task-ops
The brief said the campaign should feel modern and premium, speak to ambitious professionals, and drive awareness. The designer produced three routes. All three were rejected, politely, with the note that they were not quite it. Nobody can say what it is, because the brief did not narrow anything. It described a mood and an audience, which together rule out almost nothing, and then asked for work to be judged against a standard that only existed in one person's head.
A creative brief is not a summary of the project. It is a set of constraints, and its quality is measured by how much work it makes impossible. A brief that could produce any of forty plausible executions has not briefed anybody. The fields below exist to remove options: one message, one audience, one primary format, one person who approves, and a written list of what must and must not appear.
Three failures account for the majority of rejected first rounds, and none of them are about writing style.
The brief describes the product rather than the change wanted. Pages of feature detail arrive with no statement of what a person should think, feel or do differently. Feature lists can be attached as background. They cannot substitute for an objective, because they do not rank anything.
The message is a list. When a brief carries five key messages, the work has to carry five, which means it carries none with force. Whoever wrote the list usually knows which one matters and did not want the argument with the colleague who owns the second one. That argument is cheaper before the work starts than after.
The approver is unnamed or plural. Creative reviewed by a committee converges on the safest option, then gets criticised for being safe. One named approver, with others consulted in writing beforehand, produces sharper work and shorter timelines. This is the field most often left vague, and the one that causes the most rework.
A fourth, quieter failure: the brief is written after the deadline has been promised. Constraints then arrive as surprises during production, and the brief turns into an argument about what was assumed.
| Field | What it must contain | Failure mode |
|---|---|---|
| Objective | The change wanted, stated as behaviour | Awareness, with nothing measurable behind it |
| Audience | One primary group, described by situation | Demographics that rule nothing out |
| Single message | One sentence, the thing to remember | A list of five, so none lands |
| Reason to believe | The evidence that makes the message credible | Adjectives instead of proof |
| Deliverables | Each asset, with format and dimensions | Named loosely, so extra versions appear later |
| Mandatories | Logo, legal lines, claims, disclaimers | Discovered at the final review |
| What to avoid | Executions already tried or ruled out | Omitted, so a rejected idea returns |
| Approver and route | One name, plus review rounds and dates | Committee review with no stated end |
| Deadline and budget | The real date and what it covers | A date without the review time in it |
Two of these fields are worth expanding, because they are where the rework hides.
The deliverables field should list every asset with its format and size, including the ones people forget: the animated version, the square crop, the email header, the version with no text for a partner to overlay. Those variants are usually assumed to be trivial. They are not, and when they surface at the end they are the reason a launch slips by a week.
What to avoid is the field most templates leave out. If a route was tried last year and rejected by the person who now approves, writing that down saves a round. It also protects the team from being asked, in a review, why they did not consider an idea that was in fact forbidden in a meeting they did not attend.
The example below is for a landing page and supporting ads. It is short on purpose.
Objective. Increase completed trial signups from paid traffic. Current completion rate on the existing page is the baseline, and the target is a measurable improvement within eight weeks of launch.
Audience. Operations managers at companies of twenty to two hundred people who currently track work in spreadsheets, are the person who gets asked for status, and have been told to fix reporting without adding headcount. They are not buying a philosophy. They are trying to stop chasing updates.
Single message. Status becomes something that can be read rather than collected.
Reason to believe. Three specific screens showing the same data as a board, a schedule and a report. One customer quote naming the weekly hours no longer spent collecting updates.
Deliverables. One landing page, desktop and mobile layouts. Three static ads at 1080 by 1080 and 1200 by 628. One fifteen second silent animation. One email header at 600 pixels wide. Copy for all of the above, in one document.
Mandatories. Product name in the first screen. Pricing statement as approved by the commercial team, exactly as written. Trial terms link in the footer.
What to avoid. Stock photography of people in meetings. The word effortless. The comparison chart approach used in the previous campaign, which the approver has ruled out.
Approver and route. One named approver. Two review rounds, feedback consolidated into a single document each time, comments returned within three working days.
Deadline. Live date fixed. Working back, that means final assets seven days before, which means first round two weeks before that.
Read that against the mood brief in the opening. It rules out a great deal, which is what makes it usable.
The fields overlap. The politics do not, and templates that ignore this produce documents that work in one setting and quietly fail in the other.
Agency work has a brief before the brief. The client brief arrives as a request, and it is usually incomplete: the objective is expressed as a deliverable, the audience is everyone, the deadline is already public. The agency's job is to write the internal brief that the creative team actually works from, then confirm the differences back to the client in writing. Skipping that step means the team works from a document the client wrote for themselves, which is how a project ends up with two unstated objectives.
Agency briefs also need commercial fields that in-house briefs do not: the number of rounds included, what counts as a new round, the rate for out of scope work, and who pays for a reshoot. These belong with the brief rather than only in the contract, because the person managing the rounds is rarely the person who negotiated them.
In-house briefs have the opposite problem. There is no contract, so there is nothing to point at when a stakeholder adds a requirement in week three. The protections that work are procedural: a named approver, a written list of who was consulted before the brief was signed, and a rule that new requirements become a separate request rather than a revision. In-house teams also face a specific failure that agencies do not, which is being briefed verbally in a corridor. A short written brief for a two day job is still worth the ten minutes, because the corridor version cannot be reread.
| Agency | In-house | |
|---|---|---|
| Source of the brief | Client request, rewritten internally | Requesting team, often verbally |
| What limits changes | Contracted rounds and rates | Procedure and a named approver |
| Biggest risk | Two objectives, one unstated | Requirements added without a route |
| Usual missing field | Who pays for the extra version | A written brief at all |
The most expensive part of a creative project is often a constraint that existed from the start and was not written down. Four kinds recur.
Channel specifications. Character limits, aspect ratios, file size ceilings, safe areas for text. These are knowable before design begins and are routinely discovered afterwards.
Legal and regulated claims. Any comparative claim, price, guarantee or statistic needs its source attached in the brief. A claim that cannot be sourced should be removed before the work is made around it.
Brand rules that are actually rules. Most brand guidelines contain a mixture of law and preference. Marking which is which prevents a review where a valid execution is rejected for breaking a rule nobody can defend.
Localisation and reuse. If the asset will be translated or handed to a partner, text length and layout decisions change from the first sketch. Retrofitting is worse than planning for it.
The practical fix is a short pre-brief pass where somebody collects the specifications for every named deliverable before the brief is signed. It takes under an hour and removes the most common cause of late rework.
It is also worth writing down which constraints are negotiable. A live date set by a trade show cannot move, while a date set by a monthly meeting usually can. A character limit imposed by a platform is fixed, while one imposed by an internal preference is not. When every constraint is presented as immovable, the team has no way to trade, and the thing that gets traded instead is the quality of the work or somebody's weekend.
A brief that lives as an attachment in a message thread stops being the reference within days, because feedback arrives in three places and the current version becomes whatever the last person said. The brief needs to sit with the work.
The arrangement that holds up is simple: the brief is one item on the project board, each deliverable is its own item with the format and date on it, and review comments go on the deliverable rather than into chat. Feedback then has a location, and a designer picking up the second round is reading a thread attached to the asset rather than reconstructing it from messages. Teams whose board, schedule and discussion sit in one place get this for nothing, and teams whose tools are split spend the difference in status meetings.
Two rules make review survivable. Feedback is consolidated before it is sent, by one person, into one document or one thread, with contradictions resolved first. And a round has a defined end, after which further requests are new work. Without those, three rounds becomes nine, and the deadline is defended by the team working weekends. A short setup path is enough to get the brief, the deliverables and the review threads sitting together on one board.
Take the current template and cut the key messages field down to one sentence, add a what to avoid field, and put a single name in the approver field. Then list every deliverable with its format and dimensions before the brief is signed, including the variants that always appear late. Keep the brief and the review comments attached to the work rather than in a message thread, because a brief nobody rereads is a mood board with headings, and Pinateca is free for up to five people and ten boards.
An objective stated as a change in behaviour, one primary audience described by situation rather than demographics, a single message in one sentence, the evidence that makes it credible, every deliverable with format and dimensions, mandatories such as legal lines and approved claims, executions to avoid, one named approver with the number of review rounds and dates, and the real deadline including review time. Anything that does not narrow the work can be attached as background instead.
One page for most work, two if the deliverable list is long. Length is not a virtue, and a brief that runs to six pages tends to be read selectively, which defeats the purpose. The test is whether a designer could rule out obviously wrong directions from it without asking a question.
A project brief covers scope, schedule, budget and responsibilities for delivering the work. A creative brief covers what the work has to say and to whom, plus the constraints it must respect. They are often combined for small jobs, but merging them on larger ones buries the message and the audience under logistics.
Whoever owns the outcome writes the first draft, and whoever will produce the work reviews it before it is signed, because they see the missing specifications and the assumed variants. In agency settings the client request should be rewritten as an internal brief and the differences confirmed back in writing, since working directly from a client's own document is how a project acquires a second unstated objective.
Name one approver, consolidate feedback into a single thread or document before it is sent, resolve contradictions before the designer sees them, and define when a round ends so that later requests become new work. Most revision spirals are caused by parallel feedback from several people rather than by weak creative, and the brief is where the route is set.