In the previous piece we argued that pricing has to move toward outcomes. This is where it gets difficult.
The idea sounds flawless in a sales meeting: charge for results, align the agency's interests with the client's, stop arguing about hours and start talking about impact.
Then the first measurement quarter arrives and the real conversation shows up: the client closed three large deals and nobody can say how much of that belongs to the agency.
That is where most performance pricing models fall apart. Not for lack of goodwill, but because they were signed before the hard part was worked out.
Three conditions before signing
For outcome-based pricing to work, three things have to be true at the same time: measurement, attribution, and control.
If an agency runs paid media for an e-commerce business with a short funnel and traceable conversions, tying part of the fee to results is relatively simple. The distance between what the agency does and what happens in the business is short.
Now picture a B2B agency working with a company whose sales cycle runs nine months.
Marketing creates demand. An SDR opens the conversation. A rep carries the opportunity forward. There are several meetings. Months later, the contract is signed.
Whose revenue is that?
The answer is rarely obvious. And if it wasn't agreed in advance, it turns into a negotiation at the worst possible moment, when there is already money on the table and both sides feel they are right.
That is why the real challenge of performance pricing isn't commercial. It's building a shared measurement architecture.
What should be written down before the contract
Before the contract exists, both parties should have agreed, in writing, on seven things.
The KPI. What exactly the indicator is. Not "more sales," but which event, in which system, under which definition.
The baseline. What it is measured against. An average of recent months, the same period last year, an agreed starting point. Without a baseline there is no improvement, only opinion.
The source of truth. Which platform wins when two numbers disagree. The CRM, the analytics platform, the billing system. This prevents arguments that have no end.
The attribution window. How long a lead counts as generated by the agency. In long cycles, this single definition drives much of the economic outcome.
Each party's responsibilities. If the client doesn't follow up on leads within 48 hours, if the sales team is understaffed, if materials get approved late, the number moves. Better to say so upfront.
External variables. Price changes, seasonality, competitive moves, product changes. What happens to the agreement when the ground shifts.
The moment of calculation. When the period closes, when payment is settled, what happens to deals that land right on the edge.
Without those rules, the supposed alignment turns, fairly quickly, into an argument about attribution.
Pricing doesn't define how much you charge. It defines how you work
There is something that usually goes unnoticed when a pricing model is designed.
A pricing model doesn't only determine how much an agency bills. It determines which behavior gets rewarded inside the organization.
Charging a percentage of media spend indirectly rewards increasing that spend.
Charging purely by lead volume rewards volume, even as quality drops.
Charging by the hour penalizes efficiency, at exactly the moment when technology makes it possible to be far more efficient.
Charging a retainer with no objectives disconnects compensation from results, and over time it disconnects the conversation with the client as well.
Designing pricing is, at bottom, designing incentives.
The right question isn't how much we should charge. It's which behavior we want to reward in our team, and which outcome the client wants to reward.
When those two answers line up, the model starts to run on its own.
Not every agency should charge the same way
Moving toward value-based models doesn't mean every agency has to adopt the same formula.
The model depends on the kind of value each one creates.
A creative agency usually works better with projects and value-based pricing, because its contribution is hard to reduce to a single indicator.
A performance agency can carry a larger variable component, because it works close to the data.
A strategy consultancy can charge for projects, advisory, and value created.
A technology agency can combine implementation, subscription, and outcome pricing.
An integrated agency probably needs several of these at once.
This evolution doesn't lead to a single pricing model. It leads to a more flexible architecture, where each type of work is priced in the way that best reflects the value it produces.
Start with what you already know how to measure
There is a lower-risk way into this.
Instead of redesigning the entire commercial model at once, pick one account where measurement is already reasonably clear and test a small variable layer there. A percentage of the fee tied to one indicator both sides look at every week.
That exercise tends to reveal something uncomfortable and very useful: how much of what the agency delivers today isn't being measured by anyone.
And that discovery, more than any pricing formula, is usually what changes the conversation with the client.
Back to Blog