How PageLint scores a landing page
What gets audited
PageLint audits copy, not design. The engine fetches your page, extracts the rendered text (headlines, body, CTAs, navigation, social proof), and runs it through 55 checks grouped into 9 lenses. Each check asks one specific, falsifiable question about your copy. English-language pages only, for now.
The 9 lenses
- Clarity — Can a first-time visitor tell what you sell, for whom, in five seconds?
- Relevance — Does the copy speak to the audience it names — their words, their problems?
- Value — Are benefits concrete and quantified, or vague claims anyone could make?
- Trust — Proof, specificity, and credibility signals where skepticism peaks.
- Motivation — Does the page build desire before it asks for action?
- Friction — Anything that makes the next step harder than it needs to be.
- Distraction — Competing CTAs, off-message copy, exits placed before conversion.
- Dark Patterns — Manipulative tactics that convert short-term and burn trust long-term.
- Schwartz Modifiers — Market-sophistication adjustments from Eugene Schwartz’s framework.
The free quick audit runs the Clarity and Trust lenses (16 checks). The full report runs all 55 checks across all 9 lenses.
How the score is computed
Every check returns a severity: passing, low, medium, or high. Each severity carries a fixed credit weight: passing earns 100% of a check’s credit, low 75%, medium 40%, high 0%. Your score is the earned credit divided by the maximum possible, scaled to 0–100 — per lens, and overall. No curve, no AI vibes-score: the same page always produces the same arithmetic.
Checks that cannot run on a given page (for example, testimonial checks on a page with no testimonial section where one is not expected) are shown as informational and excluded from the denominator, so you are never penalized for a check that does not apply.
Where the checks come from
Every finding cites its source. The check library is built on primary sources and published research: David Ogilvy’s advertising principles, Robert Cialdini’s persuasion research, Eugene Schwartz’s Breakthrough Advertising(market sophistication and awareness stages), and conversion research published by CXL Institute, Nielsen Norman Group, and Baymard Institute. When a finding says “quantify the claim,” the citation shows you which principle it traces to — so you can judge the advice, not just trust it.
The teardown gallery
The public gallery at pagelint.dev/audits is produced by the same engine, same checks, same weights — not edited by hand. Each teardown page shows the date it was generated. Nobody paid to be on the list, and nobody can pay to change a score.
Limitations
- Copy only — PageLint does not evaluate visual design, layout, or brand.
- English-language pages only.
- A score is a diagnostic, not a guarantee: it measures adherence to documented copywriting principles, which correlate with — but do not promise — conversion.
- Findings are generated per page snapshot; if you change the page, re-run the audit.
Versioning
The check library and weights are versioned in code. When checks are added or reweighted, this page is updated and new audits use the new library; existing reports keep the scores they were generated with. Material scoring changes are noted here.
Questions about a specific check or a score you disagree with? Email [email protected] — built by one person, and I read everything.