Marketing Agency Pricing Models: Retainer vs. Project vs. Performance

Most Marketing and Advertising Agencies default to one pricing model for every client, usually whichever one the owner learned first, which is almost always retainer. That is a mistake. The three marketing agency pricing models, retainer, project, and performance, each answer a different question about the engagement, and each carries a different risk profile for the agency delivering the work. Picking the wrong one for a given client is one of the fastest ways to quietly erode margin, because the symptom rarely shows up as a bad client. It shows up six months later as a thinning P&L nobody can explain, and by then the fix means renegotiating a relationship instead of structuring one correctly from the start.

The three marketing agency pricing models

Each model answers a different question. Retainer answers how much capacity are we reserving for you this month. Project answers what does this specific deliverable cost to build. Performance answers what is a unit of your result actually worth to you. A healthy agency book usually runs 60 to 80 percent retainer revenue, 10 to 25 percent project work, and, where performance pricing is offered at all, under 15 percent of total revenue. That cap is not arbitrary. Performance fees move with client results, and an agency that lets them climb past 15 percent of revenue is exposed to swings in cash flow it does not control. The mix across the whole book matters more than any single deal.

Retainer pricing: the default for a reason

The client pays a fixed monthly fee for an agreed scope of ongoing work, regardless of exactly how many hours it takes in a given month. For small and mid-sized clients, retainers typically run $2,500 to $8,000 a month. Mid-market accounts run $8,000 to $25,000. Enterprise accounts can run $25,000 to $75,000 or more, usually spanning multiple channels or a dedicated team. The number that actually matters is the effective hourly rate, monthly retainer divided by hours actually delivered. A $6,000 retainer against 50 delivered hours is $120 an hour. Let scope creep push that to 80 hours, often without anyone noticing month to month, and the same $6,000 fee is now paying $75 an hour, frequently below fully loaded cost once payroll, software, and overhead are allocated in. Retainer pricing rarely fails because the fee itself was set wrong at the start. It fails because nobody tracks hours against it every month, and when that revenue actually gets booked as earned is a separate question worth getting right too (see retainer revenue recognition for marketing agencies).

Project pricing: when scope has a defined edge

Project pricing fits work with a defined edge: a website rebuild, a rebrand, a campaign launch, a one-time audit, a CRM migration. Price it cost-plus. Estimate the hours, multiply by fully loaded hourly cost, and add a 40 to 60 percent margin on top. A rebuild estimated at 120 hours at a $65 fully loaded hourly cost is $7,800 in cost. At a 50 percent margin, the price quoted to the client is $11,700. The risk in project pricing sits almost entirely in the estimate. Run 20 percent over on hours with no change order in place, and that 50 percent margin can fall below 20 percent before the project even wraps. A locked scope document with a defined change-order process is not paperwork for its own sake. It is the only thing standing between the bid you quoted and the margin you actually keep once the client asks for "just one more round" of revisions.

Performance pricing: upside, and the real risk

Performance pricing pays on a result instead of a fee for time or a deliverable. The two common structures are a percentage of managed ad spend, typically 15 percent under $50,000 a month in spend, sliding down to 10 percent or less above $250,000 a month as the base grows, and a revenue share, usually 5 to 15 percent of incremental revenue measured against an agreed baseline. Hybrid structures are common and often the smarter move: a base retainer that covers cost of delivery, plus a performance bonus layered on top for upside, for example a $4,000 base plus 10 percent of revenue above an agreed benchmark. The real risk in performance pricing is not the client walking away, it is the agency's own cash flow. A bad month for the client's results is a bad month for agency revenue, and payroll is due regardless of how the campaign performed. Performance fees should sit on top of a base that already covers delivery cost. They should never be the thing standing between the agency and meeting its own payroll.

Which model fits which client

The decision is mostly about scope certainty and who controls attribution, not about what feels more sophisticated to offer. An open-ended, ongoing scope, like always-on social management or SEO, points to retainer. A defined deliverable with a clear start and end, like a website build or a brand refresh, points to project. A metric the agency can measure and attribute cleanly, usually paid media spend or direct-response revenue rather than something fuzzier like brand lift, points toward performance, and only with a client who will not dispute the attribution once results come in either direction. Most agencies running healthy margins are not committed to one model across the whole book. They are blending a retainer base that covers fixed delivery cost for every client, a project rate for anything outside that scope, and a performance kicker layered on top for the specific accounts where results are clean enough to measure. A fractional CFO for marketing agencies can model what that mix should look like client by client, and flag it the moment a given account's effective rate drifts below target. For the fuller playbook on moving margin once pricing is set correctly, see how to increase agency profit margins.

Revisit the model as the client changes

The model matters less over time than the discipline of revisiting it. A retainer priced correctly eighteen months ago, against half the client's current ad spend and half the current scope, is not correctly priced today, even though nothing about the contract has technically changed. The same holds in the other direction: a performance deal signed when a client's baseline revenue was small can turn into an underpriced retainer-equivalent once that baseline has tripled. Review pricing model and pricing level together, on a set schedule, instead of waiting for a renewal conversation to force the question.