How to write an ad creative brief

Why briefs fail before anyone opens a design file

The usual failure is not a missing field. It is an objective written as an activity. "Six social assets for the autumn launch" tells a designer what to produce and nothing about what the work has to do, so the first round comes back as a guess and the second as a compromise between two guesses.

The other failure is late change. Briefs get rewritten after they have been briefed in, usually because someone senior saw the first drafts and only then decided what they wanted. Every one of those changes costs a round.

• An objective that names a business outcome, not a deliverable: more trial signups from cold traffic, not six banners.
• One audience, described by what they currently believe or do, not by an age bracket.
• One named approver, so sign-off is not a committee vote.
• The live date and the budget, because both change what is worth making.

Write the proposition first

The single-minded proposition is the one thing a viewer should walk away believing. It exists to direct the work, not to appear in the ad, and a copywriter will rewrite the words anyway. If the sentence needs a comma to hold two ideas together, it is two propositions, and the ads will show it.

Most briefs fail this test on the first attempt. A quick check: if any competitor in the category could claim the same sentence without lying, you have written a category description.

• One sentence, one idea, no lists.
• Plain language, not brand vocabulary.
• Specific enough that a competitor could not run it unchanged.
• A reason to believe attached: a number, a guarantee, a demo, a review count.

Name the angle and the hook, leave the layout open

An angle is the argument the ad makes. A hook is what a viewer registers in the moment before deciding whether to keep scrolling. Naming both gives direction without designing the ad.

Over-prescription is where briefs quietly go wrong. If the brief specifies the photo, the headline and the colour, the review meeting turns into an argument about whether the designer followed instructions rather than whether the ad works. Describe the problem and the constraints, and leave the execution open.

• Two or three angles worth testing, each stated as a claim: price, speed, risk removal, social proof.
• One hook per angle: the image idea or line that carries the first half second.
• What is fixed: logo, product shot, brand colours, legal footer.
• What is off limits: claims legal will not clear, comparisons you cannot substantiate.

Put the formats and their constraints on the page

Formats belong in the brief because they change composition, not just the export settings. Meta's feed recommendation is 4:5 at 1080 by 1350, Stories and Reels are 9:16 at 1080 by 1920, and carousel cards are square at 1080 by 1080. A layout that only holds together in one ratio gets cropped by whichever placement spends the most.

Vertical placements also put interface on top of the image. Keep anything that has to be read out of roughly the top sixth and the bottom third of a 9:16 frame, where the profile row, the caption and the call to action button sit. Meta retired the 20 percent text rule back in 2020, so a heavy overlay will not block delivery, though the company still recommends keeping image text light.

• The exact ratios the campaign needs, in the order they matter.
• Where the logo and the offer sit in each ratio.
• Copy limits: primary text on the Meta feed truncates around 125 characters and headlines clip near 40.
• Any placement you are deliberately excluding, so nobody designs for it.

Brief a set, not a single ad

One ad is not a test. A brief should ask for a set: three to five distinct concepts, each rendered in the ratios the campaign will actually run. Distinct means different arguments, not one layout in four colourways, because near-identical variants cannot tell you which idea did the work.

Volume has a ceiling worth knowing. Meta's own guidance for a standard ad set is six or fewer active ads, since the marginal value of the seventh is small. Advantage+ shopping campaigns work differently and are built to take a much larger pool and sort it themselves.

Production is the step that used to make a set expensive. Once the brief is settled, Aduarius takes the written direction and returns each concept as square, vertical and horizontal creatives in one pass, so the review stays about the ideas instead of about resizing.

• Three to five concepts per brief, one argument each.
• Every concept in every ratio the campaign will run.
• One control: the current best performer, kept in the set as a baseline.
• A naming convention agreed before export, so reporting can group by concept.

Say in advance how you will judge it

A brief with no success metric hands the decision to whoever talks last in the review. Write the metric and the number on the page: the cost per acquisition to beat, or the hook rate a concept has to clear in its first week.

The same line settles how long a test runs. Deciding up front that each concept gets a fixed budget or conversion count prevents the familiar pattern of killing an ad on day two because the first few impressions came in expensive.

• The one metric that decides the winner.
• The number to beat, taken from the current best ad rather than an industry average.
• The spend or conversion count each concept gets before a decision.
• Who reads the result, and when.

Feed the result into the next brief

The last section of a brief is the first section of the next one. Record which angle won, which hook died and what the losing ads had in common, then attach that to the following round. Teams that skip this rediscover the same losing idea every quarter and call it a fresh test.

It is also the cheapest source of new angles. A comment thread under a winning ad, a support ticket, a review that phrases the benefit better than anyone in the building would have: those belong in the insight line of the next brief.

• One line per concept: what it argued, what it returned.
• The customer language collected since the last round.
• Anything the platform changed, such as a new placement, ratio or policy.
• The angles now retired, so nobody briefs them again by accident.