Running design critiques that sharpen instead of flatten

Every studio has a critique. Most have a bad one — the kind where a senior designer squints at a screen, says "it feels busy", and a junior designer spends the weekend flattening perfectly good work. Critique is the highest-leverage hour in a design practice, and it is routinely run with less structure than a standup.

I've run the Thursday critique at Brassfern for six years. Here's the whole method — the roles, the timing, the rules, and the failure modes we designed them against. Steal it; rituals should travel.

What critique is for (and what it isn't)

A critique has exactly one job: increase the quality of decisions available to the person presenting. It is not a design-by-committee session, not an approval gate, and not a place for seniors to restate their taste with the room's authority behind it. The presenter leaves with more clarity than they arrived with, or the hour failed — regardless of whether any particular screen got "better".

That framing matters because it moves the burden. The presenter owns the outcome. The room owns the thinking.

The setup: roles before opinions

Three roles, named every week, rotating across the team:

The presenter brings work in doubt, not work done. Two or three problems maximum, each introduced with: the constraint, the decision they've narrowed to, and the specific question they want answered. "Does this hierarchy survive a 375px screen?" is critiquable. "Thoughts?" is an invitation to a personality contest.

The facilitator keeps time, enforces the rules, and — critically — is not the most senior person in the room most weeks. Rotating facilitation is how a studio teaches judgment instead of deference.

The room (four to seven people, never more) follows one rule above all: be specific or be silent. More on which below.

Non-presenters read the brief materials in advance. A critique where half the room sees the work for the first time in the session is a critique that spends forty minutes on context and ten on craft.

The timing: fifty minutes, hard stops

0:00–0:05 — Context. The presenter states the problem, constraints, and their questions. No defending, no apologising, no "this is really rough so…" — rough is the point.
0:05–0:15 — Silent looking. The room examines the work without the presenter's voice in their ear, writing observations on a shared board. This one change — silence before speech — killed our loudest-voice-wins problem overnight.
0:15–0:35 — Round the room. Each person gives their single strongest observation, specific, tied to the presenter's stated question. Facilitator parks tangents visibly ("good thread; not today's question").
0:35–0:45 — Open discussion, presenter leads follow-ups. This is where screens get argued over properly, with the permission structure established.
0:45–0:50 — The presenter summarises what they're taking away and one thing they're deliberately not acting on. Naming the ignored feedback out loud matters: it confirms the presenter owns the decision and the room owns only the thinking.

The rules that do the work

1. Specific or silent. "Make it pop" is banned. So is "I don't love it." Every observation must name the element, the behaviour, and the effect on the user or the goal: "The tertiary button at 13px competes with the price at mobile widths; I read the button first and the price second, which is backwards." If you can't be specific, you're reacting to taste, and taste reactions get parked.

2. Critique outcomes, not authors. The sentence shape is "the screen does X, which means the user sees Y" — never "you did X". It sounds minor. It changes the whole emotional weather of the room, and it's the rule that makes juniors willing to present, which is the entire ballgame.

3. Questions before corrections. "What happens on a 375px screen?" teaches. "You need a mobile breakpoint" instructs — once, shallowly. Facilitators reroute corrections into questions for the first few months; after that the habit sticks.

4. The ladder of conviction. Feedback gets tagged as it lands: blocker (this breaks a user goal), consideration (worth testing or weighing), or stirring (an offering, take or leave). Untagged feedback all arrives at the same volume, and presenters treat "maybe try a serif" with the same weight as "the primary action is invisible to colour-blind users". The ladder fixes the mix.

5. Disagreement is data. When two seniors disagree, the facilitator doesn't arbitrate — the disagreement itself is the finding. Two defensible directions mean the presenter has a genuine design fork, and the right move is usually a tiebreaker test, not a louder debate.

The failure modes these rules exist against

The flattening. A talented designer's strange, alive homepage enters critique; it leaves beige. Six rounds of "what if we just…" sanded every edge away. Defence: the presenter, not the room, holds the pen — and the facilitator protects any decision the room can't argue against specifically.

The taste tribunal. Seniors restating house style as universal law. Defence: rotate facilitation; tag feedback on the ladder of conviction; make taste-based comments say so ("this is preference, not principle").

The hostage situation. Critique as approval gate, so no one presents until the work is polished armour that the room can't penetrate. Defence: critique early work explicitly. Rough work presented with sharp questions is celebrated; finished work presented for blessing is sent back with "bring us the next problem".

The one-page checklist

Print this. Pin it. It's the whole method:

☐ Presenter states: constraint, narrowed decision, specific question.
☐ Room of four to seven; facilitator named and not the most senior.
☐ Five minutes silent looking before anyone speaks.
☐ Every observation names element, behaviour, and user effect.
☐ Feedback tagged: blocker / consideration / stirring.
☐ Questions before corrections; taste admitted as taste.
☐ Disagreement recorded as a fork, not resolved by rank.
☐ Presenter summarises takeaways and one thing they won't act on.
☐ Fifty minutes, hard stop. Unfinished business books its own slot.

Run it for eight weeks before judging it. The first two sessions feel stiff — the rules are doing social work the room isn't ready to do for itself yet. By week six, people start saying "blocker or consideration?" to each other without the facilitator, and that's when you know the culture has it.

What changes when critique works

The visible change is work quality, and it's real — the Brightmarsh onboarding flow survived a genuinely brutal Thursday session in week two, and the version that shipped is unrecognisably better for it. But the structural change matters more: critique done well is how a studio transmits taste as reasoning. Juniors absorb not just what good looks like but what arguing towards good sounds like. That is the difference between a studio with one good designer and a studio with a standard.

It's also — no small thing — a recruiting asset. Every strong designer we've hired in the last three years has mentioned sitting in on a Thursday critique during interviews. People want to work where feedback is sharp, fair, and about the work. Build the ritual that deserves them.

Key takeaways

Critique exists to widen the presenter's options, not to approve work or enforce taste; the presenter always holds the pen.
Structure is the kindness: named roles, silent looking before speech, and a fifty-minute hard stop protect both the work and the juniors.
"Specific or silent" plus the blocker/consideration/stirring ladder turns vague reaction into usable signal.
Senior disagreement is a finding, not a fight — record the fork and test it.
Judge any critique ritual after eight weeks, and watch for the tell: participants enforcing the rules on each other.

FAQ

How do you run critique with remote teams?

The same, plus stricter mechanics: shared written board for the silent-looking phase, cameras optional but looking mandatory, and the facilitator reads the board aloud in round-robin so no voice gets buried under latency. Remote critique done tightly is often better than an in-person free-for-all.

What if there's only one designer?

Critique against constraints instead of colleagues: present the work to yourself in writing — problem, narrowed decision, question — then sleep on it and answer in the morning. Better still, build a reciprocal arrangement with a designer at another studio and trade critique hours monthly. Our approach to solo engagements does exactly this with client-side product people.

How is critique different from a design review in sprint?

A review approves; a critique explores. Reviews check the work against requirements and are mostly about risk. Critique checks the thinking against the goal and is mostly about possibility. Conflate them and you get slow reviews and shallow critiques.

What do you do when a client is in the room?

Same rules, lighter ladder, more facilitator care — and never embarrassed a client into silence. Client feedback follows the ladder too: their "I don't like green" is stirring-grade; their "our members call this a levy, not a fee" is a blocker. The rules dignify everyone's input while pricing it correctly.