Digital Marketing

How to Do a CRO Audit: 50 Checkpoints That Lower Your Conversion Rate

Talha Aslan 17 min read 3 views

Short answer

What is a CRO audit and how do you do one step by step?

A CRO audit is a structured review of one page or one funnel that checks 50 points across tracking, intent match, value proposition, trust, forms, mobile, and speed. You finish with a ranked list that tells you what to fix now, what to test next, and what to leave for later.

Keep reading for the full details and examples

First, a CRO audit replaces guessing with a sequence. Most teams look for one answer to "why does this page not convert?" In practice, lost conversions usually come from several small leaks that add up.

In this guide we count all 50 checkpoints, one by one. Also, each checkpoint is a question plus a short note on how to check it. However, we do not promise numbers or guarantee results. For example, for most items, all you need is a phone, an analytics account, and half a day of focus.

What should you prepare before you start the audit?

Pick one page and one goal first. Because auditing everything at once means finishing nothing. The goal can be a form submission, a phone call, or a purchase.

Four tasks are enough for preparation:

  1. Write down the page you will audit and its single main goal.
  2. Then note sessions and conversions for the last three months.
  3. Calculate the current rate: open the conversion rate calculator and enter sessions and conversions.
  4. Finally, open the page on a phone and on a desktop, and take screenshots.

So these four steps are the "before" photo of your CRO audit. After you make changes, you will look at the same numbers again. However, if the rate looks low, do not panic. The goal is to find leaks, not to find someone to blame. If you are chasing a recent drop, our guide on what to do when your conversion rate drops gives the order of work.

Also agree on ownership. Write down who can access which data, who receives the findings, and who decides at the end. Otherwise everyone likes the findings, and nobody ships them. In short, this small note prevents weeks of confusion.

In which nine layers does a CRO audit move?

Run the CRO audit in nine layers, because each layer answers a different question. The table below shows the layers, the question each one asks, and the checkpoints it covers.

LayerQuestion it asksCheckpoints
TrackingCan we trust the numbers?1 to 6
Intent matchDoes the right visitor reach the right page?7 to 12
Value propositionDoes the visitor understand what you offer?13 to 18
Page structureCan people scan the page easily?19 to 24
TrustDoes the visitor trust you?25 to 30
Forms and checkoutIs the action step free of friction?31 to 37
MobileIs the experience the same on a phone?38 to 42
Speed and technicalIs the page fast and sound?43 to 47
Testing and processCan we keep what we learn?48 to 50

Check the total: 6 plus 6 plus 6 plus 6 plus 6 plus 7 plus 5 plus 5 plus 3 makes exactly 50. Also, the order matters too. If tracking is broken, you cannot trust any later finding, so always start with layer one.

Is your tracking accurate? The first six checkpoints

The tracking layer is the foundation of the CRO audit. Therefore, any improvement built on a broken count goes to the wrong place. First, walk through six points in order.

  1. Is the main conversion defined? First, make sure the goal is measured as an event. Instead, a page view is not a conversion.
  2. Does the conversion event really fire? Submit a test yourself and confirm that the event shows in your reports.
  3. Is anything counted twice? Then check whether one form submission records two conversions.
  4. Is the traffic source split correctly? Confirm that campaign links use consistent tags; a UTM builder keeps this tidy.
  5. Did you exclude your own and bot traffic? Also make sure team visits do not inflate the numbers.
  6. Is the baseline conversion rate on record? Note the rate before the audit so that you can compare afterward.

Skipping this layer is tempting, because the flaws on the page are easier to see. However, a decision built on a broken count is a guess. For a full review of your analytics setup, see is your GA4 setup correct: 50 checks.

Does visitor intent match the page? Checkpoints 7 to 12

This layer asks whether the right person lands on the right page. For instance, the page can be great, but a visitor with the wrong intent still will not convert. Specifically, look at six points.

  1. Does the ad or search promise equal the page headline? This topic deserves its own guide, so read our piece on message match.
  2. Does the headline carry the searched phrase? The words the visitor typed should appear in the headline or the first sentence.
  3. Does every campaign go to the right page? For example, a special offer that lands on a generic homepage is a loss.
  4. Have you compared conversion by traffic source? In practice, organic, paid, and social visitors often react differently.
  5. Do you separate new and returning visitors? A first-time visitor needs trust and explanation, while a returning one needs a fast action.
  6. Does the page serve a single intent? A page with three goals often serves none of them well.

The most common mistake we see in this layer is sending all traffic to one page. In other words, the ad promises one thing and the page says another. As a result, that gap quietly drains budget. So write down which source goes to which page before you do anything else.

Is the value proposition clear on the first screen? Checkpoints 13 to 18

The value proposition is the first answer to "why should I choose you?" Visitors scan before they read, so the answer must sit on the first screen. These six points test it.

  1. Does the first screen say what you offer within seconds? First, show the screen to a friend briefly and ask what they understood.
  2. Does the benefit sentence describe an outcome instead of a feature? Write "your page loads without making people wait" rather than "fast server."
  3. Is the main call to action clear? The button text should say what the visitor gets.
  4. Are the offer and terms on the first screen? The deal from the ad should not hide on the page.
  5. Is your difference from competitors in one sentence? Because if everyone says the same thing, the difference disappears.
  6. Do the visuals support the promise? For example, an unrelated stock photo lowers trust.

In fact, the cheapest way to test this layer is to show the page to someone who does not know it. Their first reaction is more honest than yours, because you already know what the page says.

Therefore, most findings in this layer are content and design decisions. If the page structure, first screen, and mobile flow need a rebuild, working with a team that offers web design services is usually faster than patching items one by one.

Can people scan the page easily? Checkpoints 19 to 24

People scan pages instead of reading them. The structure layer looks at where the eye goes and where it gets stuck. Specifically, six points make scanning easier.

  1. Does the heading hierarchy make sense? Check that H1, H2, and H3 follow the flow of the content.
  2. Are paragraphs short? Because long blocks turn into a wall on mobile.
  3. Does the call to action repeat down the page? So on a long page, one button is not enough.
  4. Are there distracting links? Also, extra exits on a landing page can split attention.
  5. Is the navigation menu simple? On focused pages, a long menu can distract.
  6. Does the page end in a dead end? Then someone who scrolls to the bottom should have a next step.

Structure problems usually come from presentation, not from the text itself. For instance, a good paragraph turns unreadable when it has no heading and no white space. Squint at the page once, because the elements that stand out show up immediately.

Also, to watch scrolling behavior directly, heatmaps and session recordings help. For tool choice, read which CRO tools you need: Clarity, Hotjar, or VWO. For a primer on one of them, see what Microsoft Clarity is.

Does the page earn trust? Checkpoints 25 to 30

A visitor looks for trust before handing money or personal data to a stranger. The trust layer checks whether the page answers that search. Look at six points.

  1. Are references and reviews real? First, never use invented or unverifiable reviews. If you have none, show none.
  2. Is contact information easy to reach? Phone, email, or a form should be one tap away.
  3. Are privacy and cookie notices clear? A short privacy note next to the form supports trust.
  4. Are membership, certificate, and award claims accurate? Instead, show only what you really hold.
  5. Is process, delivery, and return information clear? Because uncertainty grows hesitation.
  6. Do you have an about page and team information? A site that shows the people behind it looks more credible.

Note: trust elements cannot be invented. If you want to add proof, collect real proof first. A fake review or an inflated claim is risky both ethically and under platform policies. In short, a missing proof point beats a fake one.

Do your forms and checkout create friction? Checkpoints 31 to 37

Most lost conversions happen after the decision. However, the visitor is convinced, but the form or checkout tires them out. Seven points look for this friction.

  1. Is the number of fields limited to what you need? Also, every extra field raises the chance of abandonment.
  2. Are labels clear? Use persistent labels instead of placeholders that vanish when people type.
  3. Do error messages guide the visitor? Instead of "invalid input," say what to fix and how; our guide on writing form error messages gives examples.
  4. Does browser autofill work? Make it easier instead of blocking fields.
  5. Does a multi-step form show progress? People should know how many steps remain; for the decision, see multi-step form vs single page form.
  6. Does checkout hide surprise costs? For example, an extra charge on the last step breaks trust.
  7. Is there a thank-you page and follow-up after submission? That page serves both measurement and the visitor's relief.

Google's web.dev forms course points the same way: first use analytics to find the problem form page, then improve it step by step with events and goal funnels. Read it in the web.dev forms course. Also remember that form changes can pull results in two directions, so measure each change before you stack the next one.

Is the mobile experience equally good? Checkpoints 38 to 42

Mobile visitors are the majority on many sites, yet their conversion rate usually differs from desktop. This layer looks for friction on the phone. Check five points on a real device.

  1. Does the first screen show the action on mobile? The headline, offer, and button should sit on screen without scrolling.
  2. Are tap targets big enough? Because tiny buttons cause mistaken taps.
  3. Is the keyboard type right? A phone field should open a numeric keypad, and an email field should open an email keyboard.
  4. Do sticky bars cover content? A fixed menu or cookie banner can hide the button.
  5. Can you fill the form with one hand? Long dropdown lists are a pain on mobile.

In practice, a real device teaches more than any simulation. Also, try more than one phone, because the same layout can break on a different screen size. To test mobile friendliness quickly, you can use the mobile friendly test. For a broader look at why mobile converts less, see the mobile conversion gap.

Do speed and technical problems block conversions? Checkpoints 43 to 47

A slow or broken page wastes even the best copy. For example, a late-loading hero image can push the button out of sight. The technical layer looks for problems the visitor never names but still feels. Check five points.

  1. Are you close to the Core Web Vitals targets? web.dev publishes recommended targets for these three metrics. Because they can change over time, read the current thresholds there and compare them with your field data.
  2. Are images sized correctly? Also, oversized files slow loading.
  3. Do third-party scripts slow the page? Then remove every tag you no longer use.
  4. Are there broken links or long redirect chains? A visitor should never land on a 404.
  5. Does the whole page run on HTTPS? Because mixed content warnings damage trust.

For details, read the web.dev Core Web Vitals article and the Google Search Central page experience guide. The guide states that Core Web Vitals do not guarantee top rankings, yet good scores still make a site more satisfying to use. To set limits, read what a performance budget is.

What does the process layer include? Checkpoints 48 to 50

An audit that sits on a shelf loses its value. Therefore, the process layer turns findings into action and learning. Three points are enough.

  1. Do you keep a hypothesis and test log? Record every change, the reason, and the result.
  2. Did you calculate the sample size for each test? An A/B test calculator shows whether a test can reach a meaningful result.
  3. Does every finding have an owner and a date? A finding without an owner is a finding that never ships.

That completes all 50 points: six for tracking, six for intent, six for value proposition, six for structure, six for trust, seven for forms, five for mobile, five for speed, and three for process.

Which layers matter most for service sites, online stores, and B2B?

All fifty points do not carry the same weight on every site. So set a priority order by site type. The table below is a field-experience starting frame, not a guarantee or a fixed recipe.

Site typeLayers to check firstWhy
Service siteIntent, value proposition, trust, formsVisitors want to trust you before they leave a request
Online storeTracking, forms and checkout, mobile, speedCheckout friction and mobile experience hit sales directly
B2BValue proposition, trust, forms, processDecisions take long and involve several people

For example, a service site has no checkout step, but the phone and form button do the same job. Therefore, instead of checkpoint 36, you look at how visible the contact details are. On B2B sites, visitors rarely decide at once, so process and return paths gain weight.

How do you score a sample page across the 50 points?

First, scoring makes audits comparable. A simple system is enough: give each point a 0, 1, or 2. First, zero means a problem exists. Second, one means partly fine. Third, two means no problem.

The example scenario below is fictional and is not a client case:

  1. Score a fictional service page across the nine layers, point by point.
  2. Divide each layer's total by that layer's maximum score.
  3. Put the two lowest-scoring layers at the top of your fix list.
  4. Repeat the same scoring after your changes.

This method gives a numeric answer to "which layer is weakest?" However, the scores are your judgment, not a measurement. So do not make the score itself a goal. Use it only as a prioritization aid.

Which questions can you ask visitors during the audit?

In short, numbers tell you what happened, and visitors tell you why. Therefore, give part of the audit to real people. Even watching five people try a short task often surfaces surprising findings.

These questions make a good start:

  • What does this page offer, in your own words?
  • What do you see as the next step?
  • What information is missing for you to decide?
  • Did anything make you hesitate?

However, take care not to lead the answers. Also ask the questions after you show the page, not while you show it. The first impression is more honest than the second look.

How do you prioritize the findings?

First, do not try to fix everything at once. Instead, rank each finding by impact and effort. The table below is a simple decision frame.

Finding typeImpactEffortWhat you do
Broken conversion trackingVery highLowFix it now
Missing offer on the first screenHighLowFix it now
Unneeded form fieldsMedium to highLowFix it now and monitor
Headline and promise changeHighMediumTest it
Full page redesignHighHighPlan it and phase it
Small visual tweaksLowLowLeave for last

In this table, "impact" and "effort" are your estimates, not measured values. Still, ranking moves a team from debate to decision. To learn how to test findings, read getting started with A/B testing.

How do you report the audit findings?

Also, keep the report short and action-oriented. A twenty-page document is one nobody ships. Use the same pattern for each finding: what we saw, where we saw it, why it matters, what we suggest, and who will do it.

Give the report three parts:

  • Summary: The five most important findings and the suggested first step.
  • Detail: Findings and screenshots for each of the nine layers.
  • Timeline: What to do now, what to test, and what to plan.

A screenshot or a short recording convinces more than a long explanation. Then everyone who sees the finding understands the same thing. As a result, the meeting moves from "is this true?" to "which one do we do first?"

Which tools do you need for a CRO audit?

However, you do not need an expensive tool stack for an audit. Four functions are enough: analytics, behavior observation, speed measurement, and test math.

  • Analytics: To see conversion events, sources, and the funnel.
  • Behavior observation: To watch real use through heatmaps and recordings.
  • Speed measurement: To see Core Web Vitals values.
  • Test math: To check sample size and significance.

To see your starting numbers quickly, return to the page where you can calculate your conversion rate and enter current sessions and conversions. You will compare them with the numbers after the audit. If you want to run a funnel analysis, read finding conversion leaks with GA4 funnel exploration.

Which findings do you fix directly, and which do you test?

Do not test obvious bugs; fix them. A broken button, a failing form, or a miscounting tag is not a test subject. Instead, these are defects. Save testing for cases where you face two reasonable options.

A simple split works:

  • Fix now: Tracking errors, broken elements, unneeded form fields, and broken mobile layout.
  • Test: Headline text, button text, how you present the offer, and form steps.
  • Observe first: Anything where you do not yet know why people do not convert.

Also, run tests only when you have enough traffic. With low traffic, a test takes long and can mislead. In that case, logging the change and watching the result is more realistic.

How often should you repeat a CRO audit?

Repeat it after a major redesign, after a campaign structure change, or when your conversion rate drops unexpectedly. Beyond that, a full audit every few months is a reasonable rhythm for most sites. There is no exact interval, so decide by how fast your site changes.

A lighter check can run all the time. Review the tracking layer and the speed layer monthly, because those two layers break silently most often. For example, a tag can disappear by accident, and a new script can slow the page.

What mistakes do audits most often contain?

CRO audits tend to go wrong in the same places. So knowing them up front saves time.

  • Starting without verifying tracking: Every analysis on wrong numbers is wasted.
  • Scanning the whole site instead of one page: Scattered findings are findings that never ship.
  • Looking at desktop only: Most visitors are on phones.
  • Skipping prioritization: You cannot fix fifty items at once.
  • Not writing the result down: What you only discuss gets forgotten.
  • Treating the audit as a one-time job: The site changes, so the audit should change too.

The biggest mistake is to see the CRO audit as a problem list. In fact, an audit is the first step of a learning loop. Every finding is followed by a change, a measurement, and a new observation.

When should you get expert help after the audit?

First, you can do most of the checklist yourself. A phone and some time are enough for tracking, form fields, and mobile checks. However, if the findings call for rebuilding the page structure, the first screen, and the mobile flow, the work becomes design and implementation work.

At that point, then, you hold a concrete list, which makes working with a team easier. At Talha Aslan and team, we rank with you which findings truly affect conversions, then we plan how to rebuild the page. We do not guarantee results, because conversions also depend on offer, price, traffic quality, and competition.

If you want to review your audit findings with a specialist and rebuild the page together, you can reach our team through our business website design page. In the first step, we sort out together which items on your list are real blockers.

Core guide: What Is Digital Marketing? Its types, examples and how it is done, on one page with an interactive film that travels through the brain.

Frequently Asked Questions

What is a CRO audit?
A CRO audit is a structured review of a page or funnel that finds what lowers its conversion rate. It covers tracking, intent match, value proposition, page structure, trust, forms, mobile, speed, and process. You finish with a ranked list that shows what to fix now, what to test, and what to postpone.
How do you do a CRO audit?
Pick one page and one goal, record the current conversion rate, and then check the nine layers in order: tracking, intent, value proposition, structure, trust, forms, mobile, speed, and process. Rank the findings by impact and effort, fix obvious bugs, and test the uncertain ones. Report the result briefly with screenshots and a timeline.
How long does a CRO audit take?
The time depends on page complexity and the data you have. For a single page, preparation plus the 50-point check often takes one to a few working days, and longer when analytics setup is broken. We do not promise an exact duration, because it varies with your site and your team.
Which tools do I need for a CRO audit?
Start with web analytics, one behavior tool that offers heatmaps and recordings, a speed measurement, and an A/B test calculator. You do not need an expensive stack. You can also run the mobile checks with a phone. What matters is not the number of tools but whether you turn the findings into changes.
Do I have to A/B test every finding?
No. Fix broken elements, tracking errors, and unneeded form fields without testing. Test choices between two reasonable options, such as headline text, offer presentation, or button copy. Tests need enough traffic, so on low-traffic pages it is more realistic to log each change and watch the result.
How often should I repeat a CRO audit?
Repeat it after a major redesign, after a change in campaign structure, or after an unexpected drop in conversion rate. A full audit every few months suits most sites. Check the tracking and speed layers more often, for example monthly, because they tend to break silently and affect every other number.
  • cro audit
  • conversion rate optimization
  • checklist
  • landing page
  • form optimization
  • mobile conversion
  • page speed
  • ux audit
Share:
Talha Aslan

Google Partner digital marketing expert. Hands-on with SEO, Google Ads, web design and e-commerce projects since 2012; every post here comes from that experience.

Next project

Get a quote for your project.

Your brief goes straight to Talha Aslan and team: strategy led by Talha, delivery by an experienced team. We reply the same day and send the scope, timeline and price as a written quote within 1 business day.