retainer or project.

agencies default to retainers because predictable revenue is a nicer business. clients default to projects because a fixed number is a nicer purchase order. neither default is right; the work decides. the test is simple: can you describe the finish line in a sentence?

it is worth being blunt about incentives here. an agency proposing a retainer for what is obviously a project is protecting its revenue, and a client insisting on a project for what is obviously ongoing work is protecting its budget approval process. both are understandable and both produce a bad arrangement, so it pays to name the mismatch early rather than discover it in month four.

get it wrong in either direction and the cost is real. a retainer bought for one-off work quietly overpays for standing capacity you are not using. a project bought for continuous work underpays for the first month and then stops, leaving you to re-brief a new team every quarter and pay the onboarding tax each time.

— 01the two models, side by side.

the difference is not price per hour. it is what you are buying — a standing capacity that keeps producing, or a defined deliverable with an end date.

that distinction decides who carries risk. in a fixed-price project, an overrun is the agency's problem, which is why project scopes are written carefully and change requests exist. in a retainer, the risk is shared and softer: some months you get more than you paid for, some months less, and the arrangement only works if both sides accept the averaging. clients who audit a retainer month by month against an hourly rate always conclude they are being overcharged, because that is not what a retainer is.

the second difference is memory. a retainer accumulates context — your approval habits, which formats your audience saves, which claims legal will not allow. that context is unbilled and it is most of why month six costs less to produce than month one. project work resets it every time, and the reset is invisible on the invoice.

— 02buy a project when.

the work has a natural end. a brand identity. a website. a launch campaign. a playbook. a product shoot. these have finish lines, and a fixed price protects you: if it takes longer than expected, that is the agency's risk, which is exactly the right place for it.

projects are also the sane way to test a new agency. one well-scoped project tells you more about how a team works than any number of credentials calls — and if it goes badly you are out one deliverable, not a year.

scope projects tightly enough to be finishable. "a website" is not a scope; "eight pages, two rounds of revisions, copy supplied by you, live by a named date" is. every fixed-price project that goes wrong goes wrong on the same thing — an undefined finish line, argued about after the money is spent.

— 03buy a retainer when.

the output is continuous and the value compounds — content every month, community answered every day, paid campaigns iterated weekly, seo built over quarters. these have no finish line, and trying to buy them as projects produces the worst of both: stop-start work, no institutional memory, and a re-onboarding tax every time.

the honest requirement for a retainer is a minimum term, usually three months, because month one is mostly foundation. an agency offering month-to-month from day one either front-loads nothing or is pricing the risk into your fee.

what a minimum term should not mean is a lock-in beyond it. three months to prove the foundation is fair; a twelve-month contract with no exit is a financing arrangement dressed as a partnership. ask for the notice period in writing and check who keeps the assets when it ends.

there is a middle model worth knowing about: a small retainer for continuity plus fixed-price projects for the big pieces. the retainer covers the work that must happen every month — calendar, community, reporting, iteration — and each launch, shoot or rebuild is quoted separately. it is more admin than a single monthly fee, and it is the arrangement least likely to leave you paying standing rates for one-off work.

— 04the trap in the middle.

the expensive pattern is a retainer used to deliver a queue of projects: a website in month one, a brand refresh in month two, a video in month three, all inside a monthly fee. it feels flexible and quietly costs more, because you are paying for standing capacity while consuming one-off work, and nothing compounds.

if your last three months with an agency were three unrelated deliverables, you wanted three projects. say so, and get three fixed prices.

the reverse trap is quieter: a retainer that has drifted into maintenance. same twelve posts, same formats, same report, no argument for two quarters. that is not a pricing model problem, it is a scope that stopped being reviewed — and the fix is a quarterly reset where both sides are allowed to propose stopping things. long sales cycles make this worse, which is why a b2b digital marketing agency retainer should be reviewed on pipeline quality rather than monthly output.

if this is the problem: retainer vs project, what a good content retainer includes.