Travel agents are beginning to notice signs of algorithmic pricing across direct booking channels. While personalised, dynamic fares were promoted as a way to deliver better prices through NDC, the rapid evolution of self-learning pricing systems is raising new questions about transparency, competition and who ultimately controls the customer relationship.
When NDC entered the market, airlines promised it would offer cheaper, personalised fares. Today that promise is showing cracks, said Rachael Penaluna, MD of Sure Maritime Travel.
“Airlines told agents that NDC was gospel, that NDC was always going to be the cheapest and the best way to do things. But due to more advanced algorithms, we are seeing that sometimes the fare on the GDS may be more competitive,” said Penaluna.
She noted that algorithms may also be tracking search behaviour on direct channels and potentially hike fares when repeated searches indicate strong buying intent.
“We should be entitled to search the price of a fare, as agents and consumers, as many times as we want, without algorithms hiking prices to take advantage of our intent.”
Collusion by code
In a recent article, The Risks of Algorithmic Collusion, IATA warned that increasingly sophisticated pricing algorithms could create competition risks if they begin adjusting prices in response to market conditions – or even each other – in ways that disadvantage consumers.
The concern extends beyond traditional yield management. As airlines gather more customer data, personalised pricing based on individual purchasing behaviour is also attracting scrutiny from regulators in the US, UK and Europe.
The local picture
It appears that South African airlines are not yet using AI-driven algorithms nor advanced CRM profiling data.
In an interview with Travel News, FlySafair’s Chief Marketing Officer, Kirby Gordon, said the bigger question may not be whether airlines adopt these technologies first, but whether other travel retailers do.
In the travel industry, some of the most advanced mechanisms for managing pricing, inventory and market dynamics across channels are currently being used by OTAs, pointed out Gordon, and the most advanced customer-profiling-based pricing mechanisms are developed by financial institutions that have also started reselling travel products.
“The concern with personalised pricing is that information about an individual could be used to estimate that person’s willingness to pay and extract the maximum possible value from the transaction,” said Gordon. “For an airline, developing that level of understanding of the consumer is actually quite an intimidating prospect. An airline may know something about its customers and their past interactions, particularly when it operates a loyalty programme, but that knowledge can still be comparatively limited and frankly clunky.”
Gordon believes banks, insurers and loyalty businesses may ultimately be better positioned than airlines to use advanced customer data.
“Their knowledge of customers, and their ability to construct individualised propositions, and the information that they have about us, may ultimately exceed that of the airlines whose products they distribute,” said Gordon.
With this in mind, Gordon posed the question: could data-rich financial and loyalty businesses eventually become more influential as retailers of travel because they understand individual customers better than agents or suppliers do?
Gordon speculated that as advanced customer-profiling and self-learning algorithms come into play in the South African travel industry, the role of airlines and suppliers may be only to manufacture and fulfil the travel product, while other intermediaries, who adopt this technology faster, take ownership of the customer relationship and the offers.
“I do not know the answer to this question, but it is an intriguing possibility,” added Gordon. “If the story of air travel had once been a move away from the agent to the airline direct, is this then the start of it moving away from the airline again to a new type of agent?”
Learning curve for regulators
Most of the current conversation about algorithmic pricing is about airlines using AI to extract more value from consumers, or on pricing systems learning to coordinate in ways that could reduce competition, said Gordon, so understandably, the emphasis is on laws and safeguards that protect consumers.
“My instinct is that the spirit of South Africa’s existing law may be capable of establishing the right protective parameters,” he said,” but the more immediate challenge is equipping regulators to identify, audit and prove anti-competitive behaviour when it is occurring through extraordinarily fast and complex systems.”
“One advantage for South Africa and our travel industry is that we will have many international lessons from which to learn,” said Gordon.
“However, we should be cautious about assuming that this automatically protects us. Nevertheless, financial, securities, commodities and other trading markets are likely to encounter many of the relevant technologies, strategies and risks before airlines and travel companies do. Their regulators are therefore also likely to gain experience with algorithmic supervision, market manipulation, auditability and system controls ahead of us. I suspect (nay hope) that aviation will be able to learn from that experience.”
