Where Value Migrates When Interfaces Die

The previous essay ended on a promise. When the interface commoditises, value does not disappear. It migrates. The migration has happened in search, in social, in cloud infrastructure, in retail distribution, and it has gone to the same kind of place in each. Before arguing what that means for the experiences industry, it is worth laying out the pattern in the open, because the pattern is the argument.

Four cases

Search. By the mid-2000s, the front-end of search had converged toward a generic shape: a single box, a ranked list of results. The interface ceased to be a site of meaningful differentiation. Value migrated to the index underneath, to the crawl coverage, the ranking signal, and the auction mechanism for monetising intent, none of which a new front-end could cheaply replicate.^¹

Social. By the late 2010s, social-platform interfaces had converged toward a common shape: vertical infinite scroll, algorithmic ranking, short-form video as the dominant unit. Value migrated to the attention graph underneath, to the infrastructure for identity resolution, audience modelling, and ad placement that no new interface could assemble at comparable scale. New interfaces kept being built. The position underneath did not move.^²

Cloud. By the early 2020s, the consumption surface had largely converged toward a standard shape: container orchestration, infrastructure-as-code tooling, a common vocabulary for object storage, queues, and serverless functions. The interface became portable. Value migrated to what the interface could not make portable, which was the physical compute, networking, and storage at scale underneath. A new provider can replicate the APIs. Replicating the physical footprint is a different kind of project.^³

Retail. By the mid-2010s, the e-commerce front-end had converged toward a pattern any competent team could assemble: search, product page, cart, checkout, review. What did not converge was the layer underneath that made the front-end’s promises deliverable. Warehousing density, last-mile logistics, demand forecasting, returns processing. Value migrated to that layer. The dominant competitive position in e-commerce was won at the fulfilment layer, not at the storefront, and that position has been structurally durable in a way no front-end advantage has been.^⁴

Four transitions, four different industries, the same structural move each time.

The rule

The rule these cases illustrate is simple, though it is more frequently observed than stated plainly.

An interface sits between demand and supply. When interfaces become cheap to build and user switching costs are low, the margin available at the interface compresses. Whatever sits immediately underneath the interface, and is expensive or slow for a new entrant to assemble, is where the compressed margin reappears. Value moves to the layer that holds demand or supply coherent at a scale new interfaces cannot easily reassemble for themselves.^⁵

This is a structural observation rather than a strategic one. Nobody plans for it. It happens because interfaces that need a particular layer in order to function end up paying for that layer, one way or another, and the layer that multiple interfaces need accrues the returns that any individual interface cannot hold. The interface becomes a commodity. The underlying layer becomes an infrastructure position.

Two things are worth noting about the rule as it has played out historically. The first is that the migration is rarely obvious in real time. The players in each of the four transitions above continued to treat the interface as the site of competition well past the point at which the competitive surface had moved underneath. The second is that the migration is not reversible once the underlying layer reaches sufficient scale. A cloud competitor cannot out-scale the leading hyperscaler by releasing better APIs. A search competitor cannot out-index the dominant index by releasing a better front-end. The position, once held, is durable.

Why this applies to experiences

The experiences industry has, in its modern history, treated the interface as the primary site of competition. Booking flows, conversion funnels, loyalty schemes, brand campaigns, search-engine marketing, product-detail page optimisation. An enormous share of operational and strategic effort across distributors and suppliers has gone into the interface. If the argument of Essay 8 is correct, that interface surface is about to commoditise. Agents do not care about conversion design. Agents do not respond to brand equity. Agents do not absorb confusion on behalf of the user, which means the interfaces that competed on their ability to manage human confusion cannot compete on that ability when the counterparty is a machine. What agents care about is whether the supply underneath can be queried, priced, and transacted against reliably and at scale.^⁶

This is the same structural move the four prior transitions described. The interface competes itself toward a generic shape. Whatever sits immediately underneath, and holds supply coherent at scale, accrues the value that the interface can no longer hold. The experiences version of this pattern is not that booking pages disappear. It is that the page becomes less defensible than the supply state beneath it.

Where the pattern says the value goes

In the experiences case, the layer underneath is the one Essay 8 described. It is the layer that can present, across the supply base, a machine-readable product catalogue that is current rather than staged, an availability surface that is authoritative rather than eventually consistent, a transaction protocol that completes deterministically, and a reachability graph that does not depend on someone keying updates into an extranet once a week.

Whoever holds that layer, across enough of the supply base to matter, holds the position value migrates into as the interface commoditises. This is not a strategic claim. It is the same claim the pattern has already made in four other industries.

What is different about experiences is that the layer in question does not currently exist at industry-wide scale. Essay 6 traced why the channel-manager category is structurally partial: it solves for shared integration cost and consolidated reseller access, not for real-time authoritative availability, not for end-to-end data integrity, not for neutral settlement.^⁷ The layer the pattern says value migrates into is therefore either still to be built, or is about to be built in a shape that locks the industry into a particular path.

Two paths

There are, structurally, two shapes that layer can take.

The neutral layer. A layer that suppliers and distributors both connect to on commercial terms that are independent of whether they compete with each other. The layer holds the supply coherence. It does not own the supply and it does not own the demand. Its position in the graph is analogous to the position the Billing and Settlement Plan occupies in aviation settlement, or that cloud compute occupies underneath software-as-a-service, or that card-scheme infrastructure occupies underneath consumer payments. The value accrues to the layer, and the layer is available to every participant on symmetric terms.^⁸

The vertically integrated supply surface. A configuration in which coherence is achieved inside a single commercial perimeter rather than across the industry. The layer exists, and is capable of presenting the supply surface an agent needs, but it is owned by a participant whose commercial interest is not neutral to the outcomes of other participants connecting to it. This is structurally analogous to how e-commerce marketplaces resolved for the long tail of independent sellers. The layer exists. Using it is conditional on terms set by a participant with its own interest in where demand flows.^⁹

Both paths are consistent with the migration pattern. Both have happened in other industries. Neither is a prediction.

The asymmetry

These paths are not symmetric outcomes for the industry. They produce meaningfully different shapes.

In the neutral-layer outcome, the fragmented supply base retains access to agent-driven demand without being absorbed into any one participant’s commercial perimeter. An independent operator in a small European city stays an independent operator, reachable by every agent, on terms negotiated with a neutral layer rather than with a competitor for the end customer. Distributors stay distributors. The competitive surface between them moves back toward the things distributors can actually compete on, which are curation, trust, bundled experience, and demand generation within specific traveller segments. In the vertically integrated outcome, access to agent-driven demand is conditional on sitting inside whichever platform holds the coherent supply surface. Suppliers whose inventory is not visible inside that platform are unreachable to the agent. The platform’s commercial terms become the industry’s commercial terms. The long tail of independent operators either consolidates, gets absorbed into private labels, or loses reach. It is possible to believe one outcome is more likely than the other. It is also possible to hold no strong view. What is not really possible, if the pattern from the four prior transitions holds, is to assume the migration does not happen at all. The migration is what always happens. The question is only into what shape of layer it resolves.

Close

The experiences industry is inside a transition of exactly the sort that has played out several times already in adjacent industries. The commoditisation of the interface is underway. The migration of value to the layer underneath is the structural consequence. The only variable is the shape of the layer the value migrates into, and that variable is being decided now, in the specific technical and commercial choices the industry is making about how supply is going to be presented to its next counterparty.

The choices are being made distributedly, by participants who do not always recognise that the cumulative result is a structural settlement. A reservation system that does not invest in real-time availability is contributing to the settlement. A distributor that builds its own supply integrations rather than connecting to a shared layer is contributing to it. A neutral-layer initiative that fails to reach critical mass is contributing to it by default. None of these decisions appears, locally, to be a choice about the industry’s shape. Cumulatively, that is what they are.

The next essay argues that the point at which the cumulative choices crystallise into a settlement that is hard to undo is close at hand. The industry is at its inflection point.