Marketing Agency Pricing Models: Retainer, Project, or Sprint

Three proposals are open on your desk and they refuse to line up. One quotes a monthly retainer. One quotes a flat project fee with a delivery date on it. One quotes a two-week sprint against a single deliverable. The scope descriptions rhyme, the promises are close enough, and the invoices are three different shapes with no obvious way to stack them next to each other. So you do the human thing: you find the number, pick the one that hurts least, and assume the shape is just paperwork.

The shape is not paperwork. Before anyone does an hour of work, the pricing model has already decided what your vendor is paid to care about. A retainer pays them to stay attached. A project pays them to ship and move on. A sprint pays them for one defined outcome and nothing else. You are not choosing a price. You are choosing an incentive, and it runs underneath every decision the vendor makes for as long as the engagement lasts.

What are the main marketing agency pricing models?

There are five in common use: retainers (a fixed monthly fee for ongoing work), fixed project fees (one price for a defined build with a delivery date), sprints (a flat fee for a single deliverable in a short window), hourly billing (you buy time instead of outcomes), and performance or commission (the vendor takes a cut of the results). Most agencies lead with the one that suits their cash flow, then reverse-engineer a reason it suits you.

The three that matter for most founders are the first three. Hourly billing rewards the vendor for being slow and punishes them for being fast, so it survives mainly in staff augmentation and legal work. Performance pricing sounds perfectly aligned until you notice it quietly pushes the vendor toward the easiest attributable wins and away from the patient work that never reports cleanly. Retainer, project, and sprint are where the real decision lives, and they are genuinely different instruments, not three labels for the same thing.

What each model actually pays the vendor to do

Read the invoice as an incentive and the differences get loud. A retainer rewards staying busy, because next month's fee depends on the relationship continuing. A project rewards shipping, because the vendor gets paid on delivery and wants to close it out. A sprint rewards one specific outcome, because the fee is bolted to a single deliverable and nothing gets paid for dragging it out.

None of those incentives is evil. Each one is honest about a different job. The trouble is the mismatch. Put an open-ended operation on a project fee and the vendor ships something plausible, then vanishes before it is tuned. Put a one-time build on a retainer and you have signed up to pay monthly, indefinitely, for a thing that was finished in week five. Even the sprint has a failure mode: pick it for work that genuinely needs a hand every week and you get one clean deliverable followed by silence, because the model was never built to stay. The retainer's specific failure has a name we have written about at length in the retainer trap: the incentive is to make the relationship load-bearing, so the knowledge stays in someone's head and the dashboards stay proprietary, and a well-run retainer and a well-designed dependency look identical from the inside until the day you try to leave.

A retainer pays a vendor to still be needed next month; a fixed price pays them to be done. Choose the one whose incentive you actually want steering your marketing.

Should I pay a retainer or per project?

Pay per project when the job is a defined outcome with an end: a website, a rebrand, a launch, a tracking rebuild. Pay a retainer only when the job is a real ongoing operation that needs a hand on it every week: paid media that has to be watched, a content engine that has to be fed, a pipeline that degrades without maintenance. If you cannot name what a retainer produces this month that is different from last month, you are renting attention, not buying work.

The rule of thumb is almost embarrassingly simple. Ask whether the work has a finish line.

• If the work ends the moment a thing exists (a site, an identity, a launch, an automation), it is a project or a sprint. You want a fixed price and a delivery date.
• If the work only ends when you stop paying (media buying, ongoing optimization, a growth program rolling out across new markets), it is a retainer. You want a named monthly output you can check off.
• If nobody can tell you which of those two it is, the scope was written to be unfalsifiable, and that is its own answer.

Marketing retainer vs one time project: match the model to the job

The correct model is a property of the job, not of the agency. A one-time change is a project or a sprint. A continuous operation is a retainer. The mismatch happens when a vendor sells you the shape that pays them best instead of the shape your work actually has.

A launch is the cleanest case for a fixed sprint: one window, one outcome, a hard finish line. When we ran the launch for Salt & Sun, the job was a defined push with a defined result, on the order of a 300% lift in launch-week engagement, not an open-ended arrangement that renews until someone remembers to cancel it. You do not put a launch on a retainer, because a retainer has no structural reason to ever call it finished.

Continuous growth is the opposite shape. Taking Magna Pest from four locations to eleven was not a single deliverable; it was an acquisition system that had to be watched, fed, and adjusted month after month as each new market came online. That is exactly what a retainer is for, and it passes the test: it produced something new every month, and you could name what that something was. Same firm on the vendor side, two completely different instruments, because they were two completely different jobs.

But a retainer feels safer

It feels safer because it feels like insurance: someone is always there, always watching, always one Slack message away. That instinct is worth saying out loud, because it is the exact instinct the trap is built on. A retainer with no nameable monthly output is the most expensive way to buy nothing. Four thousand dollars a month to "stay engaged" is forty-eight thousand a year for a standing call and a report, and the invisible part of that bill is the twelve months you spend feeling covered instead of getting the one build that would have actually moved the number.

The cost of choosing wrong is rarely a blowup. It is slow. You do not notice a mismatched model failing; you notice, a year later, that you paid continuously and own nothing you could hand to the next person. The fix is not to distrust retainers. It is to buy one only when the job is genuinely ongoing, and to insist that whatever it builds lands in accounts with your name on them, so the value compounds on your side of the table instead of the vendor's. That is the whole argument for owning your acquisition engine rather than renting one a month at a time.

How we price it, in the open

We sell all three shapes on purpose, and we try to talk you into the smallest one that fits.

1. Sprint: $5,000 flat, two weeks, one deliverable. For a defined outcome you can point at and a finish line you can hold us to.
2. Website from $8,000 (about four weeks) and Brand plus Site from $15,000 (about six weeks). Fixed-scope projects, owned outright, no monthly fee to keep them online. The full breakdown of why those numbers are what they are lives in what a website should cost.
3. Growth retainer at $2,500 a month and Infrastructure retainer at $5,000 a month. We recommend these only when there is a real ongoing operation to run, and every one of them produces a named monthly output you can check.
4. AI systems scoped to the build, because "custom" here means the work genuinely differs, not that the price is a mystery.

All of it sits on the pricing page with actual numbers, which in this industry still counts as unusual. If you are holding three proposals you cannot rank, that is the exact conversation to have out loud. Book a call, bring the quotes, and we will tell you which model the job actually needs, even when the honest answer is the smallest and cheapest one. It is the same discipline we lay out for vetting anyone you are about to hire in how to hire a marketing agency without getting burned: read the incentive first, then the price.