to model and you have the machine doing that,” Fiig says.