How to Write a Design Rationale That Survives a Client Review

A rationale is the document that travels with creative work — the writing that explains why the thing looks the way it does. It exists because creative work sent without it gets reviewed as taste, and taste is an argument nobody can win.

The failure is not that people write bad rationales. It is that they write appreciations: paragraphs explaining why the work is good. A client who receives an argument for beauty will answer with their own opinion about beauty, and you have just agreed to have the conversation on the only ground where your expertise counts for nothing.

Convert taste into constraints

The whole job of a rationale is to move every decision off the axis of preference and onto the axis of agreement. Each decision should trace back to something already established: the brief, the audience, a measurement, a technical limit, a legal requirement, an existing brand system.

Taste: “We chose a clean, modern sans-serif that feels confident and approachable.”

Constraint: “The brief said most of your readers arrive on a phone, so the type is set larger than a print equivalent, and the family was chosen because it stays legible at that size and covers the accented characters your German market needs.”

The second version is not more defensive. It is harder to argue with, and — more usefully — it tells the client what a valid objection would look like. If they disagree, they now have to disagree with the phone-first assumption, which is a productive conversation instead of a circular one.

The five sections

The problem and the constraints you were given. Restated in your own words, at the top, before any work appears. This is the paragraph the client silently agrees with, and every later decision is anchored to it.

The decisions. Each in the same three-part shape: what we did, what it is for, and what we rejected and why. Four or five decisions, not twenty.

What we are unsure about and want your view on. Named specifically. This is where you want feedback, and saying so is what stops feedback landing everywhere else.

What is not up for debate, and why. Accessibility minimums, legal requirements, the existing brand system, anything already signed off. Short and unapologetic. A client who does not know which parts are fixed will assume everything is negotiable.

What happens with feedback. The form you want it in, the deadline, who consolidates it, and how many rounds remain. Round counts belong in the scope document, but restating the number here, at the moment somebody is about to spend one, is what makes it real.

Show the option you rejected

A design presented alone looks arbitrary. The same design presented next to the obvious alternative, with one sentence about why it lost, looks decided.

This is not the same as presenting three options and asking them to choose — that is abdication, and it invites a committee. Show the direction you are recommending, and show the credible alternative as evidence that you considered it and can say what it costs. The client’s instinct to suggest the obvious thing has already been answered before they can voice it.

Ask for the feedback you want

“What do you think?” is an invitation to taste. Replace it with questions that have a subject: “Does the headline read the way you would say it out loud?” — “Is anything here going to be a problem for your compliance team?” — “Which of these two do you want to test first?”

Specific questions get specific answers and, more importantly, they define the surface. A client asked five pointed questions is far less likely to return a paragraph about the colour green.

What to cut

Adjectives about the work: clean, modern, bold, elegant, striking. Every one of them is an opinion wearing a description’s clothes.

Mood-board vocabulary. “Warmth”, “energy”, “a sense of craft” — if you cannot trace it to a constraint, it does not belong in a rationale.

The word “just”. “We just felt” gives away that there was no reason.

A defence of anything nobody has objected to. Rationale written pre-emptively for every element reads as anxiety and buries the four decisions that matter.

Process narrative. How many directions you explored is about your effort, not their decision.

What the reading pattern tells you before the meeting

A rationale is unusual in that it is normally read shortly before a conversation, which makes the timing of the read genuinely informative about the meeting you are walking into.

Read at length the day before the review: you are meeting a client who has formed specific views. Bring the rejected option and expect detail.

Opened for a few seconds an hour before: the review is the review. Plan to present it rather than to discuss it, and do not assume any section landed.

Read twice, days apart: it is being weighed, or shown to somebody else. Either way there is an internal conversation happening that you are not in.

Two honesty notes, because this is a document where over-reading the data is tempting. You cannot tell which section somebody lingered on — there is no per-section map, only duration for the document as a whole. And a duration figure is only meaningful if it was measured while the document was actually on screen; a number derived from how long a tab was open will tell you somebody spent an hour on a rationale they left open during lunch. Why the naive method can only exaggerate is worth ten minutes before you rely on any reading time. What is and is not measured is listed with its limits on the features page.