Inputs
01URL plus conversion context
You provide a landing-page URL and conversion context (offer, audience, or goal framing in the tool UI). Context helps interpret message match and CTA language; it does not connect to your ad accounts or analytics.
- Best fit: paid or organic landing pages with a single primary action (demo, trial, download, contact).
- Weaker fit: multi-step apps, login walls, or pages whose entire offer only loads after heavy client rendering.
What we fetch
02A single public landing-page URL
The Worker fetches HTML with SSRF protections (no private networks, credentials, or non-http(s) schemes). Client-rendered apps may under-report if critical copy is not in the initial HTML.
- Headline and supporting copy structure
- CTA labels, placement cues, and competing actions
- Proof blocks (logos, quotes, metrics language when present)
- Form fields and friction proxies visible in markup
- Message-match signals vs the context you supplied
Dimension weights
03Six conversion pillars, each weighted
| Dimension | Weight |
|---|---|
| Headline outcome clarity | 15% |
| Audience specificity | 10% |
| Value proposition | 12% |
| CTA clarity & visibility | 15% |
| Offer comprehension | 10% |
| Social proof | 10% |
| Objection handling | 8% |
| Page hierarchy | 8% |
| Form friction | 7% |
| Mobile experience | 5% |
| Pricing / commitment clarity | 5% |
| Message match (traffic source) | 5% |
Dimensions are conversion heuristics for public HTML. A high score means the page exhibits patterns associated with clearer conversion paths in this model not that your live CR will match a benchmark table.
Worked interpretation
04Turning a grade into an experiment backlog
Example: a paid search landing page scores 54. Headline and hero are specific (B), but CTA is generic “Submit,” proof is below the fold only, and the form asks for seven fields including budget. Message match to the ad keyword theme is partial.
- Fix friction before poetry: reduce form fields and rename the CTA to the outcome (“Get the audit”) often beats rewriting brand adjectives.
- Proof placement: move one concrete metric or logo row adjacent to the primary CTA; the model rewards visible proof near the ask.
- Message match: if ads promise “SOC 2 for SaaS,” the H1 should not open with a vague mission statement. Align first screen language with the traffic promise.
- Do not invent revenue loss: without traffic and AOV inputs, the tool will not claim “you are losing $X.” Pair with the Marketing ROI Calculator when you want scenario math.
Progressive reveal
05From free to unlocked
- Submit URL and conversion context.
- See overall score and dimension scores free.
- See top three opportunities free.
- Unlock full evidence and markdown with email.
- Continue with Search & AI Visibility services or related free tools.
Validation
06What “good” means here
- Scoring is rule- and evidence-based on HTML patterns, not a neural estimate of conversion rate.
- Missing analytics, heatmaps, or session data reduce confidence they are never faked as passes.
- Grade bands stay strict so “A” remains uncommon for sparse or multi-CTA brochure pages.
- Use before/after runs on the same URL after shipping experiments; compare dimensions, not vanity overall alone.
What we never claim
07Explicit limits
- Your actual conversion rate or revenue loss without traffic/value inputs.
- Heatmaps, session replays, or CRM funnel drop-offs.
- Full multi-step checkout or app-only flows.
- Field Core Web Vitals from HTML alone.
- Statistical significance of any A/B test.