EvidenceNotAesthetics [principle] v1.0.0

Design decisions affecting user behavior, conversion, or accessibility must be grounded in evidence (user research, A/B data, analytics, heuristics) — not in aesthetic preference alone. Aesthetic preferences are valid inputs but they cannot override evidence of user harm or metric regression.

Implications

• An aesthetic objection ('I prefer this layout') must be paired with evidence or a testable hypothesis to carry weight in design review.
• Evidence from one context does not automatically transfer — test the evidence in the actual product context.
• Heuristics (Nielsen's 10) are evidence-level inputs. Personal opinion is not.

EvidenceNotAesthetics [principle] v1.0.0

Design decisions affecting user behavior, conversion, or accessibility must be grounded in evidence (user research, A/B data, analytics, heuristics) — not in aesthetic preference alone. Aesthetic preferences are valid inputs but they cannot override evidence of user harm or metric regression.

Implications

• An aesthetic objection ('I prefer this layout') must be paired with evidence or a testable hypothesis to carry weight in design review.
• Evidence from one context does not automatically transfer — test the evidence in the actual product context.
• Heuristics (Nielsen's 10) are evidence-level inputs. Personal opinion is not.

Rationale

Aesthetic preferences are ubiquitous, cheap to generate, and often conflicting. Evidence is scarce, expensive to produce, and convergent. When preference and evidence conflict, evidence wins. This principle prevents the common failure of 'it looks better' overriding 'users are failing this task'.