CRO Audit Checklist: 30 Checks Before You Touch Your Landing Page

TL;DR

• 30 binary pass/fail checks, grouped by page section. Run them in one sitting, before you change anything.
• Six groups: hero, offer, proof, CTA and friction, message match, speed and mobile.
• Every check is a yes/no question with a one-line reason it matters.
• Median landing page conversion is 6.6% (Unbounce, Q4 2024 data). If you're well under that, start here.
• The checklist catches mechanical failures. It can't judge your offer. More on that at the end.

If you want the full audit process, with analytics setup and prioritization, this isn't that post. That's the pillar: how to do a CRO audit. This post is the checklist itself.

A CRO audit checklist is a fixed list of pass/fail questions you run against a page before changing it. It exists to stop you from redesigning on gut feel. Each check is binary: the page passes or it fails, and every fail is a candidate fix you can rank later.

I analyzed 521 SaaS pages to build pages.report. These 30 checks are the failures I saw over and over. Not "is your value proposition compelling?" fluff you can't score. Concrete questions with yes or no answers.

Grab your landing page. Open it in an incognito tab. Score each check honestly.

Hero and headline (checks 1-5)

Your hero gets about 5 seconds. These five checks decide whether the visitor reads section two or bounces.

• 1. Can a stranger say what you sell after 5 seconds? Show the page to someone cold. If they can't repeat what it is, nothing below the fold matters.
• 2. Does the headline name an outcome, not a category? "AI-powered platform for modern teams" is a category. Outcomes convert; categories describe.
• 3. Does the hero say who it's for? A named audience filters in the right buyer. Vague heroes convert nobody because they're written for everybody.
• 4. Is one primary CTA visible without scrolling? If the visitor has to hunt for the next step, you've added friction at the exact moment intent peaks.
• 5. Does the hero visual show the product doing the thing? Abstract illustrations prove nothing. A real screenshot or short clip is your first piece of evidence.

Offer and pricing (checks 6-10)

Buyers abandon when the deal gets fuzzy. In Baymard's running meta-analysis of checkout studies, extra costs were the top abandonment reason at 40%, and 18% quit over forced account creation (Baymard Institute). That's checkout data, but the lesson transfers: surprises kill conversions.

• 6. Can a visitor find the price without filling out a form? Hiding pricing behind "Contact us" filters out self-serve buyers who were ready today.
• 7. Are all costs visible before commitment? Setup fees, per-seat math, usage overages. Surprise costs are the single biggest abandonment driver in Baymard's data.
• 8. Does the offer state exactly what the buyer gets and when? "Get your first report in 5 minutes" beats "Get started." Specifics make the promise believable.
• 9. Is there a risk reversal near the price? Free trial, money-back guarantee, "no credit card required." One line that removes the fear of being wrong.
• 10. Can a visitor pick a plan in under 30 seconds? If your pricing table needs a spreadsheet to compare, the buyer defers the decision. Deferred decisions die.

Proof and trust (checks 11-15)

19% of shoppers in Baymard's meta-analysis abandoned because they didn't trust the site with their card information (Baymard Institute). Trust isn't a vibe. It's placed evidence.

• 11. Is there at least one number above the fold? "40,000+ customers" or "4.8 on G2." A number is checkable; "loved by teams" is decoration.
• 12. Does every testimonial have a name, face, and specific result? Anonymous praise reads as fabricated, even when it's real.
• 13. Are trust signals sitting next to the CTA? Security badges, review scores, and guarantees work at the moment of hesitation, not three scrolls away.
• 14. Is your newest proof dated within the last 18 months? A 2021 case study or a logo wall of dead startups signals a stale product. Date your proof or refresh it.
• 15. Does the page show the product working, not just claim it? Screenshot, demo, before/after. If you can't show it, buyers assume you can't do it.

CTA and friction (checks 16-21)

Six checks here because friction is where I see the most silent losses. The copy is fine; the asks are wrong.

• 16. Is there exactly one primary action on the page? Two competing CTAs split intent. Pick the one action that matters and demote everything else.
• 17. Does the CTA name the next state, not the abstract action? "See your report" beats "Submit." The button should describe what the click buys.
• 18. Does the CTA repeat down the page? A convinced reader at section five shouldn't have to scroll back up to act.
• 19. Does your form ask only for what this step needs? Every extra field is a toll. Phone number on a demo request? You just taxed your best leads.
• 20. Can the buyer proceed without creating an account? Forced signup drove 18% of abandonments in Baymard's data. Earn the account after the value, not before.
• 21. Have you removed exit ramps near the CTA? Full nav bars, footer link farms, and social icons all offer an easier click than buying.

Message and search intent (checks 22-25)

The page can be flawless and still fail if it answers a question nobody asked. These four checks test the match between what brought the visitor and what they found.

• 22. Does the page mirror the query or ad that sent the visitor? If the ad says "invoice software for freelancers" and the hero says "financial operations platform," you've broken the scent trail.
• 23. Does the page use at least one verbatim customer phrase? Pull it from a support ticket, sales call, or review. Buyers trust pages that sound like their own complaints.
• 24. Does the page answer the top three objections before the final CTA? Price, effort, "will it work for me." Unanswered objections don't disappear; they walk.
• 25. Is it clear who this is NOT for? A page that disqualifies the wrong buyer converts the right one harder. Trying to please everyone reads as evasive.

Speed and mobile (checks 26-30)

Speed is the least glamorous group and the most measurable. Portent's 2022 study found ecommerce conversion at 3.05% for pages loading in 1 second versus 0.67% at 4 seconds (Portent). Older data, but the direction has never reversed. And Google and Deloitte's 2020 "Milliseconds Make Millions" study found a 0.1s mobile speed improvement lifted lead-gen form-page progression by 21.6% (web.dev).

• 26. Does the page load in about 2 seconds on a real phone? Test on cellular, not office wifi. Portent's data shows conversion falling off a cliff past 2 seconds.
• 27. Is the primary CTA visible and tappable on a phone without pinching? Most of your traffic is probably mobile. Check the button size with your thumb, not your cursor.
• 28. Does anything shift while the page loads? A CTA that jumps as images load causes misclicks and rage taps. Layout shift is a conversion bug, not a polish issue.
• 29. Are your images compressed and properly sized? Oversized hero images are the most common cause of slow largest-contentful-paint. Fix the biggest file first.
• 30. Does your form behave on mobile? Right keyboard type per field, autofill working, no fields hidden by the keyboard. Type your own email on your own phone and see.

If you want session recordings to verify the mobile checks, Microsoft Clarity is free with no traffic limits. The full stack comparison lives in the CRO audit tools post.

How to score it

Count your fails. Almost nobody passes clean. Expect to fail a handful; that's normal.

Don't fix in checklist order. Fix in impact order: message match and offer fails first, then friction, then speed, then polish. A fast page with the wrong message is just a fast failure.

Short on time? There's a compressed version of this process: the 15-minute landing page audit covers the single-page essentials when you can't run all 30.

What a checklist can't catch

Every check above is binary on purpose. That's the strength and the limit.

A checklist can't tell you if your offer is weak. It can verify your price is visible; it can't tell you the price is wrong for the market. It can confirm your headline names an outcome; it can't tell you it's an outcome nobody wants. Those are judgment calls, and they're usually the actual problem on pages that pass 25 of 30 checks and still don't convert.

The other blind spot: you wrote the page. You know what every section means, so you can't see where a stranger gets lost. Self-audits catch mechanical failures. They rarely catch comprehension failures.

So run the 30 checks first. Fix the fails. If the number still won't move, that's the point where fresh eyes beat another pass with the same eyes. A CRO audit from someone outside your team, on video, walking your page section by section, catches the things you've gone blind to. That's what I do, and it exists precisely because checklists have a ceiling.

Either way: checklist first. It's free, it takes one sitting, and it clears the mechanical failures so the judgment calls are all that's left.