Where the Untested Assumptions Live

Two deep wells and the shallow language between them

There's a structural gap on every client project and the brief sits directly on top of it.

The designer is fluent in branding, marketing and visual language, and usually knows very little about the client's specific market. The business owner knows the market in depth and usually can't describe what they want in visual terms. Two deep wells, and the language available for spanning them is thin and imprecise.

Figure 1 · Two wells, one thin bridge

So the brief gets written in the vocabulary that both sides can produce, which is adjectives. Modern. Trusted. Premium. Human. Not corporate. Those words let the meeting reach agreement, which is what they're for. They don't carry a shared picture, which is what everyone assumes they do.

Niche agencies close some of this gap because they've built market knowledge over dozens of projects in one industry. That knowledge still doesn't tell them what "premium" means to this client's particular audience.

The words that cause the most damage

Any adjective describing how the work should feel is a candidate. A few are reliably worse than the rest, because they sound specific and aren't.

Elevated. Reads as luxury to one person, as restraint to another, as simply better-executed to a third. We put the word to an audience and the split was visible.

Modern. Almost always means "current," and current is a moving target that different people locked in at different years.

Bold. Means high contrast to a designer and means brave positioning to an executive. Those are unrelated instructions.

Approachable. Usually a negation. It means not intimidating, and the thing being negated is different for everyone in the room.

Not corporate. Pure negation with no positive content at all, which means the brief has instructed the team to avoid something without indicating where to go.

Figure 2 · One word, several pictures

None of these are bad words. They're all fine as inputs. The problem is that they enter the brief as settled decisions when they're still open questions, and nothing downstream ever reopens them. It's the same gap that shows up when a client speaks a different language from the team: the words match, the meanings don't.

Where the bill arrives

The cost isn't paid at the brief stage. It's paid at every review after it.

Work comes back. Somebody says it doesn't feel premium enough. There's no way to resolve that, because "premium" was never defined past the word, so the conversation becomes a contest of taste and the most senior person's version wins. The team revises toward that version. Two rounds later somebody else says it lost the approachability, and the same thing happens in the other direction.

That's how a compromise direction gets built: not through a bad decision, but through a series of reasonable-sounding corrections toward different unstated pictures. Every round degrades the original signal a little more, and by launch the work is a merge of several people's private definitions of words nobody checked.

Then it goes live and the audience, who was never consulted about any of it, responds to whatever ended up on the page.

Resolving the adjectives before the brief closes

The move is to convert the contested words into visual evidence at the front of the project, while it's still cheap.

Do the business work first. The brief still needs its foundation, and none of what follows replaces it. The 21 questions about market, offering, audience, audience pain points, internal pain points and what matters to employees still get asked. That's where the understanding comes from.

Pull the adjectives out and list them. Every word in the draft brief that describes how the work should feel. There are usually between four and eight, and two or three of them are carrying most of the risk.

Build element boards against them. Not directions, not logos, not layouts. Design elements: line, form, texture, color, type. Two boards of twelve images is a common shape, sometimes three or four, crossed against each other to get a true sampling. A complete concept asks the audience to do the design work and produces a bad read, because they see the whole rather than the parts and how those parts could recombine.

Put it to the real audience. The people who buy, not the people who approve. Both, if the segments might disagree, which is worth knowing on its own. The method is audience perception mapping: the audience responds to the images instead of describing them, and the pattern shows where the group agrees, where it splits, and why.

Write the brief with the results in it. Now "premium" has a set of images attached to it, and the images came from the market rather than from the room. Results come back inside 48 hours, so this fits inside a normal kickoff rather than adding a phase.

Figure 3 · Resolving the brief, in order

The direction still emerges later, from the creative team's work. What changes is that the adjectives governing the work now mean something specific that everyone has seen.

What a resolved brief does to the review

The review conversation changes shape entirely, and this is where the time gets paid back.

"It doesn't feel premium enough" becomes a checkable claim. The brief contains what premium looked like to several hundred people in this market. Either the work is consistent with that or it isn't, and both parties can look at the same thing while deciding.

More usefully, it changes who has to win the argument. Without evidence, the designer is defending taste against seniority, which is a losing position however good the work is. With evidence, the designer is holding a documented read from the client's own audience, and anyone who wants to move away from it needs a reason strong enough to override it.

The creative team ends up with more standing, not less. They can walk into a room of stakeholders who don't understand the value of the work and show them something nobody can argue with. The ideas are still the team's. They've just been checked against the people the ideas are for.

Frequently asked questions