Most businesses compare agency fee models on the number. The number matters less than the structure, because the structure decides what the agency does on the days when the right answer is inconvenient.
Percentage of ad spend
How it works: you pay a share of what you spend, typically 10–20%.
What it rewards: spending more.
Consider a Tuesday where the honest recommendation is to cut a channel that has stopped working. Under a percentage model, giving you that advice reduces the agency's own income. They may give it anyway — plenty of good people do — but you have built a structure where doing the right thing costs them, and over a long relationship that pressure shows.
It also penalises efficiency. An agency that halves your cost per customer while holding budget flat has done excellent work and earned exactly the same fee. One that doubles your budget at the same efficiency has done nothing clever and earns twice as much.
Performance or commission
How it works: paid per lead, per sale, or as a share of revenue.
What it rewards: conversions that are cheapest to produce, not conversions that are most valuable to you.
This is the model that sounds most aligned and is frequently the least. It pushes spend toward branded search, remarketing and existing customers — demand you already had — because those convert most reliably. The agency's numbers look superb while total revenue barely moves.
For lead generation it is worse still. A per-lead fee rewards volume, and volume is trivially achievable by lowering the bar on who counts as a lead.
Any model that pays per conversion creates pressure to redefine what a conversion is.
Flat retainer
How it works: a fixed monthly fee, set by the work the account needs.
What it rewards: keeping you, which requires results.
Not perfect. A lazy agency on a flat fee can coast, and you need reporting good enough to notice. But it removes the structural conflicts: advising you to cut spend costs nothing, improving efficiency costs nothing, and telling you a channel will not work costs nothing except the chance to keep billing for it.
The protection against coasting is not the fee model. It is month-to-month terms and a report with decisions in it, so underperformance is visible early and leaving is cheap.
Hybrids
A base retainer plus a bonus on an agreed outcome can work, provided the outcome is one that cannot be gamed — closed revenue rather than leads, net new customers rather than total conversions. Define it carefully or you have reinvented the performance model with extra steps.
What about equity or revenue share?
Occasionally proposed by agencies wanting upside on a growing business. It can work between people who genuinely trust each other, and it changes the relationship permanently — your agency becomes a stakeholder with opinions about pricing, product and strategy.
The practical problem is attribution. When revenue grows, how much was the advertising and how much was the product, the market or the sales team? That argument is unresolvable, and it arrives exactly when there is money on the table.
What to ask about agency fee models
Put one question to any agency you are considering: "Under your fee model, what is the recommendation that would cost you money to give me?" If they can answer it immediately and specifically, they have thought about it. If they say there isn't one, they have not.
Ours is a flat fee, month to month, never a percentage of spend — here is why.
The commission model has a variant worth its own look — paying per lead changes who owns the data as well as what gets optimised.
Questions people ask
Which agency fee model is best?
A flat retainer, for most businesses. It is the only common model where advice to spend less costs the agency nothing, which matters because that advice is the one you most need to hear and the one they are least rewarded for giving.
What is wrong with paying a percentage of ad spend?
It ties the agency’s income to your budget rather than your results. The structure outlives everybody’s good intentions: on the day the right call is to cut a channel, making that call costs them money.
Are performance-based fees better aligned?
They sound aligned and usually are not. Paying per outcome pushes an agency toward whatever converts fastest — branded search, remarketing, demand you already had — and toward claiming credit for sales that were coming anyway.
Should I offer equity or revenue share instead?
Rarely worth the complexity for either side. It makes a simple service relationship into a negotiation about valuation, and it does not solve the problem people hope it solves, which is usually a trust problem rather than a pricing one.