Skip to main content

roast.page Research

Original data on what actually converts.

We score landing pages for a living. These studies pull patterns from the resulting dataset — headlines, CTAs, AI-built pages, mobile-vs-desktop, trust signals. No vendor takes. No vibes-based advice. Methodology included on every page.

Updated May 24, 2026 · See the headline report

Copy

Why Landing Page Headlines Fail: A 1,200-Page Study of What Works

We scored the H1 of 1,200 live landing pages. Outcome-driven headlines beat feature-driven headlines by 14 points on average. Length matters less than specificity. Full data, by industry.

11 min read · 1,200 H1 headlines, hand-classified across 6 industries. May 2026.

+14 pts

Gap between outcome and feature headlines

Conversion

The CTA Tax: How Generic Button Copy Costs SaaS Pages 31% in Conversion

Pages with one named CTA outscore pages with three generic CTAs by 31%. We pulled every button label from 1,200 pages. The patterns the top quartile share.

9 min read · 8,400 CTA buttons across 1,200 pages. Tagged by verb, length, and placement.

31%

Score gap: one named CTA vs three generic ones

AI & Builders

Do AI Page Builders Make Better Landing Pages? A 240-Page Audit

We scored 240 landing pages built with Framer AI, Mixo, Durable, Wix ADI, and Unbounce Smart Builder against 240 hand-coded pages. AI-built pages averaged 9 points lower. The patterns repeat.

13 min read · 240 AI-builder pages vs. 240 hand-coded pages, matched by industry and traffic tier.

−9 pts

AI-built pages vs. hand-coded average score

Mobile

The Mobile Conversion Gap: Why 6 in 10 Landing Pages Lose on Mobile

Same pages, two viewports. The average page scored 9 points lower on mobile than desktop. The fixes — CTA placement, type size, viewport meta — take less than a day.

10 min read · 1,000 pages scored on both desktop (1280×800) and mobile (390×844) viewports.

−9 pts

Average mobile-vs-desktop score gap

Trust

The Trust Signal Hierarchy: What Visitors Actually Look At Above the Fold

Logos, customer counts, named testimonials, star ratings — which trust signals correlate with the highest scores? We ranked them. The answer surprised us.

10 min read · 1,200 pages tagged for 11 distinct trust signals. Bivariate correlations against overall and dimension scores.

r = 0.54

Correlation between quantified proof and overall score

About this research

Where the data comes from

Every page we publish here is built on landing pages we scored ourselves. roast.page runs each URL through three pipelines — screenshots at desktop and mobile viewports, an HTML extractor that captures every heading, button, and trust signal, and Google PageSpeed Insights for the Core Web Vitals. A Claude vision model then scores eight weighted dimensions on a 1–10 scale and writes the recommendations.

Studies sample from this scored dataset. Where we make claims like “outcome-driven headlines score 14 points higher,” we mean: we hand-classified the H1 of every page in the sample, ran a two-sample comparison, and report the mean difference. Where a finding is correlational, we say r. Where the sample is small or skewed, we say so. We don’t round inconvenient numbers up.

Citations are welcome. Please link back to the specific study, not the homepage, so readers can verify the methodology. We update quarterly as the dataset grows.

Also from roast.page

Related work