Evidence-earned scoring
01Starts at zero, not fifty
- Exact phrase in hero/title → verified (full credit).
- ≥70% meaningful tokens → strong (most credit).
- 40–69% meaningful tokens → partial (calibrated fraction).
- Below 40% → fail (zero credit, not a partial guess).
Generic tokens like "marketing," "teams," or "B2B" are tokenized but never treated as a full match by themselves.
What we measure
02Five positioning pillars
- Audience clarity in hero / title
- Offer / value proposition specificity
- Differentiation vs commodity claims
- CTA alignment with desired action
- Proof language proximity
Limits
03What this model does not cover
- Single-page HTML — not brand research interviews or competitive teardown depth.
- Missing audience/offer inputs leave those dimensions not scored (not free points).