Agencies bill in three main ways — a monthly retainer, a fixed project fee, or a performance-based fee — and each puts the risk and the incentives in a different place. The retainer buys ongoing capacity, the project buys a defined deliverable, and performance ties the fee to results. Whichever model you choose, the clause that matters most is the same: the one that separates the agency's management fee from your ad spend, so you always know how much you are paying the agency versus how much reaches the platforms. For performance deals, a second clause matters just as much — a fair commission clause that defines KPIs and attribution before the results are counted.
Operator guidance from an agency that bills on these models — not legal advice; check specifics with a lawyer for your situation.
TL;DR
- Three models: retainer (ongoing capacity), project (defined deliverable), performance (fee tied to results); plus hybrids.
- Each puts risk somewhere different — the retainer on the client, the project on scope, performance on the agency (which prices for it).
- Always separate the management fee from ad spend. Ad spend is your cost paid to the platform, never the agency's revenue.
- A percentage-of-spend fee (often ~10–25%) creates an incentive to overspend unless it is capped, tiered or paired with accountability.
- A fair performance clause defines KPIs and attribution up front, exports a signed baseline, and excludes factors the agency doesn't control.
- Ownership and handover clauses apply regardless of billing model — you own the accounts and data.
- Match the model to the work and the risk you want to hold, not to whichever number looks smallest.
The three billing models
| Model | How it works | Best when | Who bears the risk | Failure mode |
|---|---|---|---|---|
| Retainer | Fixed monthly fee for ongoing scope | Continuous work, evolving needs | Client (pays regardless of month-to-month output) | Scope creep or under-delivery |
| Project | Fixed fee for a defined deliverable | One-off, well-defined work | Agency (fixed price, variable effort) | Ongoing needs billed piecemeal |
| Performance | Fee tied to results (leads, revenue, % of spend) | Measurable outcomes, aligned goals | Agency (paid on results) — priced accordingly | Attribution disputes; incentive misalignment |
Most agency relationships are retainers, because marketing is continuous and a retainer buys reliable capacity. Projects fit defined, bounded work (a website, a launch, an audit). Performance models sound the most aligned — the agency only wins when you win — but they carry the most measurement complexity, because "results" must be defined precisely enough to bill against, which is where disputes start.
Hybrids are common and often the fairest: a base retainer that covers capacity plus a performance component that shares upside. The base keeps the agency stable enough to do good work; the performance element aligns incentives without making the whole fee hostage to attribution.
The clause that matters most: fee vs ad spend
Whatever the model, the non-negotiable clause is the separation of the management fee from ad spend. The management fee is the agency's revenue; ad spend is your cost, paid to the platform. A contract that blends them into one "media budget" number hides how much you pay the agency versus how much reaches Google or Meta, and it makes the relationship impossible to evaluate.
The contract should:
- Show the management fee and ad spend as separate line items.
- Say who holds the payment method — ideally you pay the platforms directly on your own billing profile, so the spend and its history stay yours (the ownership reasoning is in who owns your Google Ads, Meta and GA4).
- Never book ad spend as agency revenue — money paid to a platform is your cost.
The percentage-of-spend trap
A common performance-adjacent model is a management fee set as a percentage of ad spend, often in the region of 10–25%. It is simple and scales with the account, but it carries a built-in conflict: the agency earns more by spending more, whether or not the extra spend is efficient. Left unchecked, it rewards inflating the budget rather than improving the return.
It is not automatically wrong — it is a reasonable way to price management effort that scales with account size — but it needs guardrails:
- A cap or a tiered rate, so the percentage falls as spend rises and the incentive to inflate weakens.
- Pairing with performance accountability, so the agency is judged on results (CAC, ROAS, pipeline), not just on spend managed.
- Transparency on who controls the budget decision — the client sets the budget, the agency recommends, so "spend more" is a recommendation to approve, not a lever the agency pulls alone.
The fair performance clause
A performance clause needs a measurement schedule, not a slogan. It should define the eligible outcome, source system, baseline, attribution or incrementality method, validation window, refunds and cancellations, new-versus-existing customer treatment, currency, tax, and dispute process. The agency should not control the only system that determines its fee.
Hybrid pricing is often more stable: a base fee pays for minimum team capacity and governance, while a variable component rewards a business outcome inside the agency's influence. Caps and floors protect both parties when demand, inventory, pricing, sales follow-up, or tracking changes.
Worked example: a bonus based on qualified pipeline should define qualification fields in the CRM, exclude duplicates and existing opportunities, wait through the agreed validation period, and reconcile reversed or rejected records. “10% of revenue generated” without those rules invites conflict rather than alignment.
Performance-based fees live or die on how "performance" is defined, and a vague definition guarantees a dispute. A fair performance clause pins down the ambiguity before any money is at stake:
- Define the KPIs precisely — what counts as a lead, a qualified lead, a conversion, revenue — with the exact events and thresholds.
- Agree the attribution methodology up front — which model, which window, which platform is the source of truth — because attribution is where "results" are contested (the reasons are in marketing attribution for executives).
- Export a signed baseline — a snapshot of performance at the start, attached to the contract, so improvement is measured against an agreed starting point, not a disputed memory.
- Exclude external factors the agency doesn't control — seasonality, a viral moment, a pricing change, a PR event, a platform outage — so neither side is rewarded or penalised for things outside the work.
- Say what happens if tracking breaks — a fallback measure or a pause, so a broken pixel does not become a billing argument.
The performance clause is where a contract earns trust. An agency willing to define KPIs, fix the attribution method and sign a baseline is confident in its work; one that keeps "results" vague is leaving itself room to claim credit it did not earn.
Scope, term and ownership apply to every model
Three clauses are model-independent and matter regardless of how you are billed:
- Scope and scope-creep protection — a clear SOW defining what is included and how extra work is quoted, so a retainer does not quietly expand or a project does not get nickel-and-dimed.
- Minimum term, notice and auto-renewal — a fair notice (around 30 days) and a transparent renewal, not a long lock-in with a narrow cancellation window.
- Account and data ownership + handover — you own the accounts, data and creative regardless of billing model, with a defined offboarding SLA.
These are covered in full in the marketing agency contract checklist; the billing model changes how you pay, not who owns what or how you leave.
Glossary
- Retainer — a fixed recurring fee for ongoing scope and capacity.
- Project fee — a fixed fee for a defined, bounded deliverable.
- Performance fee — a fee tied to results (leads, revenue, or a share of outcomes).
- Percentage of ad spend — a management fee set as a share of media budget.
- Management fee — the agency's revenue, separate from ad spend.
- Baseline — a signed snapshot of starting performance, against which improvement is measured.
- Attribution methodology — the agreed rules for crediting results, fixed before billing.
How to choose the model
The right model depends on the work and the risk you want to hold:
- Continuous, evolving marketing → retainer (or retainer + performance hybrid).
- A defined, one-off build → project fee.
- A measurable, isolatable outcome with clean attribution → performance, with a fair clause.
- Uncertain or new relationship → a base retainer with a modest performance component, so both sides share risk while trust is established.
The wrong way to choose is by the headline number. A cheap retainer with vague deliverables can cost more than a higher one with clear scope; a pure performance deal can be expensive if the agency prices heavily for the risk it carries. Judge the model on how it aligns incentives and where it puts the risk, not on which invoice looks smallest.
How Space Ads approaches billing
We bill across these models, and our view is that the model matters less than two clauses: the fee/ad-spend separation and, where performance is involved, a fair commission clause. The failure we see in accounts we inherit is a blended number where the client never knew how much was fee and how much was media, or a performance deal with undefined KPIs that became an argument the moment results were counted.
Our approach is to keep the management fee and ad spend as separate line items with the client paying platforms directly where possible; to cap or tier any percentage-of-spend fee and pair it with real accountability; and, for performance components, to define KPIs and attribution up front, sign a baseline and exclude external factors. That is the billing layer of performance marketing; the deeper budget question — how much to spend in the first place — is in what % of revenue to spend on marketing, and when the need is senior ownership of the whole commercial relationship, a fractional CMO fits.
Stop doing / Do instead
| Stop doing | Do instead |
|---|---|
| Blending fee and ad spend into one number | Separate them as line items |
| Uncapped percentage-of-spend fee | Cap or tier it, pair with accountability |
| Vague "results" in a performance deal | Define KPIs, attribution and a signed baseline |
| Choosing the model by headline price | Choose by incentive alignment and risk |
| Ignoring external factors in performance | Exclude seasonality, PR, pricing, outages |
| Treating billing and ownership together | Own accounts and data regardless of model |
FAQ
What is the difference between a retainer, project and performance fee?
A retainer is a fixed recurring fee for ongoing capacity and scope; a project fee is a fixed price for a defined, bounded deliverable; a performance fee is tied to results such as leads, revenue or a share of outcomes. Each puts the risk in a different place — the retainer on the client, the project on the agency's effort, performance on the agency's results — and hybrids combining a base retainer with a performance element are common.
How do marketing agencies charge for ad spend?
Ad spend should be separate from the agency's fee: it is your cost, paid to the platform (Google, Meta), never the agency's revenue. A contract should show the management fee and ad spend as distinct line items, and ideally you pay the platforms directly on your own billing profile so the spend and its history stay yours.
What percentage of ad spend do agencies charge?
Management fees set as a percentage of ad spend often fall in the region of 10–25%, though it varies with account size and complexity. The percentage model carries a conflict — the agency earns more by spending more — so it should be capped or tiered and paired with performance accountability, and the client, not the agency, should control the budget decision.
How do you structure a fair performance clause?
Define the KPIs precisely (what counts as a lead, conversion or revenue), agree the attribution methodology and window up front, export and sign a baseline of starting performance, exclude external factors the agency doesn't control (seasonality, PR, pricing, outages), and specify what happens if tracking breaks. Vague "results" guarantee a dispute; a defined clause builds trust.
Should ad spend be included in the retainer or separate?
Separate. Blending ad spend into the retainer hides how much you pay the agency versus how much reaches the platforms and makes the relationship impossible to evaluate. Keep the management fee and ad spend as separate line items, and ideally pay the platforms directly so you retain the spend and its history.
Which billing model is best for a marketing agency relationship?
It depends on the work and the risk you want to hold: a retainer for continuous, evolving marketing; a project fee for a defined one-off build; performance for a measurable, cleanly attributable outcome; and a base-retainer-plus-performance hybrid for a new relationship where both sides share risk. Choose by incentive alignment and where the risk sits, not by the smallest headline number.
Key takeaways
- Agencies bill on retainer, project or performance models, plus hybrids; each places risk differently.
- The key clause in every model separates the management fee from ad spend — spend is your cost, not agency revenue.
- Percentage-of-spend fees (often 10–25%) need a cap or tier and performance accountability.
- A fair performance clause defines KPIs and attribution up front, signs a baseline, and excludes external factors.
- Scope, term and ownership clauses apply regardless of model; choose the model by incentive alignment.
Sources and further reading
- Google Ads Help — About manager accounts and billing
- IAB / industry — Agency remuneration and fee models
- ANA — Agency compensation and transparency guidance
Continue learning
- What should be in a marketing agency contract? A founder's checklist
- Who owns your Google Ads, Meta and GA4 when you leave an agency?
- What % of revenue should you spend on marketing?
- Marketing attribution for executives
- How much does a fractional CMO cost?
- Performance marketing, billed transparently
Continue reading
What Should Be in a Marketing Agency Contract? A Founder's Checklist
A marketing agency contract should make three things unambiguous: who owns the accounts and data, how money and ad spend are separated, and how you exit with everything you built. This clause-by-clause checklist covers what a fair contract says — from an agency that signs and honours them.
Your First Senior Marketing Hire: CMO, VP, or Fractional?
When founder-led marketing hits its ceiling, the instinct is to hire a CMO. Often that is too senior, too early, and too expensive — a VP who still executes, or a fractional lead for direction plus specialist hands, fits better until the company is large enough for a full CMO to have leverage.
Who Owns Your Google Ads, Meta, and GA4 When You Leave an Agency?
You should own your ad accounts and data; the agency is a revocable manager, not the owner. But access is not ownership — the billing profile and account structure decide. This is the platform-by-platform mechanics of what you keep, what you can lose, and how to check before it's too late.
