Channel Managers Are a Partial Answer

Essay 5 ended by naming the channel manager as the clearest partial answer the experiences industry has produced. Partial is the operative word. The channel manager is not glamorous. It does not appear in industry keynotes under its own name. It does not attract the attention that consumer-facing marketplaces do, or the attention that reservation systems attract when they touch the operator’s daily work directly. And yet much of the industry’s non-direct booking flow passes through this layer, and has done so since the category emerged as a distinct piece of infrastructure.

If the infrastructure question is going to be asked properly, the category that sits in the middle of most current traffic is the place to start. The purpose of this essay is to describe, carefully, what the channel manager is, what problem it solves, what problems it was not designed to solve, and why it has ended up carrying many of those additional problems anyway. None of this is criticism of the category. A solution that fits its original problem well has no obligation to fit every problem that shows up afterwards. The question is how much of the weight the infrastructure of this industry now needs to carry can continue to be carried by a category built for a narrower purpose.

What the channel manager is, precisely

A channel manager is a routing tool. It is not a reservation system. A reservation system holds inventory, takes bookings, manages refunds, and produces the operator’s daily manifest. A channel manager does none of those things. It routes content, availability, bookings, and cancellations between a reservation system on the supplier side and a set of resellers on the distribution side. Its purpose is singular,^¹ though its commercial neutrality is not: some channel managers operate as standalone routers, others sit inside reservation-system or distribution-side organisations. Three kinds of participants use it, and the distinction matters for what the category is asked to do.

Suppliers. Operators running their own reservation system use a channel manager to expose inventory to many resellers without building and maintaining a direct integration to each one. The alternative is point-to-point integration: the operator’s engineering team builds one connection to one reseller, another to another, and so on. Each integration is a line item that consumes engineering time, carries a deprecation lifecycle, and breaks when either side changes. The channel manager collapses this cost by replacing many operator-to-reseller connections with one operator-to-channel-manager connection, which then fans out to many resellers.

Res-techs. Other reservation systems that serve a book of operators use a channel manager for a structurally similar reason. A reservation system used by a hundred or a thousand operators does not want to maintain direct integrations to every reseller either. By routing its operator base through a channel manager, it shares the integration-build and maintenance cost across the aggregate and gains, by aggregation, easier access to resellers that might not otherwise negotiate individually.

Resellers. A large reseller that wants access to thousands of operators does not want to integrate with every one of them directly. A channel manager offers the reseller a consolidated API surface through which many operators can be reached. This is the category’s second economic good: consolidated access.

The singular purpose is routing. The two economic goods the category delivers, shared integration cost and consolidated access, follow from that purpose.^²

What the channel manager solves well

These two economic goods are real. Neither has been replaced by any other category since channel managers have been in widespread use. No subsequent layer of the stack has made point-to-point integration cheaper, and no subsequent layer has given resellers a better route into the fragmented supply base at scale.^³ The channel manager’s durability is not an accident of incumbency; it is a consequence of solving a specific problem well enough that the solution has not been dislodged.

It is important to state this clearly before the rest of the essay becomes about what the category does not do, because the rest of the essay will be about what the category does not do. Both descriptions can be true at the same time. A category can solve its original problem well and still be asked, by the market’s evolution, to solve problems it was not designed for.

What the channel manager was not designed to solve

Four structural problems sit beyond the category’s original design envelope. None of them are failures of execution. They are features of scope.

Real-time availability consistency across deep chains. Every system in the booking chain has its own latency budget. The channel manager sits in the middle of that chain, and each hop in either direction adds to the budget. When the chain is two hops deep, the compounding is manageable. When the chain is three or four hops deep, as it often is in practice, the time any individual boundary has to complete its work shrinks toward zero. This is not a failure of the channel manager; it is a structural property of routing. A router routes, and cannot eliminate the cost of the routes it carries. Genuinely sub-second availability consistency across a deep chain is not achievable by adding more routing. It would require the chain to get shorter, not longer.^⁴

End-to-end data integrity across boundaries. Each end of a channel-manager connection has its own schema. The supplier’s reservation system describes its inventory in one product vocabulary: age bands, pricing categories, unit types, meeting-point definitions. The reseller expects its own vocabulary on the way in. The channel manager negotiates a translation between them at the point of integration. That translation is a choice. When the reseller’s schema does not map cleanly onto the supplier’s, the channel manager picks a mapping. The mapping is an engineering decision made once and carried in every subsequent booking that passes through it. It is not possible to route heterogeneous schemas without picking compromises, and picking compromises is the work.^⁵ Integrity, in the sense of no drift between what the operator thinks it is selling and what the reseller thinks it is buying, is bounded by the quality of the mappings and by the rate at which either schema changes beneath them.

Distribution economics, at structural scale. This limitation deserves more care than the others. Channel managers do influence distribution economics at the margin. Some negotiate preferential commission terms on behalf of their book of operators, offering resellers volume in exchange for rate. Some offer standardised payment and refund frameworks that their larger customers use as a baseline. These are real contributions, and they are not trivial. But the underlying commercial relationships, the commission cut itself, payment timing, refund policies, voucher-redemption logistics, chargeback handling, remain negotiated out of band, directly between suppliers and resellers, or between their parent organisations. The channel manager carries the booking through. It does not carry the commercial framework. In other words, the category shapes economics at the edges without reshaping them structurally. The distinction matters, because the industry’s distribution-economics problems are not edge problems; they are structural.

Neutral settlement. Essay 5 argued that the Billing and Settlement Plan on the international side and the Airlines Reporting Corporation on the US side were preconditions for the aviation clearinghouse, not afterthoughts.^⁶ Experiences has no equivalent, and the channel manager has not become one. Money does not flow through the channel manager. Settlement is peer-to-peer between reseller and supplier, or between reseller and an intermediary separate from the connectivity layer entirely. This is not an oversight of the category; it was never part of the category’s design. A routing tool routes data. A settlement layer routes money. They are different pieces of infrastructure, with different legal, regulatory, and commercial requirements. Building one does not imply building the other. The consequence reaches Essay 5 directly. If a neutral inventory layer without a neutral settlement layer is only half of what made the aviation pattern work, the closest thing the experiences industry has produced to a neutral layer does not address the other half at all.

Each of these four is a structural feature, not a flaw. Together they define the shape of the category’s original envelope.

Why the category gets pushed past its scope

A category designed to do one thing well sometimes ends up doing many things imperfectly, because two pressures pull on it simultaneously and there are no alternative layers to which it can delegate.^⁷

The external pressure is the absence of adjacent layers. In adjacent industries, a newly emerging distribution problem might be absorbed by a different piece of infrastructure: a standards body, a settlement network, a regulator, an institutional intermediary. The experiences industry has few of these. When a new distribution-connectivity problem appears, the only layer close enough to absorb it is, in most cases, the channel manager. Content-freshness management on behalf of operators who do not automate content, schema reconciliation between resellers with conflicting product models, pricing synchronisation under dynamic availability, workflow support for cancellations and refunds, each of these has been handed to the channel-manager category over time, mostly by default.

The internal pressure is competitive dynamics within the category itself. A channel manager that implements more reseller-specific features, more pricing logic, more content-management sophistication, more workflow assistance, tends to win commercial deals against one that does less. The economic incentive is to accumulate scope, not to stay lean. A category pushed in this direction will absorb adjacent problems, even those it was not designed for, as long as customers reward the absorption with spend.

Both pressures are rational, and together they explain why the channel manager today carries a great deal of weight that the category in its original form did not. That weight is real work, done by real engineering teams, resolving real operational problems the rest of the stack does not resolve. Calling this a feature of the category’s evolution is analytically accurate. Calling it the category’s long-term equilibrium may not be.

The structural ceiling

A category can be pushed some distance past its original design envelope before the strain becomes visible from outside. That distance has been considerable in the channel-manager case. The category has absorbed schema heterogeneity, content-management responsibilities, pricing-synchronisation complexity, and workflow functions its original design did not contemplate. Each absorption was individually reasonable. The cumulative effect is that the category now sits under commercial and operational pressures it was not built to hold.

The signs of structural strain are becoming visible now. Chains of intermediation have grown deeper, not shallower, because new resellers often reach the supply base through another layer of aggregation rather than directly. The latency budget has not grown more generous in response. Translation burden has grown with the diversity of product vocabularies, because the industry has not converged on one and shows no sign of doing so.^⁸ Larger resellers, responding to the compounding cost of deep chains, have invested in proprietary supply stacks that let them bypass parts of the channel-manager layer and reach operators more directly, through their own connectivity specifications, their own integrations, or their own acquisitions of reservation-system vendors upstream.^⁹ A category whose largest customers are building infrastructure to get around parts of it is a category showing structural strain.

None of this is a failure. The channel manager’s core economic goods, shared integration cost and consolidated access, continue to be real, and continue to be delivered. The strain is at the edges of the scope the category has absorbed, not at the centre of what it was built for.

Close

The channel manager is the clearest partial answer the experiences industry has produced to the distribution-connectivity problem. Partial is the honest word. The category solves two real economic goods and solves them durably. It was not designed to solve real-time availability consistency across deep chains, end-to-end data integrity under heterogeneous vocabularies, distribution economics at structural scale, or neutral settlement. It has been pushed to carry some of those additional problems, with meaningful success at the edges and visible strain at the centre as scope has accumulated.

Whether a layer emerges that sits underneath or alongside the channel-manager category, carrying weight it was not built for, is the structural question the rest of this arc takes up. The next essay turns to the arrival of AI-mediated discovery and agentic distribution, and asks what that wave, now reshaping adjacent travel categories, actually changes about the architecture described so far. Less, in some ways, than the current hype cycle suggests. More, in ways the industry has not sat with, than it is ready for.

The channel manager works. The question is what it cannot carry, and what that implies for an industry whose infrastructure needs are growing faster than any one category was built to absorb. Whatever layer emerges next must build on what the channel manager already solved, not pretend it failed.