‍


I recently listened to Sir Martin Sorrell discuss the future of agency pricing. His argument was simple: as AI makes production faster, agencies should move away from billing for time and instead charge for outputs. It's difficult to disagree. But the conversation left me wondering whether we're solving the wrong problem. Changing the unit on an invoice is easy. Changing the economics behind it is something else entirely.

The efficiency paradox‍

For decades, agency economics have begun in roughly the same place: people, time and cost. Whether the commercial wrapper is a retainer, project fee or production estimate, the spreadsheet underneath usually adds labour to overhead and margin.

AI is now disrupting that calculation. Experienced strategists, creatives and media specialists can research, develop and produce strong work in a fraction of the time. That should reward expertise and better systems. Under a time-based model, it can do the opposite.

Imagine two agencies. Agency A delivers a positioning strategy in four weeks. Agency B delivers work of equal quality in four days. Which should cost more?

Our industry’s accounting logic still tends to favour Agency A because the effort is more visible. Yet Agency B may have invested years in talent, tools and workflows that compress time without compromising quality. If greater capability leads to fewer billable hours, are we saying the better an agency becomes, the less it should be paid?

Output is not the same as outcome

This is where the vocabulary matters. An output can be counted: a strategy, a campaign platform, a film or a set of assets. An outcome is the business effect: sales, market share, customer acquisition, brand preference or profit.

Output-based pricing may solve part of the AI problem because it no longer punishes speed directly. Outcome-based pricing goes further by linking agency reward to client performance. In theory, that creates alignment. In practice, attribution becomes the fault line.

Suppose a campaign coincides with an additional $10 million in revenue. How much of that value belongs to the creative idea? How much to media, distribution, pricing, product availability, sales execution, market conditions, timing or luck?

No CFO can isolate every variable with complete confidence. Agencies may influence an outcome without controlling it, while clients may control decisive factors the agency cannot see. A pure outcome model can therefore ask one party to carry risk created by many others.

That leaves an uncomfortable possibility: some models described as outcome-based are still calculated from time and materials behind the scenes, then presented in more progressive language.
‍
‍

The CFO’s challenge

Our own CFO often brings the debate back to first principles. Strip away the terminology, the AI narrative and the commercial packaging. Is every creative output still, ultimately, the cost of people plus overhead plus margin? It is a difficult challenge to dismiss. Agencies need a defensible cost floor. Clients need transparency and predictability. Neither requirement disappears because the industry adopts a new label. But cost is not the same as value. If pricing never moves beyond cost, agencies cannot capture the upside created by better judgment, distinctive intellectual property or radical efficiency. If pricing ignores cost altogether, the model may become impossible to operate or govern.
‍

What is the future?

Perhaps the future is neither time-based nor purely outcome based. It may be a more dynamic model built around value, control and shared risk.

A base fee could pay for access to the agency's capability: its talent, systems, intellectual property and judgment. Output pricing could apply where deliverables are clearly defined. A value premium could reward speed, originality or the importance of the problem being solved. And an outcome incentive could apply where the agency has genuine influence, reliable data and an agreed method of attribution.

The mix would change according to the assignment. A production brief may suit output pricing. A performance campaign may support an outcome component. A transformation project may require a capability fee because its value cannot be reduced to a set of assets or a single commercial result.

This is less dramatic than declaring the death of the billable hour. But it may be more realistic.

The future may not belong to one universal pricing model. It may belong to agencies and clients that can agree, before the work begins, what is being valued, who controls the result, how success will be measured and how the risk should be shared.

‍

So, what is an idea worth?

This is the question no pricing model fully resolves. Would anyone have accurately priced ‘Just Do It’ or ‘Think Different’ before those ideas entered culture? Their value became visible only after people experienced them. The invoice, however, needed a value before the outcome existed.

That may be the central contradiction of agency economics: ideas are priced prospectively but valued retrospectively.

AI makes that contradiction harder to ignore. When execution becomes faster and more abundant, the scarce things are no longer hours or assets. They are judgment, originality, accountability and the ability to make the right decision before the market reveals the answer.

Perhaps that is what agencies should learn to price and what clients should be willing to reward.

Until then, changing the language from time to outputs or outcomes may improve the model, but it will not settle the deeper question: what is a great creative idea actually worth?

‍