Critique For All
A framework for turning feelings into actionable design claims.
A good critique makes expertise legible.
Critique is one of the highest-leverage things a product team can do. When it works, it sharpens thinking, surfaces blind spots, and produces decisions other people can learn from. When it does not, you get polite nodding, unresolved tension, or outcomes driven by whoever talks last.
The difference is not talent or intention, it’s structure.
The concept is simple. Critique should produce knowledge, not just consensus. Every design or research choice encodes assumptions. Those assumptions can be named, tested, and refined. This framework makes that process explicit.
Use it for design system decisions, product design reviews, research plan critiques, final report evaluations. Scale it up for high-stakes decisions or down to a single principled comment.
What this is
Evaluation based on principles, not preference.
The framework evaluates reasoning, not artifacts. A design that works but cannot be explained scores lower than one with clear, principled rationale, because inarticulate excellence cannot be taught or reliably reproduced.
When work scores low on articulation but succeeds in practice, that is a signal to extract the implicit logic, not penalize the practitioner.
What it does
Evaluates the reasoning behind a decision
Names principles and tradeoffs
Produces documented, teachable knowledge
What it replaces
Voting on which option feels best
Undeveloped opinions
Verbal feedback that disappears
This framework provides eight evaluation criteria, each with scoring rubrics, critique sentence formats, and guidance tuned to design and research work. It includes a shared vocabulary of 89 principles organized by decision domain, so reviewers have a standardized language to name what they see. And it includes scoping guidance so teams know when to run the full rubric and when a three-sentence comment is enough.
It evaluates reasoning, not just artifacts. A design that works but can't be explained scores lower than one with clear, principled rationale, because inarticulate excellence can't be taught, refined, or reliably reproduced. When work scores low on articulation but succeeds in practice, the framework treats that as a signal to extract the implicit logic, not penalize the practitioner.
This framework evaluates design and research reasoning — the quality of thinking behind decisions, not just the artifact produced.
A design that works well but cannot be explained scores lower than one with clear, principled rationale. A research study that produces findings but cannot articulate its methodology scores lower than one with transparent, defensible choices.
This is intentional: inarticulate excellence cannot be taught, refined, or reliably reproduced. The goal is to make expertise legible.
When work scores low on articulation but appears to succeed in practice, treat this as a signal that the underlying principles need to be extracted, not that the work is wrong. The critique process is to surface implicit logic, not penalize practitioners who haven't yet formalized their reasoning.
The instrument
Eight criteria.
Every piece of work is evaluated against these criteria, with rubrics and critique sentence formats tuned to design and research separately. Weights shift by context and always total 100.
Purpose clarity
One clear job
Tradeoff articulation
What was sacrificed
Principle grounding
Reasons, not preference
Falsifiability
A testable claim
Complexity justification
Earn every addition
Inheritability
Others can learn it
Contextual fit
Works in its constraints
Evidence sufficiency
Research only
The vocabulary
Nine categories.
A shared vocabulary of 89 principles, organized by the type of problem, so a reviewer can name what they see instead of reaching for "something feels off."
Perception
What the eye takes in
Interaction
How it responds to action
Behavior
How people actually act
Findability
Whether it can be found
System
How the parts hold together
Message
What the words convey
Inclusion
Who gets left out
Ethics
Whose interest it serves
Evidence
What you can claim
How principles hold
Three kinds of principle.
Not every principle carries the same weight. The framework separates what predicts, what a craft agrees on, and what you owe.
Descriptive. Predicts behavior. Trades freely against other goals.
Conventional. Agreed craft discipline. Holds by commitment.
Normative. What you owe. A floor, not a weight.
How to use it
Scale it up, or down to one comment.
Use the full rubric for high-stakes decisions, or a single principled sentence when that is enough. At every scale, the goal is the same: make expertise legible.
5 min
Name the principle, state the tradeoff, suggest a direction.
20 min
Read the purpose and scoping guidance, plus one worked example.
1 hour
Score a piece of your own work against the rubric.
First session
Pick two to four criteria, not all eight.
Scoping Guidance
When to use this framework
Recommended Approaches
Design system patterns, components, or architectural decisions
Full rubric with scoringMajor product design decisions (new flows, significant redesigns)
Full rubric, adjust weights for product context
Research plans, methodologies, reports
Full rubric, adjust weights for research contextRoutine design reviews (iterations, minor updates)
Critique Structure Template only (skip scoring)Quick feedback (copy, color, spacing tweaks)
Principle-based comment only: name the principle, state the tradeoff