What this page covers
Error reporting, test design, and what to disclose publicly.
Who it is for
Teams reviewing vendor claims and planning their own validation.
Last reviewed
March 24, 2026.
Ask For The Reference Point
“Accuracy” can mean agreement with hand measurements, recommendation acceptance, or purchase outcomes. Without a reference point, the number is too vague to be actionable.
Average Error Matters More Than A Marketing Ceiling
A public page should explain median or average error, not just cite a best-case scenario. It should also state whether the metric applies equally across categories and body types.
Brand Mapping Changes The Result
Even a strong measurement estimate can produce weak recommendations if the catalog mapping is shallow or outdated. Validation should cover both estimation and brand-size interpretation.
Public Disclosures Build Trust
Landing pages do not need to publish proprietary datasets, but they should explain cohort size, category scope, and where uncertainty is highest.
Minimum Methodology Checklist
- What was the comparison baseline?
- How large and how diverse was the cohort?
- Which categories and brands were included?
- How were outliers and low-confidence results handled?
- Was the result reproduced after rollout?
Methodology
This page is intentionally framed as a review checklist rather than a numeric claim page. That makes it safer for readers and more useful for teams who need to ask better questions before trusting a headline percentage.