The CRO Audit Checklist: 40 Things to Check Before You Buy More Traffic

Before you buy more traffic, check whether the pages you’re sending it to can convert it. A CRO audit is a structured review of every step between a visitor’s first click and a conversion, looking for broken tracking, friction and missing persuasion. The logic is simple arithmetic: revenue is traffic multiplied by conversion rate multiplied by order value. Doubling conversion rate has the same revenue effect as doubling traffic, without doubling your media bill. Below are the 40 checks I run on every audit, grouped by area, followed by how to prioritize what you find with ICE or PIE.

Before you start

Gather this first, or the audit turns into opinion:

• GA4 funnel data for the last 90 days, split by device and channel
• Your top 10 landing pages by paid spend and by sessions
• Heatmaps or session recordings on key templates, collected in line with consent rules
• Sales call notes, support tickets and reviews, where customers tell you their objections
• At least one real iPhone and one real Android phone

Mark each check as Pass, Fail or Unknown. “Unknown” usually means a measurement gap, and that’s a finding too.

Analytics and tracking (checks 1–6)

I start here because every other check depends on it. If you can’t trust the conversion data, you can’t prioritize or measure anything else.

1. Each conversion event fires once, and only once, per real conversion.
2. Every funnel step is tracked: landing page, product or pricing page, form start or add-to-cart, checkout, purchase or lead.
3. Conversion counts reconcile with your CRM or ecommerce backend within a known, explained gap.
4. Data can be segmented by device, browser, channel and landing page.
5. Internal traffic, test orders and obvious bots are filtered out.
6. Heatmaps or session recordings are running on key templates, with consent respected.

Speed (checks 7–11)

Use field data from real users (Search Console or PageSpeed Insights), not only lab scores. The thresholds below are Google’s published “good” levels for Core Web Vitals.

7. Largest Contentful Paint (LCP) is 2.5 seconds or less on mobile for your top landing pages.
8. Interaction to Next Paint (INP) is 200 milliseconds or less: buttons, filters and menus respond without lag.
9. Cumulative Layout Shift (CLS) is 0.1 or less: nothing jumps around as banners, fonts or images load.
10. Hero images are compressed, served in modern formats and correctly sized, and the main image isn’t lazy-loaded.
11. Third-party scripts are audited: unused apps, old pixels and heavy widgets are removed, and non-critical ones load later.

Message match (checks 12–17)

Message match is how closely a landing page repeats the promise of the ad, email or search result that brought the visitor there.

12. The headline matches the promise of the ad or keyword that sent the traffic.
13. A first-time visitor can tell within five seconds what you offer, who it’s for and why it’s better.
14. There’s one primary call to action per page, visible without scrolling on mobile.
15. Major campaigns land on dedicated pages, not the homepage.
16. Offer, price and terms stay consistent from ad to landing page to checkout.
17. The copy answers the top three objections you hear in sales calls, reviews and support tickets.

Forms (checks 18–22)

18. Forms only ask for fields you actually use; anything else is collected later.
19. Labels stay visible (not placeholder-only), and inputs use the right type (email, tel) so mobile keyboards match.
20. Validation happens inline with helpful messages, and entries aren’t wiped when there’s an error.
21. After submission, the next step is clear, and the confirmation is tracked as a conversion.
22. Spam protection doesn’t punish real people. Prefer invisible methods over hard puzzles.

Trust (checks 23–28)

23. Real reviews or testimonials sit near the decision point, with names and context where you have permission.
24. Proof is specific and genuine: client logos, certifications, press, awards. Never invented.
25. Pricing, or at least clear pricing guidance, is easy to find.
26. Company details are visible: contact options, address or legal entity, and who’s behind the business.
27. Guarantees, returns and cancellation terms are stated plainly, near the call to action.
28. The site looks maintained: no broken links, placeholder text, outdated years or typos on key pages.

Checkout and booking flow (checks 29–34)

For lead generation, read “checkout” as the booking or demo request flow.

29. Guest checkout is available, or the account step is as light as possible.
30. Shipping costs and delivery times appear before checkout, not as a surprise at the last step.
31. Express payment options your customers use (Apple Pay, Google Pay, Shop Pay, PayPal) are offered.
32. The number of steps is minimal and a progress indicator shows where the customer is.
33. The discount code field doesn’t send people off to search for codes and never come back.
34. Abandoned cart, checkout or booking recovery emails are live and tested.

Mobile (checks 35–40)

35. Every key flow has been completed on a real iOS and Android device, not just a resized desktop browser.
36. Tap targets are large and well spaced, and the main call to action or add-to-cart stays reachable (sticky where appropriate).
37. No intrusive popups cover the content on entry.
38. Navigation and search work comfortably with one thumb, and the menu surfaces the most important categories or pages.
39. Text is readable without zooming and nothing causes horizontal scrolling.
40. Mobile conversion rate is compared with desktop by channel. A large gap points to a mobile-specific problem worth investigating first.

How to prioritize: ICE vs PIE

An audit usually produces more fixes than you can ship in a month. A scoring framework stops the loudest opinion from winning.

ICE scores each idea from 1 to 10 on Impact, Confidence and Ease, then averages the three. It was popularised in growth teams by Sean Ellis.

PIE scores each page or area from 1 to 10 on Potential (how much room for improvement), Importance (how much valuable traffic or spend it gets) and Ease, then averages them. It was created by Chris Goward at WiderFunnel.

I use both: PIE to pick the pages, ICE to rank the fixes on those pages. Here’s an illustrative ICE scoring from a typical audit:

The pattern is typical: fixing tracking and removing friction beats a redesign almost every time, because it’s faster and the evidence is stronger.

Fix, ship or test?

Not every finding needs an A/B test:

• Broken things such as bugs, double-firing events and slow pages: fix them now. There’s nothing to test.
• Low-risk best-practice gaps such as missing express checkout or unclear shipping info: ship them and compare before-and-after results.
• Bigger, uncertain bets such as a new offer, pricing layout or page structure: A/B test, but only if the page gets enough conversions. Run a sample size calculation before you start. If the test would take months to reach a reliable result, don’t run it; make a judgement call instead.

After the audit

A sensible sequence is: tracking fixes in week one, quick wins in weeks two to four, then a testing roadmap for the bigger ideas. Re-run the checklist every quarter, and always before a significant increase in ad spend. It’s much cheaper to fix a leak than to pay for the traffic that falls through it.

Get it built

If you’d like a second pair of eyes before you scale spend, I run CRO audits that cover all 40 checks, with a prioritized fix list and hands-on implementation if you want it. See my CRO service, or get in touch.