A Better Way to Review

Good reviews make the work clearer without making the author smaller.

Start with what is visible

I try to separate questions, observations, and decisions so a review can move forward instead of becoming a debate about taste.
That one distinction has changed the tone of my reviews more than any clever checklist.

An observation describes what is visible: the button is hard to find, the error message appears below the fold, or the interaction behaves differently from the rest of the product.
A question opens a door: what happens if this request fails, or could this action be available earlier in the flow?
A decision closes a loop: we will keep this layout, use the existing component, or revisit the problem in a separate task.

When all three are mixed together, feedback can feel like a verdict.
When they are separated, the author can respond to the actual issue instead of trying to decode the reviewer's mood.

Review against the goal

I also try to review the work against its goal, not against my personal taste.
The goal might be clarity, speed, accessibility, or a reliable way to recover from an error.
That gives the conversation something more useful than preference to hold onto.

Make room for honesty

Kindness matters here, but kindness is not the same as avoiding difficult feedback.
The kindest review I can give is specific enough to help someone make the work better and respectful enough to keep the collaboration open.