Pro
Landing Page CRO Audit
Scored conversion audit, copy rewrites and A/B test plans with real sample sizes
Landing Page CRO Audit gives your agent a repeatable way to find out why a landing page is not converting and what to do about it. It is built for founders, marketers, growth and paid-media teams, and freelancers who review pages for clients.
How it works
The skill runs a first-screen test on desktop and mobile (what is it, who is it for, why care, what next) and checks message match with the ad or email that sends the traffic. It then scores 35 criteria across seven weighted areas - clarity, value proposition, friction, trust, calls to action, mobile and speed (Core Web Vitals from your PageSpeed data) - with evidence for every score, and groups the gaps into findings ranked by impact, confidence and ease. Clear errors are marked "fix now"; open questions become A/B tests with a hypothesis, primary and guardrail metrics, and a sample size and duration from the bundled calculator, including an honest warning when traffic is too low to test.
Rewrites follow a copy framework: three headline options with reasoning, subhead, primary and secondary CTA, first-screen layout, benefit blocks and objection handling. Wherever real proof is needed, the skill leaves a placeholder instead of inventing testimonials or numbers.
What you get
- Audit checklist, scoring rubric with anchors, copy framework and A/B testing guide
- Report, scoring-sheet (CSV) and test-plan templates
- A sample-size, duration and detectable-lift calculator
- A full worked audit of a fictional SaaS landing page
The score is for tracking your own page over time, not a benchmark. No conversion lift is promised.
Highlights
- 35-criterion scorecard: clarity, value, friction, trust, CTA, mobile, speed
- First-screen and message-match test against your ad or email
- ICE-prioritized findings split into fix now, test and research
- Headline, CTA and section rewrites with no invented proof
- A/B test plans with a sample-size and duration calculator
Example
Audit our Google Ads landing page for clinic booking software. ~340 visitors/day, 1.9% request a demo, 70% mobile. PSI: LCP 4.1 s, INP 180 ms, CLS 0.22.
# CRO audit - Tidewater Scheduling, Google Ads landing page Verdict: the ad promises "online booking" and a "14-day free trial"; the page says "The Future of Clinic Management Is Here" and never mentions the trial. Visitors who want to act meet "Get Started" and a 9-field form. On phones the first screen is a slow stock photo. Scorecard: Clarity 4.0/20 | Value 4.0/20 | Friction 3.0/15 | Trust 6.0/15 | CTA 3.0/10 | Mobile 4.0/10 | Speed 4.0/10 | Total 28/100 (baseline for this page, not a benchmark) Top findings (ICE): 1. First screen ignores the ad's promise (80, test) 2. Trial hidden, vague CTAs (80, test) 3. 9-field form (64, fix now) 4. Mobile first screen (48, fix now) 5. LCP 4.1 s from a 2.8 MB PNG hero (48, fix now) Traffic check: +25% needs 14,553 visitors per arm (86 days) - too long. Test 1 bundles a new first screen at +40% MDE: 6,055 per arm, run 42 days.
Rewrite this headline: 'Unlock the Power of Smarter Finance'. We make expense software for agencies with 10-50 people. Ad keyword: 'expense tracking for agencies'.
Current: "Unlock the Power of Smarter Finance" - slogan; no category, audience or outcome; no match with the ad keyword. Options: 1. "Expense tracking for agencies - receipts, approvals and client rebilling in one place" - echoes the keyword exactly; recommended for paid traffic. 2. "Bill every client expense back, without chasing receipts" - outcome-led; best if rebilling is your main differentiator [confirm]. 3. "The expense app built for 10-50 person agencies" - category + audience; use if size fit is why customers choose you. Subhead: Staff snap receipts, managers approve in one tap, and billable costs flow to client invoices. [insert real setup time or trial terms] CTA: "Start my free trial" (if you offer one) - otherwise "See it with your data". No numbers added: insert real figures only from your customer data.
Our checkout page converts at 3.2% with about 1,500 visitors a day. We want to detect a 10% relative lift. How long?
python3 scripts/sample_size.py --baseline 3.2 --mde-rel 10 --daily-visitors 1500 Result (two-sided, alpha 0.05, power 0.8): 49,777 visitors per arm, 99,554 in total -> 67 days at 1,500/day, so plan 70 days (10 full weeks). If that is too long: - Test a bolder change and size for +20%: 13,015 per arm, about 18 days -> run 21 days (3 full weeks). - Or test where the change acts more directly, with the final conversion kept as a guardrail. Rules: fix the duration in advance, run whole weeks, do not stop the first time it looks significant, and check sample-ratio mismatch before reading results. No lift is guaranteed - a neutral result is still information.
What's inside
landing-page-cro-audit/ ├── agents/ │ └── openai.yaml ├── references/ │ ├── ab-test-planning.md │ ├── audit-checklist.md │ ├── copy-rewrite-framework.md │ ├── scoring-rubric.md │ └── worked-example.md ├── scripts/ │ └── sample_size.py ├── templates/ │ ├── 01-cro-audit-report.md │ ├── 02-scoring-sheet.csv │ └── 03-ab-test-plan.md ├── LICENSE.txt ├── README.md └── SKILL.md
Install by unzipping into your agent's skills folder. Install guide →