Marketing scope of work template for agency retainers.

A marketing scope of work states what lands in the client's hands each month, what it costs, and what happens when they ask for something that isn't on the list. Retainer work makes that harder than a one-off build, because the month resets and every unpriced favor quietly becomes next month's expectation. What follows is that wording, drafted against a six-month retainer at $9,400 a month for a company selling scheduling software to dental practices, so you can read the sentences instead of guessing at the brackets.

Four things decide whether the document survives past month three: the deliverable counts, who holds the ad accounts and the media budget, what the client owes you before anything can ship, and what exactly you're promising when you say the campaigns will work. The rest of a marketing scope is furniture. Nobody has ever argued about a confidentiality paragraph, and plenty of clients have argued that a fifth article was obviously included, that the ad budget came out of the fee, and that three weeks of waiting on their brand team shouldn't have moved the calendar.

What does a marketing scope of work need to say?

Eight sections cover the whole engagement, and four of them carry the weight. The purpose section says what the client is trying to sell and to whom, in language their own sales team would recognize. The channels section names where you're working, which for the dental scheduling client meant Google Ads, LinkedIn, the blog and the existing email list. The fee and term section states the number and the length. Then come the four that matter: the monthly deliverables with a count against every line, the media budget clause, the client dependencies with a consequence attached, and the pricing for anything that arrives outside the month, with the reporting section closing the document out.

If you've downloaded a free marketing scope template and it still felt thin, this is why. The usual ones give you objectives, timeline, deliverables, assumptions, success criteria and a signature block, which is the furniture, and then leave the four clauses that decide who pays for what as blank space you're expected to fill in yourself. None of them says a word about who holds the ad accounts.

How do you write monthly deliverables a client can't stretch?

Give every line a number, a size and a destination, so both sides can count the month on the last day and reach the same total. The failure mode in marketing retainers is a deliverable list made of activities. "Content creation" is an activity, and an activity has no edges, which is why it can mean four articles in your head and seven in the client's by March. Here is the monthly list from the $9,400 retainer, written the way it should read on the page.

1. Four articles of 1,200 to 1,800 words, drafted, edited and published to the client's CMS.
2. Eight LinkedIn posts, scheduled from the client's company page.
3. Two email campaigns to the existing list, built and sent in the client's platform.
4. Up to three active Google Ads campaigns and two LinkedIn campaigns, with up to six new ad creatives a month.
5. One written report on the second business day of the following month, plus a 45-minute review call.

Revisions need a number too, because creative work has no natural stopping point. Two rounds per asset, requested within five business days of delivery, with a third round billed at $150 an hour, is a normal shape and it holds up in practice. The five-day window matters as much as the two rounds. Feedback that arrives three weeks later on an ad that has already been running is a new piece of work, and a scope that stays quiet about timing invites exactly that.

Who owns the ad accounts, and who pays for the media?

The client owns the accounts and pays the platforms directly, and this is the clause that separates a marketing scope from every other kind. The dental scheduling retainer put it in three sentences: the client holds the Google Ads and LinkedIn Campaign Manager accounts in their own name and pays media costs directly to each platform on their own card; the agency fee of $9,400 a month covers management and creative, and it's a flat number rather than a percentage of whatever gets spent; that fee holds while the combined ad budget across both platforms stays at or below $30,000 a month, and every additional $15,000 of monthly budget adds $700 to it.

Two of those sentences save real money. Fronting media on the agency card turns you into a lender with a $30,000 exposure and thirty-day payment terms, and one slow client quarter is all it takes for that to become the agency's problem. Tying the fee to a percentage of spend does something worse, because it hands you a reason to recommend a bigger budget and the client a reason to wonder whether the recommendation was honest. A flat fee that only moves when the budget crosses a number both sides wrote down keeps that conversation clean.

Account ownership matters on the way out. When the accounts sit in the agency's name, ending the relationship means either a messy migration or a client who can't get their own historical data, and both of those turn a normal ending into a bad one. Set it up in the client's name on day one and give yourself admin access instead.

What does the client owe you, and by when?

Whatever the scope names, on the date the scope names, with a written consequence when it slips. Marketing retainers stall on the client's side more often than any project type, because almost nothing ships without something the client controls. The list from the dental retainer read: admin access to Google Ads, LinkedIn Campaign Manager, GA4 and the CMS within five business days of the start date; the brand asset library and any existing customer stories in the same window; and a single named approver who responds to draft assets within three business days.

The consequence sentence is the part most agencies leave out, and it's the one doing the work. The version we'd use reads: where an asset submitted for approval isn't approved or returned with written changes within three business days, it moves to the following month's calendar and the current month's deliverable count is met by the items already delivered. That isn't a penalty, and it shouldn't be presented as one. It exists so that a client who goes quiet for two weeks in August doesn't get to describe August as a month where the agency underdelivered.

Should you promise results or promise the work?

Promise the work and the reporting, and write down why. A marketing agency controls the campaigns, the creative and the targeting, and it doesn't control the price, the product, the sales team's follow-up speed or the six other things that decide whether a lead turns into revenue. Committing to a cost per lead in a signed scope means committing to variables sitting inside someone else's company.

Clients accept this far more easily than agencies expect, as long as the four measures are named and the report actually arrives. What clients won't accept is vagueness in both directions at once, where the scope promises growth without defining it and then reports on whatever looked good that month. Naming the measures up front is also how you get an honest renewal conversation instead of a defensive one.

You can promise the work and the reporting. You can't promise a number that the client's own sales team controls.

How do you price the work that falls outside the month?

Keep a short priced list in the scope itself, so the answer takes a minute instead of a negotiation. The dental retainer named five: a landing page built and published at $2,400, a fifth article in a month at $900, an extra set of three ad creatives at $650, a new automated email sequence at $1,800, and video production quoted separately because it never repeats. Underneath sat the rule that any of them starts only once the client has approved the price in writing and the month it lands in.

That last clause is the one agencies break, almost always out of goodwill. The landing page gets started on a Tuesday because the client asked nicely and it's only a day, the approval never comes back, and by the end of the quarter there are four of them and a conversation nobody wants to have. Every extra wants its own written request, price and approval from the person the scope names, kept in one place rather than scattered across a month of email. A change request form does that on paper, and having a price ready is most of what handling scope creep without losing the client comes down to. The alternative compounds quietly, which is what scope creep costs an agency over a year of small favors.

How is a retainer scope different from a project scope of work?

A project scope describes one finished thing, and a retainer scope describes a rate of delivery that has to hold for months, and that difference changes three clauses. The deliverables become monthly counts rather than a single list, so they need a rollover rule saying whether an undelivered item carries forward or expires, and expiring at month end with the reason written down is the version that keeps the calendar honest. The term needs its own sentence, in this case six months and then month to month with 30 days' written notice on either side. And acceptance stops being a single event, since there's no final handover to sign off.

The mechanics of a build are a different job entirely. If your agency does both, the website development scope of work and the software development scope of work cover the project side, where acceptance criteria and phase sign-off do the work that monthly counts do here. The general rules of writing one that clients can't misread apply to both.

What you're actually selling when you sell a retainer

Read the deliverable list back and ask what a client is buying with the $9,400. It isn't strategy, and it isn't a promise about demo requests. It's four articles, eight posts, two emails, five active campaigns, six creatives and a report on the second business day, delivered at that rate for six months, with a named price for everything else. A retainer that can't produce that sentence is being sold on trust alone, and trust is what runs out in month four when the client starts counting.

Docket keeps the record of every extra that got approved, at what price, by whom and on what date, which matters more on a retainer than on a build because the requests arrive weekly rather than twice. At renewal that record is what turns "we did a lot of extra work for you this year" into three landing pages, four additional articles, two creative sets and an email sequence, at $13,900 the client has already approved one line at a time.

Frequently asked questions