How we score a funnel
Rubric v1 · engine 1.0.0. Every audit records the rubric version that produced it.
A score is only useful if you can see how it was made. Each funnel — yours and every competitor’s — is crawled in a real browser, measured, and scored on the same rubric. Checks are either objective (measured against a published standard, like Core Web Vitals) or judgmental (defensible heuristics, like headline specificity). Objective checks carry 70% of the total weight; judgmental ones 30%.
If a signal cannot be measured — for example, a site blocks our crawler’s mobile pass — that check is left out rather than scored as zero, and the audit’s data confidence goes down. Low-confidence audits are reviewed by an analyst before you see them.
Dimension weights
- Performance20%
- Conversion structure25%
- Messaging clarity15%
- Technical & SEO15%
- Trust & credibility10%
- Mobile experience15%
Performance · 20% of the score
Largest Contentful Paint objective
Lab LCP from the crawl. 1.0 at ≤ 1.2 s, 0.9 at the Core Web Vitals “good” threshold of 2.5 s, 0 at ≥ 4.0 s.
25%of this dimension
Interactivity (INP / TBT) objective
INP when available, otherwise lab Total Blocking Time. INP: 1.0 ≤ 100 ms, 0.9 at 200 ms, 0 ≥ 500 ms. TBT: 1.0 ≤ 100 ms, 0.9 at 200 ms, 0 ≥ 600 ms.
17%of this dimension
Cumulative Layout Shift objective
Lab CLS. 1.0 ≤ 0.05, 0.9 at 0.10, 0 ≥ 0.25.
17%of this dimension
Server response time objective
Time to first byte. 1.0 ≤ 400 ms, 0.9 at 800 ms, 0 ≥ 1800 ms.
8%of this dimension
Total page weight objective
Bytes transferred. 1.0 ≤ 0.75 MB, 0.9 at 1.5 MB, 0 ≥ 4 MB.
8%of this dimension
Render-blocking resources objective
Count of synchronous scripts and stylesheets in <head>. 1.0 at 0, 0.9 at 1, 0 at ≥ 6.
8%of this dimension
Lighthouse performance (desktop) objective
Google Lighthouse performance category score, used as-is.
17%of this dimension
Conversion structure · 25% of the score
Call to action above the fold objective
At least one call-to-action visible in the first viewport (measured in a real browser; text-position estimate otherwise).
27%of this dimension
Call-to-action clarity judgmental
Primary CTA label names the action and the outcome (2–6 words, not a generic "Submit"/"Click here").
18%of this dimension
Call-to-action focus judgmental
Number of distinct CTA labels: 1–3 full credit, 4–6 partial, more than 6 dilutes attention.
9%of this dimension
Form friction objective
Fields in the largest form: ≤ 4 full credit, 5–6 → 0.7, 7–9 → 0.4, ≥ 10 → 0.1. Not applicable when the page has no form.
18%of this dimension
Trust signals near the decision judgmental
Presence of testimonials, a guarantee and trust badges: one point each, scaled to 0–1.
18%of this dimension
Social proof judgmental
Quantified proof (e.g. “2,000+ patients”, star ratings, review counts). Two or more instances earn full credit.
9%of this dimension
Messaging clarity · 15% of the score
Headline specificity judgmental
A single H1 of 4–14 words that names a concrete outcome, number or the target keyword.
30%of this dimension
Value proposition present judgmental
Headline plus supporting subheading or description, with explicit benefit language.
20%of this dimension
Readability objective
Flesch–Kincaid grade of visible copy. Full credit ≤ grade 8, none ≥ grade 14.
20%of this dimension
Customer-focused language judgmental
Ratio of “you/your” to “we/our”, and benefit terms to feature terms.
20%of this dimension
Jargon density judgmental
Buzzwords (“synergy”, “cutting-edge”, “world-class”…) per 1,000 words. Full credit at 0, none at ≥ 5.
10%of this dimension
Technical & SEO · 15% of the score
Title tag objective
Present and 10–65 characters.
15%of this dimension
Meta description objective
Present and 50–160 characters.
8%of this dimension
Canonical URL objective
A rel="canonical" link is declared.
8%of this dimension
Structured data objective
At least one schema.org type declared (JSON-LD or microdata).
8%of this dimension
Indexable objective
HTTP 200 and no noindex directive.
15%of this dimension
HTTPS objective
Final URL is served over HTTPS.
15%of this dimension
Keyword relevance objective
Share of the claimed keyword’s terms found in title (35%), H1 (30%), meta description (15%) and body (20%).
15%of this dimension
Image alt text objective
Share of <img> elements with an alt attribute.
8%of this dimension
Lighthouse SEO objective
Google Lighthouse SEO category score (mobile).
8%of this dimension
Trust & credibility · 10% of the score
Secure connection objective
Final URL uses HTTPS.
11%of this dimension
Privacy policy objective
A link to a privacy policy is present.
22%of this dimension
Contact details objective
Phone, email, street address or contact page: any two earn full credit.
22%of this dimension
Company identity objective
Copyright line, terms link or about link: any two earn full credit.
11%of this dimension
Reviews judgmental
Review mentions, ratings or review structured data. Three signals earn full credit.
22%of this dimension
Domain age objective
Registration age from WHOIS/RDAP: full credit ≥ 2 years, none < 90 days.
11%of this dimension
Mobile experience · 15% of the score
Mobile viewport objective
Viewport meta tag with width=device-width.
22%of this dimension
No horizontal scrolling objective
At a 390 px viewport, content does not overflow horizontally.
22%of this dimension
Tap target size objective
Share of links/buttons smaller than 24×24 px (WCAG 2.2 target size). Full credit ≤ 5%, none ≥ 30%.
22%of this dimension
Legible font size objective
Share of text rendered below 12 px on mobile. Full credit ≤ 5%, none ≥ 40%.
11%of this dimension
Mobile speed objective
Lighthouse mobile performance when available; otherwise mobile LCP on the same curve as desktop LCP.
22%of this dimension
Rankings, benchmarks and disputes
Your rank is computed against the approved competitor set for your keyword and market: the top commercial search results, excluding directories, marketplaces and social platforms. Benchmarks are only shown when at least five funnels have been measured, so no individual score can be inferred.
Disagree with a score? Reply to your audit email. Every finding links to the measurement that produced it, and an analyst will review it.