Campaign brief template

A campaign brief is one page, written before the campaign exists, that says what it is for and how you will know whether it worked. Eighteen fields, a sentence each.

It is not a plan and it is not a creative brief. It is the agreement the plan and the creative are both built against, and it is the document somebody reads in week three when they want to know why a line exists.

Write it before anything is built, and get it approved before anything is built. The Approver field holds one name and a date; if nobody will put their name to the brief, that is information about the campaign.

Use it again at the end. The Definition of done field was written before anybody knew the answer, which is what makes it worth reading afterwards. A campaign that is written up against its own brief teaches you something; one that is written up against whatever number looked best teaches you nothing.

The most common way a brief goes wrong is on its first two lines, where the business outcome and the primary metric collapse into one sentence. "Improve return on ad spend" gets written in the outcome field, and from that moment the campaign has no purpose that a person outside the marketing team can recognise.

Keep them apart deliberately. The outcome is what changes in the business: stock sold through, a launch that lands, a segment that starts buying again, a cost that comes down. Write it with no metric in it at all. The metric is then the instrument you have chosen to detect that change, and choosing it becomes a real decision rather than a reflex.

The discipline pays off at the end. A campaign that hit its metric and missed its outcome is a specific, useful finding — it usually means the metric was the wrong instrument — and you can only ever notice it if the two were written down separately at the start.

The baseline and target fields ship blank in the downloads, in that order, and the order is the point. A baseline is a reading: this metric, over this window, from this account. A target is a claim about where that reading can get to. Making the claim before taking the reading is how a campaign acquires a number that nobody can defend and everybody argues about.

Choosing the window is most of the work. Too short and you have measured a fortnight of one creative; too long and the offer, the landing page and the audience have all changed inside the sample. A reasonable rule is to cover at least one full buying cycle and to stop at the last material change to the offer, then to write both the window and the reason for it beside the number.

Where there genuinely is no baseline — a new channel, a new product, a first flight — leave both fields blank and say so in the definition of done instead: review at two weeks, set the target then. A brief that admits it is a first reading is more useful than one carrying a target somebody made up to fill the cell.

These three collapse into each other more often than any others on the sheet, and separating them is the fastest improvement most briefs can make.

The offer is the transaction: what the person gets and on what terms. Free returns, a bundle, early access, the product at its usual price. Naming it as an offer even when there is nothing special about it is worth doing, because it makes visible how much work the creative is being asked to do on its own.

The message is the single thing the ad must get across, in the customer's words. One sentence. If it needs a comma-spliced second clause, it is two messages and you are about to run a campaign that says neither of them clearly.

The proof is the reason to believe the message. A number, a guarantee, a review, a demonstration, a name. A message with no proof behind it is a claim, and it will get challenged — either by the creative team at the start, or by the market for the whole flight.

Most briefs describe what should happen and say nothing about what must not. That silence is why so many campaigns need their owner in the room for every small decision.

Guardrails are the inverse of the brief: the campaigns that are not to be touched, the size of a budget move that needs a second approval, the creative that is not to be reused, the audiences that belong to another team, the floor a bid may not go under. Written properly, they let somebody else act on this brief at four in the afternoon without asking, and they let you say afterwards whether what happened was inside the terms you set or outside them.

That is a useful property whoever is doing the acting. A person reads guardrails and works within them. A rule written from the same guardrails does the same thing, at the same boundary, and leaves the same record of which line it was acting under. The difference between an action you sanctioned and an action you merely did not prevent is written in this field, and nowhere else in the brief.

The last field is the one most briefs skip, and it is the one that decides whether the campaign teaches anybody anything.

A definition of done is a sentence you can answer yes or no to on the end date. It names the primary metric, the target, the baseline it is compared against, and the window. "The spring range is sold through, or the flight ends, and the result is written up against the August baseline" is a definition of done. "The campaign performs well" is not.

Its value comes entirely from being written before anybody knows the answer. Afterwards, every campaign has a story, and the story is always assembled from whichever numbers moved. A sentence fixed in advance is the only thing that can contradict that story, which makes it the only sentence in the brief that can change what you do next time.

The CSV version of this template is one row per brief, with the eighteen fields as columns, and it is worth keeping even if every brief also lives as its own document.

Read down a column rather than across a row and the register answers questions a single brief cannot. Have the targets been getting more or less ambitious? How often is the baseline field blank? Which guardrails keep getting written, which means they should probably be a standing rule instead of a per-campaign sentence? How many briefs from the last year were ever written up against their own definition of done?

Those answers are cheap once the briefs share a shape, and unavailable when every brief is a differently organised page. Add columns at the end if you need them; leave the eighteen names alone, so the register stays readable across a year.