NDC Flatlined: Modern, and Still Renting

Eric Marketts

Last week we asked who wrote the terms. Three chokepoints, and at every one the terms were unwritten and the airline had not written them. This week the story is quieter and, honestly, worse. Because this week the terms did get written. In three ordinary business deals. And in all three, the airline signed them rather than wrote them.

Steph Nell

And none of these is a scandal. That is what makes them matter. Last week was a Congressional letter and a class action. This week is a data series nobody reads, a crowded product category, and a product launch. The boring end of the news cycle. And the boring end is where ownership actually gets decided, because nobody is watching it.

Eric Marketts

So here is the claim for the next twenty minutes. Two of the three things we are going to talk about genuinely moved this year. The order management system became the thing every vendor sells. A machine started completing bookings end to end. The third one, NDC, has not moved in two years, and the industry has been quoting a number that does not exist to cover it.

Steph Nell

And at every one of the three, the airline got the modern version of the thing and did not get to own it. Where retailing moved, somebody else owns it. Where it stopped, it stopped exactly where owning it would have started costing the intermediary something.

Eric Marketts

Welcome back to The V1 Airline Retailing Report. I'm Eric Marketts.

Steph Nell

And I'm Steph Nell. For anyone new, here is what we do. Every Monday we take the developments that actually move airline and travel distribution, and we run each one through the bull case, the bear case, and the critical take. Not the press release. The strategic read. Sources are in the episode details.

Eric Marketts

Narrative one. NDC did not scale this year. It plateaued two years ago, and the number everyone quotes to say otherwise has no source.

Steph Nell

Narrative two. Every vendor in the market will now sell you an order management system. That is not a milestone. That is a warning.

Eric Marketts

Narrative three. A machine finally booked a trip end to end this year. It did it inside somebody else's policy.

Steph Nell

And the through-line is one sentence. Modernization is not the same as ownership, and this industry keeps buying the first and calling it the second.

Eric Marketts

Let's get into it.

Eric Marketts

Before the numbers, we have to remember what NDC was sold as, because the sales pitch and the outcome are pointing in opposite directions.

Steph Nell

Go back to the original promise. NDC, New Distribution Capability, was IATA's answer to a specific complaint. Airlines felt trapped inside the old EDIFACT pipes that the global distribution systems controlled. A fare was a fare, an ancillary was invisible, and the G-D-S sat in the middle taking a fee and owning the agency relationship. NDC was going to let the airline build a rich, branded offer, push it straight to the seller, and get out from under the intermediary. The word everyone used was direct.

Eric Marketts

Direct meaning around the G-D-S.

Steph Nell

Direct meaning around the G-D-S. That was the whole emotional appeal to an airline board. You spend the money, you build the offer A-P-I, and you stop paying a toll to a company that has been marking up your content for thirty years. Disintermediation. That is the word that sold a decade of NDC investment.

Eric Marketts

Now give me what actually happened.

Eric Marketts

Here is where I have to correct something, including a number this show has repeated. The figure you have seen everywhere this year is that NDC handles about twenty-four percent of indirect airline ticket sales globally, up from about eleven percent in 2023. Europe around thirty-one percent. North American corporate volumes up triple digits year over year. We went looking for the source of those numbers this week.

Steph Nell

There is not one. We traced all four figures back to a single travel-technology vendor's blog post, published with no attribution to any research firm, data provider or industry body. Not IATA. Not A-R-C. Not T2RL. And IATA's own Distribution with Offers and Orders fact sheet, updated June 2026, contains no adoption statistic at all. Not one percentage. It says adoption is each airline's own assessment and points at the Airline Retailing Maturity index. The most-quoted number in airline distribution this year is a number no institution publishes.

Eric Marketts

So what do we actually have.

Steph Nell

One series, and it is a good one. A-R-C, the Airlines Reporting Corporation, publishes NDC share of A-R-C-settled transactions every month inside its US agency air ticket sales release. US only, agency-settled only, transactions rather than revenue. Real caveats, real audit, published every month. Here is the series. NDC crossed twenty percent in mid-2024. June 2024, twenty-one point five percent. December 2025, twenty-one point two. March 2026, twenty point eight. April, twenty point one. May, twenty-one point six. June, twenty-one point six. July 2026, twenty-one point two.

Eric Marketts

That's flat.

Steph Nell

That is two years flat. The band is roughly nineteen point six to twenty-one point six and it has not left it. Year-over-year moves are two-tenths to nine-tenths of a point, which is inside the monthly noise. A-R-C's own language for May was "steady operational expansion," which is a generous way to describe a line that does not go anywhere. And it lines up with the other two audited numbers we have. A-R-C's channel split puts corporate at seven percent of NDC volume. And a Business Travel News survey of large corporate buyers found fourteen of twenty reporting under ten percent of their air bookings flowing through NDC.

Eric Marketts

And the deal flow?

Steph Nell

The deal flow is real, and it is the explanation. Royal Jordanian went live with its NDC content on Travelport at the end of May, delivering an enhanced distribution agreement the two signed in October of last year, a five-year deal. Icelandair signed a multi-year Travelport content agreement in March that explicitly includes its NDC content. Travelport is the smallest of the big three and the most aggressive on NDC content deals, precisely because it is the smallest. When you are behind on scale, breadth of content is how you stay relevant. The independent tracker most people watch lists around eighty airlines with a live NDC A-P-I, up from about seventy-five a year ago.

Eric Marketts

Bull case.

Steph Nell

The bull case is that a plateau above twenty percent is not a failure, it is a floor, and a floor is what maturity looks like. NDC went from nothing to a fifth of US agency transactions in under a decade, against forty years of installed EDIFACT. It is in production. It settles. It services. More than eleven hundred agencies transacted NDC in a single month this summer. Nobody should expect a distribution standard to compound forever, and share is a lagging measure of a standard still filling in its plumbing. The parts that are growing are the parts that make it usable end to end: payments, order servicing, direct connects. Singapore Airlines onboarded two corporate travel platforms to direct NDC connections in four days in August. That is the maturity curve, and it does not show up in a share number.

Eric Marketts

Bear case.

Steph Nell

The bear case is that this industry has spent two years describing a flat line as momentum, and it did it by quoting a global figure that no institution publishes. That is not a rounding problem. That is a governance problem. Airline boards approved nine-figure NDC programs on adoption curves, and a curve needs a number. The only audited number stopped moving in mid-2024, so the trade press reached for one that was still going up, and nobody asked where it came from. Including, for a while, us. And underneath it, the corporate channel, where the airline's highest-yield distribution actually sits, is seven percent of NDC volume, with most large buyers under ten percent of their own air spend. The place NDC was supposed to matter most is the place it has barely landed.

Eric Marketts

Critical take.

Steph Nell

The critical take is that the plateau is the disintermediation thesis reporting its own result, and this industry has refused to read it. Look at where the line flattened. NDC climbed fast while airlines were connecting the willing: the OTAs, the aggregators, the early direct connects. Then it reached the point where further growth required one of two things. Either airlines build direct connectivity to sellers at scale, which is expensive and slow and almost nobody has done broadly. Or the content routes through the aggregator, which is fast and cheap, and is exactly what happened. Royal Jordanian and Icelandair are not outliers. They are the pattern of the year. And here is the part that should stop a room. Routing NDC through the G-D-S changes the format of a booking. It does not change who is in the middle of it. So the share number stopped moving, because share of indirect was never measuring what airlines thought it was measuring. It was measuring format adoption, and the industry read it as power transfer. The G-D-S was never afraid of NDC. The G-D-S was afraid of airlines connecting directly to sellers at scale, and NDC, in practice, became the format that let the aggregator carry the airline's best content without the airline ever building that muscle. The intermediary funded its own relevance out of the airline's modernization budget. And then the industry found a growth statistic to avoid noticing.

Eric Marketts

Monday morning?

Steph Nell

Two things. First, find whatever NDC adoption number is sitting in your board deck and trace it to a primary source. If it is a global percentage, you will not find one, and you want to know that before somebody on your board does. Use A-R-C, state the caveats out loud, US only, agency-settled, transactions not revenue, and be the person in the room quoting the number that can actually be audited. Second, pull your own channel mix and separate two figures that usually get reported as one. How much of your NDC volume goes to a seller or aggregator you hold a direct commercial relationship with, versus how much flows through a G-D-S that also owns the agency. If the second one is what grew, your NDC program modernized your content and left your distribution power exactly where it was. That is fine, if it was the plan. It is a problem if you told your board it was disintermediation.

Eric Marketts

Second narrative, and it needs one distinction that the whole modern-retailing pitch rests on and rarely explains.

Steph Nell

The distinction is between the offer and the order. The offer is what the airline shows you. The bundle, the price, the ancillaries, assembled for a shopping request. We have spent a lot of this show on the offer. The order is what happens after you say yes. It is the record of what you bought, the thing that has to be serviced, changed, refunded, settled, and honored at the airport eight months later. In the old world, that record was a P-N-R living inside the Passenger Service System, the P-S-S, the operational nervous system the airline leases from Amadeus, or Sabre, or a handful of others.

Eric Marketts

And the modern-retailing idea is to pull the order out of there.

Steph Nell

The modern-retailing idea is to pull the order out of the P-S-S and put it in a dedicated order management system, an O-M-S, that the airline controls. That is the actual heart of the whole Offer and Order project. Not the NDC message we just talked about. The order system. Because whoever holds the order holds servicing, holds settlement, holds the customer record, and holds the airline's dependency for the next fifteen years. The offer is the marketing. The order is the mortgage.

Eric Marketts

That's the setup. Here's what happened.

Eric Marketts

The development this year is not one launch. It is that the order management system became the thing everyone sells. The category page is crowded now. FLYR markets an Offer and Order Management System it says is designed to displace the legacy P-S-S outright. OpenJaw sells an omnichannel order platform and has partnered with Lufthansa Systems on end-to-end Offer and Order. And then the P-S-S-neutral field, companies whose entire pitch is that they sit on top of whatever P-S-S you already run: Datalex, Accelya's FLX ONE, Astra by TPConnects, Retailaer, PROS, alongside the big-vendor platforms, Sabre Mosaic and Amadeus Nevio. The industry consensus phrase for 2026 is that the O-M-S is becoming the, quote, gravitational centre of retailing.

Eric Marketts

Hold on. That is a category, not an event. What is the dated thing?

Steph Nell

Fair challenge, and I want to be precise about it. The crowded-category claim is a synthesis. It comes from vendor product pages and from the January twenty twenty-six modern-retailing outlook that coined the gravitational-centre line. There is no single press release announcing that the order system category arrived. But there is a dated deal this year that shows you exactly what the category is doing to ownership, and it is not an order system deal at all. On June twenty-fifth, Amadeus published that British Airways is retiring the NDC platform it built in-house, on the seventeen point two standard, after more than a decade of running it, and moving to Amadeus Altea NDC on twenty-one point three.

Eric Marketts

BA built its own, and is giving it up.

Steph Nell

BA built its own, and is giving it up. And I want to be careful here, because that is an NDC platform, not an order system. Different layer. But it is the cleanest evidence this year of the thing this entire narrative is about. The carrier that went furthest toward owning its own retailing technology looked at the next version of the NDC standard and decided the vendor should carry it. That is the direction of travel underneath the crowded order-system category. Not airlines building. Airlines choosing whose to rent.

Steph Nell

And I want to flag the phrase everyone is using, because it is doing a lot of work. The consensus is that airlines are decoupling the order from the monolithic P-S-S. Decoupling. Except the same industry write-ups add, in the next sentence, that these systems are not yet fully independent of the P-S-S. So the order is being moved out of the vendor's monolith and into a different system that still, for now, depends on the vendor's monolith. Hold that, because it is the whole narrative.

Eric Marketts

Who are the P-S-S-neutral players again?

Steph Nell

That is the interesting group. Datalex, Accelya FLX ONE, Astra by TPConnects, Retailaer, and PROS all lead with P-S-S-neutral as the selling point. Which tells you something. When half a category is advertising that it is not locked to your incumbent vendor, it is because being locked to your incumbent vendor is the fear they are all selling against. They are selling freedom from exactly the dependency this industry has complained about for a decade.

Eric Marketts

Bull case.

Steph Nell

The bull case is that this is the single most important thing that can happen to airline retailing, and it is finally happening in production, not in slideware. If the order lives in a system the airline controls, then the airline controls servicing, controls how the offer gets constructed, controls its own data, and can change its retailing without filing a change request with a P-S-S vendor and waiting eighteen months. That is real. And a crowded vendor field is good news for the buyer. Ten credible O-M-S options means competition, means pricing pressure, means an airline that missed the first wave is not choosing between two incumbents who both have it trapped. The market that story one said stayed closed, this story says is prying open at the layer that actually matters.

Eric Marketts

Bear case.

Steph Nell

The bear case is that swapping one dependency for another is not independence, and a crowded field is a consolidation waiting to happen. Read the fine print of the consensus: not yet fully independent of the P-S-S. So on day one you are running an O-M-S that still leans on the monolith you were trying to escape, which means you now depend on two systems and the integration between them instead of one. And ten O-M-S vendors is not a stable market. It is a market before its shakeout. Some of these platforms will be acquired, some will be discontinued, and the airline that bet on the wrong one will be doing a second migration in 2031, having sold the first one to its board as the migration that ended migrations. The last time this industry picked a system of record, it stayed picked for thirty years. Picking wrong at the order layer is not a software regret. It is a fifteen-year re-lock.

Eric Marketts

Critical take.

Steph Nell

The critical take is that when every vendor in a market suddenly sells the same thing, the buyer's real risk stops being availability and becomes discernment, and this industry is much better at the first than the second. Two years ago the complaint was that airlines could not get out of the P-S-S to do Offer and Order. Story one of that era was scarcity. Now the complaint should flip, because the constraint flipped. The order management system is no longer hard to buy. It is hard to buy correctly. And notice what the vendors are competing on. Not whether they have an O-M-S, they all do now, but on the word neutral. P-S-S-neutral is the new moat claim, and it is a claim about a negative, about what the system is not chained to. A decade ago the danger was staying on legacy infrastructure. Today the danger is the opposite. It is picking a modern system that quietly rebuilds the same lock-in with a nicer interface, because you were so relieved to finally have options that you did not interrogate what you were signing. The scarce thing in this market is no longer the order system. It is the airline's ability to tell a real decoupling from a rebranded one. And nobody sells that.

Eric Marketts

Monday morning?

Steph Nell

Ask your O-M-S vendor, or the one you are evaluating, one question in writing. If we leave you in three years, what leaves with us, and what stays trapped. Who holds the order records, in what format, and can we service in-flight orders on a different system without a re-platform. If the honest answer is that your orders effectively cannot move, then the word neutral in the sales deck is decoration, and you are choosing your next fifteen-year dependency, not escaping your last one. Choose it like that.

Eric Marketts

Last narrative, and it is the one almost nobody in this industry called correctly.

Steph Nell

Start with the received wisdom, because six months ago it was everywhere. The line was that AI assistants cannot actually book travel. They research a trip, build an itinerary, tell you what a good flight looks like, and then hand you off to a website to do the actual transaction. And if you look at a consumer connector, that is exactly right. Expedia's connector inside Claude searches flights and hotels worldwide and returns real-time pricing and availability. That is the whole product. It discovers. It does not transact. Tripadvisor and Viator are the same shape.

Eric Marketts

And everyone read that as a limitation of the technology.

Steph Nell

Everyone read that as a limitation of the technology, and that is the error. It is not a technology limit. It is a trust limit. And from outside, those two look identical.

Eric Marketts

That's the setup. Here's what happened.

Eric Marketts

In July, American Express Global Business Travel launched the Egencia AI Connector inside Claude. Not a research tool. It searches, books, and manages air and hotel for a traveler, within corporate travel policy. Amex GBT describes it as the first deployment of its own agent-to-agent architecture, built specifically to put full T-M-C services inside AI platforms.

Steph Nell

Now hold those two facts next to each other, because the sequence is the entire story. The consumer connector, from one of the largest travel sellers in the world, stops at the handoff. The corporate connector, from a T-M-C, completes the booking. Machine-executed booking did not show up in leisure and trickle into managed travel. It showed up in managed travel first.

Eric Marketts

And there's a capability line running underneath this.

Steph Nell

There is, and it is moving quickly. Anthropic shipped a new model on September 1. On its own automation benchmark, machine-executed task completion went from seventeen point one percent to thirty-one point four percent in a single release cycle. That is vendor-reported, so take it as directional and not audited, and note that it is still under a third of tasks completed. Nobody should be declaring this solved. But the slope is the thing, because the slope is the input to whether the other agentic bets in this industry land in one year or five. Travelport, with Cognizant and Anthropic, said in May it is automating exchanges and rebooking and embedding disruption intelligence into Trip Services, with the first customer-facing capabilities reaching market this year.

Eric Marketts

Bull case.

Steph Nell

The bull case is that this is precisely the right sequencing and the industry should be relieved by it. Corporate travel has a policy. Approved suppliers, negotiated rates, class-of-service rules, an approval chain, a duty-of-care obligation, and an auditable trail. That is a bounded problem with a defensible right answer, and a bounded problem with a right answer is the only kind you should hand to a machine that is spending real money. Proving agentic booking inside those rails before turning it loose on unmanaged consumer spend is responsible sequencing, not timidity. And airlines do well out of it. An agent that books in policy books compliant, higher-yield corporate tickets, with less leakage and lower servicing cost per transaction. If you sell corporate, a channel just got faster and cheaper to sell through.

Eric Marketts

Bear case.

Steph Nell

The bear case is that the first at-scale pattern for how a machine buys a trip is being set right now, and it is being set by the intermediary rather than the airline. Read what Amex GBT said it built. Agent-to-agent architecture, to put full T-M-C services inside the AI platform. Full T-M-C services. In that architecture the airline is a supplier. Its content arrives, gets policy-filtered, and gets selected or not by somebody else's agent against somebody else's rules. The airline is not a party to the conversation between the traveler's agent and the T-M-C's agent. It is the thing being shopped. Travelport is building the same shape at the G-D-S layer. So two intermediaries are actively constructing the agentic booking flow, and I cannot point you to a single airline that has published a position on what its offer is supposed to look like when the buyer is a machine following a rulebook.

Eric Marketts

Critical take.

Steph Nell

The critical take is one sentence, and it explains the whole inversion. An agent can be held to a policy. It cannot be held to a preference. Managed travel went first because managed travel could hand the machine a constraint set and a defensible definition of a correct booking. Consumer travel cannot. There is no auditable right answer to what a given person wants on a Tuesday, which is why the consumer connectors still stop at the handoff, and why they will keep stopping there longer than the capability curve alone would suggest. Now sit with what that means for retailing. This industry has spent a decade building toward the personalised, live-assembled, per-traveler offer. The first channel where machines genuinely buy is the one channel that exists specifically to override individual preference with corporate rules. Your richest offer walks into a filter whose entire job is to flatten it to a fare, a carrier, a cabin, and a compliance flag. And this is the third time in this episode. NDC scaled and the G-D-S kept the shelf. The order started moving and the P-S-S kept the dependency. The machine learned to book and the T-M-C owns the policy layer that decides what it is even allowed to see. The airline has now modernized the offer at three layers and has not once gotten closer to the rules that govern which offers get looked at.

Eric Marketts

Monday morning?

Steph Nell

If you sell corporate, find out whose policy engine your content passes through before an agent ever sees it, and get an honest answer to one question. When your NDC offer goes through that filter, what survives it. If your branded fares, your bundles, and your ancillaries come out the other side as a fare and a carrier code, then every dollar you have spent on offer richness is being spent upstream of a system designed to discard it. That is not an argument against selling corporate. It is an argument for negotiating what the policy layer is obligated to carry, and doing it now, while there are two agentic booking stacks in this market and neither one has locked its shape.

Eric Marketts

Here's the spine.

Steph Nell

Two layers of airline retailing moved this year and one has not moved since 2024, and the combination is the story. The order management system became a live, competitive, production category. A machine started completing bookings end to end. And NDC has sat inside a two-point band for two years, while the industry called it scaling on the strength of a global number nobody publishes. Modernization is not the problem anymore. The problem is that at all three, the ownership did not move with it. NDC stopped growing at exactly the point where growing would have meant building direct muscle instead of handing content to the aggregator. The order is moving into systems that still lean on the P-S-S. And the machine that finally books does it inside a policy engine the airline does not write. Up and to the right where somebody else owns it. Flat where the airline would have had to own it itself.

Eric Marketts

So what is the rule.

Steph Nell

Modernization is not ownership, and this industry keeps buying the first and reporting the second. A better offer is not control of the offer. A dedicated order system is not independence if the order cannot leave it. Being bookable by an agent is not owning the transaction if somebody else writes the rules that agent follows. And an adoption statistic is not adoption if nobody can tell you where it came from. Every one of these is a genuine upgrade to the product and none of them, on its own, moves the balance of power toward the airline. The upgrade is easy to buy. The power has to be negotiated, and it almost never is, because the upgrade feels like enough.

Eric Marketts

And the sharper version, for the people who have been here a while.

Steph Nell

The sharper version is that this industry has learned to modernize without ever renegotiating. It will spend nine figures to escape a dependency and then re-sign the dependency in a better font. It did it with NDC and the G-D-S. It is at risk of doing it again with the O-M-S and the P-S-S. And it is standing at the edge of the agentic channel ready to do it a third time, in a market where the rules are being drafted this year and no incumbent is forcing its hand. The technology keeps advancing and the terms keep not moving. The engineering ships. The ownership does not.

Eric Marketts

And the metric.

Steph Nell

One question, and it is the same one at every layer. When you buy the modern version of anything this year, the NDC program, the order system, the agentic channel, ask what percentage of the value you now control versus what percentage you are renting, and get the honest number before you sign, not after. If nobody in the building can produce that number, you are buying modernization and calling it ownership, and the gap between those two words is where the next decade of this industry's margin is going to leak.

Eric Marketts

If you own a channel or retailing P and L, send this to whoever approved your last modernization business case. Not to relitigate it. To read the outcome next to the pitch and see whether the words match.

Steph Nell

And the standing ask, unchanged from last week. If you are a carrier about to publish real terms for a machine buyer, on authentication or on pricing, we still want to hear from you. And now a second one. If you are an airline that actually renegotiated the terms and not just the technology, that kept control of the shelf and not only the storefront, we especially want to hear from you, because we are looking and we are not finding many.

Eric Marketts

Send us the correction too, if we got something wrong. We read them, and we read them on air.

Eric Marketts

We'll be back next Monday. I'm Eric Marketts.

Steph Nell

I'm Steph Nell. Thanks for listening.

Eric Marketts

Buy the upgrade if you want. Just don't call renting a house owning one. Stay sharp out there.