We traced the toll booth history of GDS distribution early this week. The short version was simple: NDC was airlines’ attempt to remove the toll booth, and instead multiplied the number of companies that might be running one.
That left a more important question unanswered.
A mid-size TMC in Lagos gets three pitches in the same month. One vendor claims direct NDC access to twelve airlines. Another calls itself an aggregator with thirty-plus carrier connections. A third is the buyer’s existing GDS, rebranding its NDC layer under a new product name.
All three say NDC. None of them are selling the same thing.
Most buyers only discover that in the middle of a booking, usually when something breaks and, occasionally, with a traveler already at the airport.
The ecosystem, stated plainly
It is tempting to line up airline, technology provider, aggregator and GDS as four separate categories of company. They are not necessarily four companies at all. They are functions, and a single company can occupy more than one of them at once.
The layers still run in a rough order: the airline as content owner, a technology or infrastructure provider building and operating the NDC connection, an aggregation layer normalizing multiple airlines into one interface, and the seller at the end of it. But the question worth asking is never what a company calls itself. It is what function it is actually performing between the airline and the buyer, because the label on the pitch deck and the function underneath it are not always the same thing.
At the top, content originates with the airline. Fares, availability, ancillaries and servicing rules are ultimately controlled by the airline’s own commercial decisions, whatever technology happens to sit underneath them. Some airlines build and run that technology in-house. Others contract it to a PSS or NDC technology vendor, while keeping commercial control of pricing and content themselves.
Below that, aggregators and NDC-enabled GDS platforms are often performing a similar normalization function, even though they arrive at it from different histories. Neither one originates airline content. Both connect to multiple airlines’ NDC APIs and hand the seller something closer to a single interface than twenty separate integrations. AirGateway, Duffel, TPConnects and Verteil built businesses on solving exactly that problem. Amadeus, Sabre and Travelport are increasingly solving a version of the same problem, delivered through a login screen sellers already know. The commercial models behind each are not identical, and buyers should not assume one aggregator’s fee structure looks like another’s just because the function looks similar from the outside. But the function itself, absorbing the differences between airline implementations so the seller does not have to, is doing comparable work regardless of which logo is on the interface.
Be careful with “direct”
This is where the market gets genuinely misleading, and it is worth being precise about it.
“Direct” gets used to describe at least three different things, and they are not interchangeable. Direct content means the fares and availability originate from the airline itself, rather than being resold or repackaged by a third party. Direct technical connection means the API call goes to the airline’s own NDC endpoint or the technology provider operating it on the airline’s behalf. Direct commercial relationship means the buyer’s contract, pricing and incentives are with the airline itself, not with an intermediary sitting between them.
A vendor can be technically accurate calling its connection “direct” while none of the buyer’s actual commercial terms touch the airline at all. That is not necessarily deception. It is a marketing term doing more work than the underlying relationship supports, and a buyer who does not ask which of the three meanings is actually in play has no way to know what they signed.
Why it is not a technicality
A buyer can sign an NDC contract, use it for a year, and still not know what they actually bought.
Take a TMC that gets access to twenty airlines through an aggregator. On paper, that reads as twenty airline connections. In practice, it is twenty different commercial and operational relationships wearing the same interface. One airline may support full self-service, another may restrict changes to voice-only servicing. One may expose every ancillary in its NDC offer, another may hold certain products back for its own direct channels. One may require a specific payment method the aggregator handles smoothly, another may not. A buyer who thinks they purchased twenty uniform airline connections actually purchased twenty different risk profiles, bundled into one invoice.
This is not a hypothetical risk. Alessandro Ciancimino, who leads NDC strategy at Sabre, made the same point from the operator’s side of the table in a TDN interview earlier this year. “Running a handful of airline connections in isolation is very different from supporting dozens of carriers across regions, schemas, and servicing models, under peak volume and commercial pressure,” he said. “Scale exposes the real challenges of NDC.” That is exactly the gap between what a buyer thinks they are signing and what they are actually operating once volume hits.
That is what changes depending on which layer a buyer is actually dealing with, and depending on the specific commercial model in front of them, since terms vary enough across this market that no two contracts should be assumed identical.
Negotiating leverage moves first. Dealing directly with an airline, in the commercial sense, means negotiating that airline’s own terms. Dealing through an aggregator or an NDC-enabled GDS often means negotiating a relationship with a party whose own terms with the airline are not the buyer’s to see. That is not automatically a worse deal. It is frequently an unpriced unknown, and buyers rarely account for what they cannot see.
Accountability moves second. When a booking fails, or an ancillary cannot be fulfilled, or the price at checkout does not match the offer, someone owns that failure, and it is not always obvious in advance whether that is the airline’s own implementation, the aggregation layer, or the payment layer underneath either of them. A support ticket sent to the wrong party turns a technical problem into a customer-service crisis, usually with a traveler on the other end of it.
Cost moves third. An aggregator or GDS fee is frequently layered on top of the airline’s own commercial terms rather than replacing them, and comparing that fee against an imagined “direct” cost without confirming how the two actually stack is not a real comparison. The more useful question is rarely how much NDC costs. It is what the total cost of distribution looks like from shopping through booking, payment, ticketing and servicing.
Aggregation is not the problem
None of this is an argument against aggregators or GDS platforms.
For most mid-size and smaller sellers, direct integration with dozens of airlines individually is not realistic. Maintaining separate implementations, tracking version changes, and supporting servicing flows across twenty airlines is a permanent engineering job, and the aggregation layer exists precisely to absorb that complexity so the buyer does not have to.
Jerrin Jos, Founder and CEO of Verteil Technologies, made a version of this same point in a TDN interview earlier this year, describing how the function of intermediaries shifts under retailing rather than disappearing. “Retailing is not about removing intermediaries, it is about modernizing their role,” he said, arguing that the value has moved from simply providing access toward enabling scale and integration across a fragmented ecosystem. That is a useful frame for buyers too. The aggregation layer is not a leftover from the old GDS era. It is doing a specific, ongoing job, and the buyer’s task is to understand that job clearly enough to price it.
The problem is not that this layer exists. The problem is a buyer who does not know it is there, and signs a contract believing they bought something other than what they actually bought.
What buyers should actually ask
Where does the content originate? Is it coming directly from the airline, through its technology provider, or through an aggregation layer sitting above both?
What does “direct” actually mean here? Is the vendor describing the technical connection, the commercial relationship, or both, and are they willing to say which?
Who owns servicing? If a ticket needs to be changed, refunded or reissued, who actually has the authority and the capability to do it? Jorge Díaz, CEO of AirGateway, put the industry-wide version of this problem plainly in a TDN interview earlier this year: “While many can shop and book, few can seamlessly rebook or refund without manual intervention.” That gap sits squarely at the aggregation layer, and a buyer who has not asked how their specific vendor handles it is assuming an answer rather than confirming one.
Who owns the failure? When something breaks, which party is contractually responsible for resolving it, and within what timeframe?
How does everyone make money? Are the fees, markups, incentives and transaction charges in this chain actually visible to the buyer, or only partially disclosed?
What is the total cost? Not the headline access fee. The cost from shopping through booking, payment, ticketing and servicing, all together.
How difficult is it to leave? Can the buyer move to another provider in two years without rebuilding their distribution architecture, or has the layer they signed with quietly become a dependency that is expensive to exit?
That last question introduces something worth sitting with without needing to sound alarmist about it. Every layer between a buyer and an airline’s content, whatever it calls itself, exists for a reason, and that reason is usually legitimate. The question was never whether the charge is justified. It is whether the label on the sales presentation actually describes what is underneath it.
Know where the layer sits. Know what you’re paying for. Know who owns the failure.
