Critical product review

Critical Product & UX Review

Use this skill when reviewing product specifications, UI designs, or implemented screens to ensure the interface is genuinely easy to understand and operate — never rubber-stamping a confusing workflow just because it works technically.

The Seven-Step Inspection Sequence

Evaluate every screen or feature through a concrete user task:

1. User goal: What specific outcome is the person trying to accomplish right now?
2. Prior knowledge: What does the user already know at this step, and what system jargon are we wrongly assuming they know?
3. Perception & comprehension (Norman & Krug): What must the user notice, read, understand, and decide on this screen?
4. Friction & hesitation points: Where does the user pause, second-guess a label, hunt across menus, or perform manual busywork?
5. Elimination, consolidation & defaults (Rams & Tesler): What can be removed, merged, or handled automatically by a safe system default?
6. Essential visibility: Which controls, warnings, and status indicators must remain explicitly visible?
7. Exception resilience: How does the simplified flow handle errors, empty states, permissions, and edge cases?

Workflow

1. Separate essential domain complexity from accidental UI clutter (Tesler's Law):
   • Never pretend complex domain rules do not exist, and never confuse visual emptiness (white space with mystery icons) with genuine usability.
   • Shift burden from the user to the system: pre-fill known context, calculate derived values automatically, and validate inline before submission.
2. Audit primary action hierarchy:
   • Every view should have an unmistakable primary purpose and next step.
   • Flag screens where 5 equal-weight primary buttons, duplicate toolbars, or secondary configuration panels compete with the main task.
3. Inspect before-and-after task context:
   • Always check where the user came from (entry trigger, navigation item, notification link) and where they land after completing the action (confirmation, updated list, or next logical step).
4. Separate evidence tiers:
   • Clearly distinguish observed usability defects (contradictory labels, dead-end flows, hidden primary actions), heuristic conclusions, and hypotheses to validate with users.
   • When a screen is already clear and well-structured, state so briefly instead of inventing artificial critiques.