Updated 2026-07-27
"What do you think?" is the worst question in design, and it is the one almost everyone asks. It invites an opinion, and opinions are exactly what you do not need — you need a diagnosis.
Feedback quality is mostly determined before anyone looks at your work, by how you framed the request. Here is how to frame it.
Say what the thing is supposed to do
A logo for a children's dentist and a logo for a law firm succeed at opposite things. Without the goal, every reviewer substitutes their own taste, and you get a pile of contradictory preferences.
Lead with one sentence: who this is for, what it needs to make them do or feel, and where it will appear. Reviewers can then tell you whether it works, which is a fact, instead of whether they like it, which is not.
Ask about the thing you are unsure about
You almost always know where the weak spot is. Naming it — "the hierarchy in the middle section feels muddy to me" — gets you a focused answer, and it tells the reviewer you are past the stage of needing encouragement.
Show it in context, not on a nice background
A logo centred on a soft gradient looks good in a way that tells you nothing. Show it at the size it will really appear, on the surfaces it will really appear on, next to the competitors it will really sit beside. Half of all design problems only become visible in context.
Do not explain it first
If you narrate your intention before someone looks, you have destroyed the only test that matters — whether the work communicates without you standing next to it. Let them react, then explain, then ask what changed.
How to tell good feedback from noise
- Good feedback describes what the reviewer experienced ("I read the price before the product name") rather than issuing instructions ("make it blue").
- Good feedback is specific enough to act on. "It feels cheap" is a symptom, not a note — ask what specifically made it feel that way.
- "Make the logo bigger" is usually a real problem badly translated: something is not prominent enough. Find the real problem before obeying the instruction.
- Consensus is not correctness. Five non-designers agreeing they prefer the worse option is data about that audience, not about quality — but if they all misread the same element, that is a genuine defect.
- A reviewer who only compliments is being kind, not useful. A reviewer who only criticises may be showing off. You want the one who tells you what works and why, because that is what you have to protect while you fix the rest.
What to do with it afterwards
Collect notes, then wait a day before acting. Distance turns "they hated it" back into "three people stumbled on the same section". Fix the stumbles. Ignore the preferences — including your own, when they conflict with what the piece is for.
And do not revise on every note at once. Change one thing, look again. Piling on ten fixes simultaneously is how designs turn into compromises nobody chose.