A Grid in the Sky, part III - Hackers and Pilots
Fragmentation is coming for the airline industry. OTAs must lean in to it.
This is the last post of an ongoing series called A Grid in the Sky, which looks closely at an airline industry that is becoming more fragmented as a result of technology. You may go back to part I and II in whatever order you find suitable, as they have been written as standalone pieces. However, if you haven’t read the previous posts, I encourage you to do so, and to subscribe if you want to learn more about payments in the context of the travel industry.
Have you ever noticed in the movies, that the “hacker” never leaves the keyboard when he (it’s usually a he) sits down at the computer? Before the invention of the mouse, the only way you could interact with a computer was with what’s called a Command Line Interface (CLI for short), that accepted a prearranged number of keywords and options to execute tasks. Over time, and as the mouse became more prevalent, an alternative form of interaction with the computer, with arrows, buttons, and icons, took over most consumer software. That was the Graphical User Interface, or GUI.
Despite the Internet boom that started during the 90s, the most powerful enterprise software, even today, is CLI-driven. This is because, unlike consumer software, where monetization is often indirect (via ads), enterprise software is billed directly to the companies that use it. Companies are way more demanding in terms of what can be done with software, and have the negotiation power to make it happen. As a consequence, it is strategically sound to give your customers the ability to do the job themselves, with the flexibility of a CLI, rather than provide them with a very complex, fully featured, cockpit-like GUI to lose themselves in.
Some companies, like Salesforce, have figured out a way to create dynamic and powerful GUIs that allow their customers to create their own workflows, much like CLIs, by making them more interactive. Others, like Retool, provide a mixture of the two. And there’s also Amazon Web Services, which has the capacity to serve both a flexible CLI and a fully-featured GUI.
How come CLIs haven’t disappeared yet? Because a good command-line is a very expressive way of communicating with a software system. They are a bet on text, its flexibility, stability and efficiency. GUIs make easy tasks easy, but CLIs make difficult tasks possible.
Your agent is looking at a CLI right now
Over the last weeks, I’ve been looking at a particular problem in the airline industry from two different angles without explicitly referring to it. That problem is the entrenched system in which airlines, as suppliers of flight seats, meet their customers’ demand.
In between these two groups sits a concert of companies (GDSs, like Amadeus and Sabre; the pricing publishing company ATPCo; and scheduling information providers like OAG and Innovata) that were, in the moment they were born, an astounding revolution. They leveraged software and achieved success in much the same way that Google disrupted the way people accessed the Internet, by coping with a dramatic explosion of demand with computers.
Today, though, they face a difficult conundrum. Reservation agents worldwide look nowadays at a very similar command-line interface that they’ve been using since the inception of these companies. The creativity and expertise that these systems require make humans at the counter the single most important point of failure during reservation. As a result, passengers that go to an agent at an airport are often left stranded unnecessarily, that is, despite options being available to them, only because the complexity of the tool makes it impossible for the agent to seize them.
Cracks in the Ceiling
Ryanair has spearheaded the trend that this complex web of slow moving monopolies seems to be leading us toward: fragmentation. In a curious case of a reverse Tragedy of the Commons, airlines are encouraging customers to buy tickets in their own websites, pushing back against the centralized view that GDSs used to provide. Fragmentation is also spurred by direct to consumer methods such as loyalty programs, which rely on the sunk cost fallacy of their users to buy directly from them before looking at other options.
On top of that, there’s the problem with data. One of the most prevalent narratives subsumed within AI, LLM, and chatGPT is that “data is the new oil”. However, data is not One Coherent Thing1, but innumerable different collections of information, highly specific and contextual. As a result, there is a qualitative aspect to it that cannot be captured in “how much data there is”. Data is only valuable to the extent that intersects with the rest, as a result of a consistent means of acquiring it.
The fragmentation of the airline industry isn’t making data less available; but the fact that needs to be captured from different sources makes it less likely to produce knowledge.
The end result is probably bad for the consumers: as the view of the market becomes more fuzzy, finding the optional route becomes more difficult, or at least requires more ingenuity from the customer, who over time develops a sense of distrust for the whole industry. “Are they tracking me while I’m purchasing tickets only to push the prices higher?”, “Do I get cheaper prices if I use a VPN?”.
Virtual Interlining
As the airline industry is bound to fragmentation, where do online travel agencies sit? They are up against a very difficult challenge: not only is the information more difficult to capture, but the very airlines that supply that information actively fight against any form of data collection. The idea of airlines turning into platforms is certainly not happening (which could have led to some form of accessible CLIs), and they’re using all the tools at their disposal to prevent any form of web scraping, programs that simulate users of their GUIs.
Online travel agencies, then, will have to adapt to this situation. We are probably seeing some of this already happening:
OTAs must turn big: The only way to combat the efforts to prevent web scraping is by one-up web scraping efforts. In a form of mutually-absurd-development, only more technologically capable companies can confront shifting GUIs and extract data from them successfully. For that to happen, a considerable amount of time and resources must be put into it, which forces OTAs to grow in order spread the costs throughout more customers.
OTAs must turn into Merchants-Of-Record: The fragmentation of the industry can only be fought when the incentives of the consumer and the intermediary are aligned. This, I believe, cannot happen when OTAs act as a convenient happenstance provider that finds the link that the customer has to click, and moves on. Catering to the customer experience and feeling their pain can provide the necessary incentives for the company to address the real problem, and to do that successfully.
OTAs must collaborate with airlines: To the extent that fighting the whole industry is a losing proposition, OTAs must find common ground with their suppliers. Restricted API access, rebates, affiliate programs, everything seems to be on the table if data can be accessed more reliably and cheaper.
Embedded into these ideas is payment acceptance. Reducing transaction costs at scale, owning the payment relationship and increasing the flexibility of payment options can spur this situation in the right direction for OTAs. It is by providing this powerful mechanism by which end consumers can become like the hackers in the movies, ingenious individuals that find the optimal path in seconds, all by the power of modern technology.
I’ve used this concept so much I should probably trademark it.