How to give design feedback without dictating the solution
A six-step guide for PMs on reviewing mockups: share the problem, sort your feedback, describe impact, ask first, and write down what was decided.

You open the mockup and say the first thing that comes to mind: “make the button bigger,” or “move this up.” The designer nods, makes the change, and feels a little less trusted. Meanwhile nobody has asked whether the screen solves the problem. This guide is for product managers who work with designers and are not sure where their input helps and where it turns into designing for them. By the end you will have a repeatable way to run a design review: a shared yardstick, sorted feedback, comments that describe impact, and a written record of what was decided.
- Restate the problem, the constraint and the success metric in two sentences before the review.
- Sort each comment: problem fit, constraint, or taste. Say which one it is.
- Describe what you observe and why it matters. Do not give instructions.
- Ask questions before verdicts. Offer a solution only when asked, as one option.
- End by writing down decisions, open questions and owners.

Why “make it bigger” goes wrong
A designer has spent days with the problem. They weighed options you have not seen. When you give an instruction, you skip all of that and hand them a conclusion. They can only obey or argue about pixels.
It also hides your real concern. “Make the button bigger” usually means “I am not sure a user will notice this.” That is a useful thing to say. It is a problem the designer can solve in several ways, and a bigger button may be the weakest of them.
Write the yardstick before the review
Before the meeting, write two sentences. The first names the user problem and the constraint. The second names the success metric. For example: “New users do not know what to do first, and we cannot change the sign-up form. We will know it works if more of them reach the step that drives activation.”
Send it to the designer ahead of time, or open the review by reading it aloud. If the designer disagrees with it, you have found the real conversation, and you found it before anyone argued about colour.
Sort your feedback into three kinds
Each comment you have belongs in one of three groups. Naming the group does most of the work. It tells the designer how much weight to give the comment and whether it is open to debate.
| Kind | The question behind it | How to say it |
|---|---|---|
| Problem fit | Does this solve the stated problem? | “I am not sure a user would find the next step. Here is why.” |
| Business or technical constraint | Is there a fact the designer cannot know? | “This is a fact, not an opinion: we cannot do X in this release.” |
| Taste | Do I just prefer it another way? | “This is personal preference. Ignore it if it does not help.” |
Constraints are the one place where you hold information the designer lacks. Share them, and share them early. Taste is allowed, but label it. Taste disguised as a requirement is the fastest way to lose a designer’s trust.

Say what you see and why it matters
Replace instructions with an observation and its impact. An observation is something you noticed. The impact is why it matters for the user or the metric.
That sentence points at a problem and leaves the fix open. The designer might change the layout, the wording, the order, or something you would never have thought of. Compare it with “move the invite button up.” That one closes the door.
Ask before you give verdicts
Questions show you are trying to understand the work, not grade it. Two are worth keeping in your pocket:
- “What did you consider and drop?”
- “What would a user do first here?”
The first surfaces options you did not know existed, and sometimes the idea you were about to suggest. The second makes the designer walk through the screen as a user, and often the gap you noticed shows up in their own answer.
Offer a solution only when asked
Sometimes a designer says, “I am stuck, what would you do?” Then give an idea. Label it: “One option would be this. It is not the answer.” Keep it rough. A sketch in words is enough.
If you have a strong solution in mind and nobody asked, hold it back and try describing the problem again. If the designer still does not see it, you can say, “I have an idea. Do you want to hear it?” That small question keeps the decision with them.
Close by writing it down
Finish the review with a short written note in one place the whole team can see. List three things: decisions made, open questions, and who owns each. Without owners, open questions drift, and the next review starts from scratch. Written notes also protect against the late surprise where a constraint appears after the design is done.
A worked example: the onboarding checklist
A designer shows a mockup of an onboarding checklist. You look at it and think it is a lot. Your first instinct is the weak comment on the left. The version on the right is the same concern, rewritten with the steps above.
“Too cluttered, cut three steps.”
It is a verdict and an instruction. The designer does not know why you want fewer steps, or which three, or whether the reason is the user or your taste.
“With this many steps at equal weight, a new user may not see that inviting teammates matters most, and invites drive our activation metric. Does the checklist show that?”
It names the observation, the impact and the metric. It does not say what to cut.
The yardstick for this review was two sentences: “New users do not know what to do first. No hard technical constraint is known yet, and we will know it works if more users invite teammates, which is what drives activation.” Saying “no known constraint” aloud is useful. It tells the designer the field is open.
After the comment, you ask two questions:
- “What did you consider and drop?” Perhaps the designer tried grouping steps and dropped it for a reason you can now hear.
- “What would a new user do first here?” If the answer is “probably the top item,” you can both look at what the top item is.
You also say what works. Perhaps the checklist makes progress visible, which is a real strength. Say it plainly, with a reason.
- Decision: the invite step gets more visual weight than the other steps. Owner: designer.
- Open question: can some steps be grouped or hidden until later? Owner: designer, with a proposal at the next review.
- Open question: are all the steps required for a new user? Owner: PM, who checks and reports back.
The PM did not cut three steps. The team agreed on the problem, and each open question has one person responsible for it.

Mistakes to avoid
- Feedback in a big group chat. Many voices pile on, the designer gets contradictory notes, and the thread buries the decisions. Review live or in a short written note, with a small group.
- Taste dressed as a requirement. If it is a preference, say “this is my preference.” Designers can tell, and they remember.
- New constraints late. If you know a limit, say it at the start. A constraint that appears after the design is finished feels like moving the goalposts.
- Reviewing only the polished screens. Rough work is where your input costs least and helps most. Ask to see early sketches and flows, not just the final frames.
- Never saying what works. Name what is working and why. It tells the designer what to keep, and it makes the harder comments easier to hear.
Try it on your next review
You do not need to change everything at once. Write the two-sentence yardstick, label each comment as fit, constraint or taste, and send a written summary afterwards. If you want to rehearse, take a real mockup and draft your comments before the meeting. PikoPM’s Design review tool is one place to do that. Then edit each one until it describes a problem and leaves the solution to the person whose job it is to find it.


