The Real Cost of Being the Default

Competition enforcement against large technology platforms has produced two decades of headlines, substantial fines, and remarkably little change in market structure. That gap — between the visibility of enforcement and its effect — is the most interesting thing about the field, and it has an explanation that is more about economics than about regulatory failure.

The explanation is defaults.

Why fines do not change behaviour

A penalty, however large, is a one-time transfer. If the conduct it punishes generates a recurring revenue stream that exceeds the penalty’s amortised cost, the rational response is to pay and continue, adjusting only enough to argue compliance. This is not cynicism; it is arithmetic, and any competent finance function performs it.

What changes behaviour is a remedy that alters the structure of the market — separating a business, prohibiting a contractual arrangement, or mandating access on terms the firm would not choose. These are rare, slow, and vigorously contested, because they are the only remedies that actually cost anything ongoing.

Edgewisely’s analysis of how a major ad-tech antitrust penalty could reshape markets and redefine the limits of platform power turns on precisely this distinction: the number in the headline is the least consequential part of the decision, and the structural questions underneath it are where the actual stakes sit.

The default as the asset

In most platform markets, the contested asset is not the product. It is the position of being what happens when a user does nothing.

The default search engine on a browser. The default browser on an operating system. The default payment method in an app store. The default assistant on a phone. In each case the alternative is available, often free, often good — and adoption of the alternative is a small fraction of what the quality difference would predict, because the number of users who change a default is consistently tiny.

This is why the most valuable commercial arrangements in technology are default placements, and why they are the arrangements regulators find hardest to address. Prohibiting a payment for default placement does not create competition on its own; it changes who captures the value of the default, and the default remains a default. Choice screens have been tried and produce modest, decaying effects. Users pick the familiar name and move on.

The uncomfortable conclusion is that default power is a property of human behaviour rather than of corporate conduct, and competition law is not well equipped to regulate inertia.

Where AI changes the calculation

The current platform transition is interesting because it briefly makes defaults contestable, and then, quite quickly, makes them stronger.

Contestable, because a new interaction layer reshuffles which defaults matter. If users increasingly get answers from an assistant rather than a search results page, the default search engine becomes less valuable and the default assistant becomes more so. New defaults are up for grabs in a way they have not been for fifteen years.

Stronger, because the entities best positioned to win those new defaults are the ones that already control the surfaces where defaults live — operating systems, browsers, devices. The transition is an opportunity for incumbents to re-anchor their position in a new layer, not an opening for challengers.

Apple’s decision to license external models for its assistant illustrates the dynamic from an unusual angle. Edgewisely’s account of how Apple traded fifteen years of integration orthodoxy for a shipping date is, read structurally, a story about a company choosing to protect the surface rather than the technology. Apple does not need to build the best model. It needs the assistant on its devices to be good enough that users do not go elsewhere, because the surface is the asset and the model is an input to it.

That logic is available to every incumbent that owns a surface, and it is not available to challengers who own only a model.

What this suggests about remedies

If defaults are the asset and inertia is the mechanism, three implications follow for how enforcement should be designed.

Structural remedies beat conduct remedies. A rule that says “you may not do X” requires ongoing monitoring of a firm with vastly more information than the monitor. A remedy that changes what the firm owns requires monitoring once. Regulators consistently underestimate how much of a conduct remedy’s value evaporates during the years of supervised compliance.

Timeliness dominates severity. A remedy imposed after a market has tipped is archaeology. The transition happening now — where interaction defaults are genuinely in play — is the window in which intervention could plausibly affect structure, and it will close in two or three years. Enforcement processes currently run longer than that.

Interoperability is the underused tool. The reason defaults are sticky is that switching costs are real: data, history, integrations, muscle memory. Mandated portability and interoperability attack the mechanism rather than the conduct, and they do not require a regulator to determine what a competitive market should look like. They are also, not coincidentally, the remedies that platforms resist most strenuously.

The honest counterargument

There is a serious case on the other side, and it deserves stating rather than caricaturing.

Integrated platforms deliver genuine benefits that are hard to replicate in fragmented markets: consistent security models, predictable developer economics, a single accountable party, and product experiences that work without configuration. Aggressive structural intervention can degrade those benefits, and the beneficiaries of degradation are not necessarily consumers.

There is also a real question about whether regulators can act quickly enough to matter without acting hastily enough to err. The window argument cuts both ways: intervening early in a transition means intervening before anyone knows which arrangements are durable and which are transitional. Several enforcement actions in the last decade addressed conduct that the market had already routed around by the time the decision landed.

Neither argument resolves the matter. What both suggest is that the productive debate is not “should platforms be regulated” but “which specific mechanism produces the observed harm, and what remedy attacks that mechanism rather than its symptoms.” That is a narrower conversation than the one usually held, and a more useful one.

What businesses downstream should take from this

For companies that are neither platforms nor regulators — which is almost everyone — the practical reading is about dependency rather than about policy.

If a meaningful share of your customer acquisition comes through a surface you do not control, you are exposed to a decision you will not be consulted about. That exposure does not depend on the platform behaving badly. A default can be reassigned for entirely ordinary commercial reasons, and the change will reach you as a traffic number falling rather than as a notification.

The mitigation is unglamorous: know what proportion of your demand arrives through each channel, know which of those channels are defaults you do not own, and maintain at least one route to your customers that nobody can reassign. Direct relationships, owned communications and brand recognition are all expensive and all slow to build, which is why they are consistently underfunded relative to paid channels that produce measurable results next quarter.

Companies that survived previous platform transitions almost always did so because they had one of those routes before they needed it. Companies that did not survive usually had excellent channel economics right up until the channel changed.

The default is the asset. Everything else is commentary.