Which creative brief sections are decisions and which are just description

Somebody in the Monday review notices that the market context section of the 2027 brief is identical to the 2026 one, word for word, down to a comma splice. The room shrugs, correctly. Market context describes a market that has not changed much in a year, and rewriting it from scratch would have introduced errors rather than removed them.

Then the same person scrolls down and finds the audience section is also identical. Nobody shrugs at that one, and the difference between the two reactions is the most useful thing you can know about creative brief sections.

Two kinds of section, and why the difference pays

A brief section either records something that is already true or records a choice somebody made. Call the first a description and the second a decision.

Descriptions are cheap to produce, cheap to check, and safe to inherit. Venue, hall, dates, the mandatory legal line about the trade association, where the logo files live, what last edition did on registrations and net square metres. All of it exists somewhere else already, and the brief copies it so the reader does not have to go looking.

Decisions have none of those properties. They cost an argument to produce, they can only be checked by somebody with authority, and inheriting one is how last year quietly becomes this year. There are not many of them.

Parnas and Clements made the general version of this argument in IEEE Transactions on Software Engineering in 1986, in a paper about design documentation. Their point was that the ideal rational process never happens, so the discipline lies in producing the document that process would have produced, which means writing down the decisions and the reasons even when they were reached messily. A brief is that document for an edition. The description sections are the furniture. The decision sections are the record.

What makes a section a description?

Apply one test. Could a competent person outside the room fill this section correctly by looking something up?

Venue: yes, it is on the contract. Dates: yes. Prior performance: yes, it is in the post show report. Brand assets: yes, they are in the library. Budget envelope: yes, finance set it. Mandatories: yes, legal wrote them and they change when legal says so.

Now try the same test on positioning. Nobody outside the room can look up what this edition is for, because it has not happened yet and the answer depends on a judgement about which part of the market is worth chasing. Same for audience, where a lookup gives you who came last time and the brief has to say who you are aiming at this time. Same for the single success measure, for tone where the brand allows more than one register, and for non goals, which by definition record something the team chose to skip.

That gives a clean split on most templates. Eight or so descriptions, four or five decisions. The exact boundary shifts by organisation. Deliverables can be either, depending on whether your production plan is fixed by contract or open every year. Timings are usually a description if the edition date is set and a decision if the campaign start is genuinely up for debate.

The sections that cannot be inherited

Positioning, audience, the one success measure, non goals, and tone where the brand system permits a choice. Those are the ones to guard, and the guarding is mechanical: they arrive empty, every cycle, and the form will not accept an approval until each has been touched by a named person this cycle.

Every other section can arrive prefilled, which is a feature. Retyping the venue address from memory is how a brief ends up with the wrong hall in it.

The failure mode is a system that treats all sections identically. Copy everything and the decisions come through prefilled, which nobody edits. Copy nothing and the team spends a day retyping the legal line, gets it slightly wrong, and starts resenting the brief. The mechanics of doing that copy honestly, including how to mark what was inherited so a reviewer can see it, is the roll over problem and it has its own set of traps.

Tagging a live template at a food ingredients show

Take a 12 section template and tag each section D for decision or X for description, then measure how much of each bucket arrived already written.

Background: X. Market context: X. Objectives: D. Audience: D. Positioning: D. Proposition: D. Tone: D. Mandatories: X. Deliverables: X. Budget: X. Timings: X. Approvals: X.

Seven descriptions, five decisions. Now compare each section against the same section in the 2026 brief and count the ones that are unchanged.

All seven descriptions unchanged, which is correct and saves the team perhaps two hours of retyping. Of the five decisions, three are unchanged: objectives, tone and proposition. Audience has one word different, "bakery" swapped for "bakery and snacks". Positioning has been rewritten.

So the honest score is one decision section out of five genuinely made this cycle, with a second half made. The brief reads as complete, it reads as consistent with strategy, and it commits the 2027 edition to a proposition that was written in a room in early 2025 for a market that has since had two large ingredient price shocks.

Work the two ratios separately and the divergence is obvious. Twelve of twelve sections filled is 100 per cent complete. One and a half of five decisions made is 30 per cent. The first number goes into the status report because it is the one the tool produces. The second is the one that predicts whether the campaign gets rebriefed in June.

Note also what the tagging exercise costs. Twelve sections, one letter each, perhaps six minutes for somebody who knows the show. Comparing each section against last year's text takes another fifteen. For half an hour you get a number that tells you how much of the 2027 brief is actually about 2027, and the case for turning that into a repeatable score sits with the five item quality rubric.

Why does a filled brief still read as thin to the people receiving it?

The BetterBriefs Project surveyed over 1,700 marketers and agency staff across more than 70 countries and published with the IPA in 2021. It found 78 per cent of marketers believed their briefs provided clear strategic direction, against 5 per cent of creative agencies. On clear and concise language the split was 83 per cent against 7 per cent.

A gap that wide is not two groups disagreeing about quality. It is two groups measuring different things. The writer of a brief experiences it as a form with twelve boxes, and a filled box feels like progress because filling it took effort. The receiver experiences it as a set of constraints, and an inherited decision imposes no constraint at all, because it was written for last year's problem and everybody knows it.

There is a practical consequence for review. If you read a brief section by section in order, you will spend most of your attention on the descriptions, because they come first and they are easy to check. Read the five decision sections first, in isolation, with the rest of the document closed. A brief that survives that reading is worth circulating. Most do not survive it, and knowing which five sections to open is what makes the template worth having at all.

Writing the descriptions once and then leaving them alone

The best outcome of tagging is that the description sections stop being written by hand.

Venue, dates, hall, capacity, prior registrations, prior exhibitor count, budget envelope and the mandatory lines all exist in a system somewhere. Pulling them into the brief from that system, with the source named beside the value, does two things worth having. It removes the retyping errors, and it makes the description sections visibly not the team's work, which frees the team's attention for the five sections that are.

This is the part of brief production that generalises across a portfolio, because the descriptions are structurally the same for every show even when the decisions are wildly different. A group running eight editions a year is producing roughly 96 description sections annually, most of which are lookups. A brief and review workspace that holds them centrally, of the kind covered on the publishing engine page, takes that work off the show teams entirely, and changes nothing about whether the decisions get made.

Where the split breaks down

Two places, and both are worth knowing before you build a process on this.

The first is a section that looks like a description and is quietly a decision. Deliverables is the usual one. On a show where the production list has been the same for six years it is a lookup, and on a show where somebody is deciding whether to drop the printed guide this year it is the most consequential decision in the document, sitting in a section nobody reads carefully because it has always been furniture. The tag has to be reviewed each cycle, which is a small annoyance and a real one.

The second is that inheriting a description is only safe when the underlying fact has not changed. Hall numbers change. Legal lines change. A description copied forward is a claim that nothing moved, and unlike a decision, nobody remembers to check it, because the section has never once been wrong before.

This week, open your current template and put a D or an X next to each section heading. Then look at the D sections in the most recent brief you signed and check how many words differ from the edition before. Anything at zero is a decision you did not make.