How to Do a CRO Audit: A Step-by-Step Guide (Without the $5k Agency)

A CRO audit is a structured review of a page or funnel to find out why visitors aren't converting, and what to fix first. You compare your numbers against benchmarks, watch how real visitors behave, and rank the problems by impact. That's it. No agency required.

If you're looking for "10 psychological triggers to skyrocket conversions," close the tab. This guide assumes you have a live page, some traffic, and a conversion rate that annoys you. You want a process you can run yourself, this afternoon, and a straight answer on what to fix first.

That's what most CRO audit guides won't give you. The big ones, Glassbox, Matomo, KlientBoost, all walk you through 10 to 12 corporate steps. None of them tell you what to fix first. None give you a time budget. None show current benchmarks to audit against. Some still cite an "average conversion rate" recycled from 2014 data.

I analyzed 521 SaaS pages building pages.report. Same problems, over and over. Not button colors. Broken offers, buried proof, friction nobody measured. This process finds those problems in about four hours.

TL;DR

• A CRO audit finds why visitors don't convert and ranks fixes by impact. You can run one in an afternoon, about 4 hours.
• Benchmark first: median landing page conversion is 6.6% across industries, 3.8% for SaaS, 4.2% for ecommerce (Unbounce, 41K+ pages).
• Free stack: GA4, Microsoft Clarity, PageSpeed Insights, Search Console. $0 total.
• Fix in this order: offer, headline, proof, friction, speed. Button colors last, if ever.
• At Bing, "less than a third" of tested ideas moved their target metrics (Kohavi, 2013). Auditing before testing is how you avoid burning traffic on the losing two-thirds.

What is a CRO audit?

A CRO audit (conversion rate optimization audit) is a systematic review of a website or landing page to identify what's stopping visitors from converting. It combines quantitative data (analytics, benchmarks, speed metrics) with qualitative evidence (session recordings, funnel walkthroughs) to produce a prioritized list of fixes.

The output is not a report. The output is a ranked list: fix this first, this second, ignore the rest.

That last part is where most audits fall apart. Anyone can generate 40 findings. The skill is knowing that finding #3 is worth 10x the other 39 combined. This guide bakes prioritization into the process instead of leaving it as an exercise for the reader.

One boundary before we start. This is the process. If you want a line-by-line inspection list to run against your page, that's the CRO audit checklist. If you're auditing one specific landing page rather than a funnel, the landing page audit walks that narrower job. This post owns the how and the what-order.

Know your benchmarks before you audit anything

You can't diagnose a conversion rate without knowing what normal looks like. Most founders skip this and either panic over a healthy number or celebrate a broken one.

Here are the current numbers worth auditing against, as of August 2026:

Three things to read out of that table.

First, dedicated landing pages convert roughly twice as well as sitewide averages. 6.6% median for landing pages versus 2.74% sitewide for ecommerce. If you're sending paid traffic to your homepage, that gap is your first finding, before you audit anything else.

Second, the spread inside each vertical is huge. Within SaaS, hardware pages hit 4.1% while data and infrastructure sit at 3.3% (Unbounce). Within ecommerce, food and beverage converts at 7.1% while fashion and beauty scrapes 1.3% (Unbounce). Benchmark against your subcategory, not the headline number.

Third, check the dates. The most-quoted conversion "average" in CRO guides comes from a 2014 dataset and has been recycled for over a decade. Audit against 2026 numbers or don't bother.

Ecommerce founders: your audit has an extra layer, checkout. Baymard's running meta-analysis of 50 studies puts average cart abandonment at 70.22%. The ecommerce CRO audit covers that funnel properly. SaaS founders with trial funnels: First Page Sage tracked 86 SaaS companies (their agency clients, updated September 2025). Their data shows opt-in trials converting visitors to trial at 8.5% organic, and trial to paid at 18.2%. The SaaS CRO audit works through that math. Here, we stay on the core process.

The CRO audit process: 7 steps in one afternoon

Total time: about 4 hours for a first pass on one page or one funnel. Not a week. Not a $5k engagement. One focused afternoon with free tools.

The stack: GA4, Microsoft Clarity (free, unlimited traffic), Google PageSpeed Insights, Google Search Console. All free. If you want the full comparison of what each does and what to skip, that's the free CRO audit tools post. Here I'll just tell you what to do with them.

Step 1: Define the one conversion that matters (15 minutes)

Pick one action. Purchase, trial signup, booked call, lead form. One.

Write it down, along with your current conversion rate for it. If you can't compute that rate in five minutes, that's your first finding: your tracking is broken, and fixing it comes before everything else.

Everything downstream gets measured against this one number. Audits that track five "conversions" produce reports nobody acts on.

Step 2: Pull your baseline numbers (45 minutes)

Open GA4 and answer four questions:

1. What's the conversion rate for your target action, overall?
2. What is it split by device? Mobile versus desktop gaps are the most common smoking gun.
3. What is it split by traffic source? Paid, organic, direct behave differently.
4. Where do people exit? Find the single page or step with the biggest drop-off.

Then open Search Console and check what queries actually bring people to the page. If the page ranks for questions it doesn't answer, you've found an intent mismatch, and no on-page tweak fixes that.

Write down the three ugliest numbers. Those are your suspects.

Step 3: Run the speed check (20 minutes)

Put your page through PageSpeed Insights, mobile first. Note the load metrics.

Speed is boring, which is why it's underrated. Google and Deloitte's "Milliseconds Make Millions" study (2020, 37 brands, 30M+ sessions) is dated, but still the best data we have. It found a 0.1 second mobile speed improvement lifted retail conversions 8.4%. Lead-gen form-page progression rose 21.6%. A tenth of a second.

Portent's 2022 analysis of 100M+ pageviews found ecommerce conversion at 3.05% for pages loading in 1 second, versus 0.67% at 4 seconds. That's a 4.5x difference for doing nothing except being slow.

If your mobile load is past 3 seconds, put speed near the top of the fix list regardless of what else you find.

Step 4: Watch 20 real sessions (60 minutes)

Install Microsoft Clarity if you haven't. It's free with unlimited traffic, and it changes the audit from theory to evidence.

Watch 20 recordings of visitors on your target page. Not skimming. Watching. Take notes on:

• Where they stop scrolling. Everything below that line doesn't exist.
• What they click that isn't clickable. That's a broken expectation.
• Rage clicks and dead clicks. Clarity flags these for you.
• Where mobile users pinch and zoom. That's a legibility failure.
• Where they hesitate before the form or the checkout, then leave.

This is the highest-value hour of the whole audit. Analytics tells you where people leave. Recordings tell you why. You will see something in the first ten sessions that no dashboard would ever show you.

Step 5: Walk your own funnel as a stranger (30 minutes)

Open an incognito window on your phone. Go through the entire journey: ad or search result, landing page, form, confirmation. Pay for your own product if you have to.

Count every ask. Every field, every click, every decision point. Then ask the only friction question that matters: is the page asking for something the visitor hasn't agreed to give yet?

I learned this one the expensive way. When I launched pages.report, conversion was bad. I rewrote the hero, tested headlines, changed the color scheme. Nothing moved. Then I swapped the magic-link signup button for "Continue with Google." One change. Login rate tripled. The problem was never the copy. It was asking for an email from someone who hadn't decided they wanted the product yet.

Your funnel has one of these. The walkthrough is how you find it.

Step 6: Audit the page itself (45 minutes)

Now, and only now, look at the page content. Four layers, in order:

1. Offer. Would a stranger understand what they get, for how much, in 5 seconds? Is the offer actually worth the ask?
2. Headline. Does it name the outcome and the audience? Or could it sit on a competitor's page unchanged?
3. Proof. Real numbers, named customers, specific outcomes? Or "trusted by teams everywhere"?
4. Call to action. One action, repeated? Or five competing links splitting intent?

Notice what's not on the list. Fonts, gradients, animations, button colors. Those are layer nine, and audits that start there produce redesigns, not conversions.

For the full line-by-line version of this step, run the CRO audit checklist against your page. For a single-page deep pass, use the landing page audit process. And if you want to see what this layer looks like done on real pages, the teardowns are exactly this step, annotated section by section.

Step 7: Build the fix list and score it (30 minutes)

Dump every finding from steps 2 through 6 into one list. Aim for 10 to 20 items, not 50. Then score each one on two axes:

• Impact: if fixed, how much could it plausibly move the one conversion from step 1? High, medium, low.
• Effort: hours, days, or weeks to ship?

That scoring is the whole point of the audit, and it deserves its own section.

What to fix first: the impact-vs-effort rubric

Most CRO advice fails here. It hands you 40 findings and wishes you luck. Here's the rubric I use, and the order almost always comes out the same.

Quadrant 1: high impact, low effort. Do these this week. Offer clarity, headline rewrite, cutting form fields, removing an unearned ask (like my magic-link button), fixing a broken mobile layout, adding real proof you already have but never published. These are copy and friction changes. Hours, not sprints.

Quadrant 2: high impact, high effort. Schedule these. Restructuring the page, rebuilding a slow template, building a dedicated landing page instead of dumping paid traffic on the homepage, creating proof you don't have yet. Worth doing. Not first.

Quadrant 3: low impact, low effort. Batch these. Button copy, image swaps, minor spacing fixes. Fine to do in an idle hour. Never let them jump the queue, because they feel productive and change nothing.

Quadrant 4: low impact, high effort. Delete these. Redesigns for taste. New brand colors. That animation your designer loves. If it doesn't touch offer, clarity, proof, friction, or speed, it's not a conversion project. It's decoration.

The pattern behind the rubric: fix the bones before the paint. Offer before headline. Headline before proof. Proof before friction. Friction before speed. Speed before anything cosmetic. A page with a broken offer converts badly no matter how polished it is. A page with a great offer converts despite ugly buttons.

Here's why prioritizing beats testing your way out. Ron Kohavi, who ran experimentation at Microsoft, reported in 2013 that at Bing "less than a third" of tested ideas moved the metrics they were designed to improve. That's Microsoft, with world-class infrastructure and oceans of traffic. Dated, yes. Still the canonical number, and nobody has published better.

If most ideas fail even there, a solo founder with 3,000 monthly visitors can't afford to test guesses. The audit exists to make sure the few tests you can run are aimed at quadrant 1 and 2 problems. When you're ready for that stage, here's how to run A/B tests on landing pages without fooling yourself. But testing is what comes after the audit. It's not a substitute for one.

DIY vs paid CRO audit: the honest math

Should you pay someone for this? Honest answer: usually not first. Sometimes yes. Here's the actual math.

DIY: $0 and one afternoon. Everything above runs on free tools. For a first audit, DIY wins on one huge dimension: nobody knows your customers, your offer, or your margins like you do. An outsider can spot friction. They can't tell you whether your offer is right.

DIY has two real weaknesses. First, you can't see your own page with stranger eyes. You wrote the headline. It's obvious to you. That's exactly the problem. Second, you don't have a pattern library. If this is the first funnel you've audited, every finding looks equally important, and the prioritization step gets shaky.

Agency audit: four figures and weeks of waiting. Worth it when real money flows through the funnel, think five figures a month in paid spend, and you need research depth: user testing panels, full-funnel analytics rebuilds, testing roadmaps. Overkill for a single landing page or an early-stage product. You'll pay for a 60-page deck when what you needed was five fixes.

The middle: a fixed-price human teardown. Full disclosure, this is what I sell. I do these by hand for $99 at pages.report: a recorded video audit of your page, delivered in 48 hours, running essentially the process in this post plus the pattern library from 520+ annotated pages. It exists for one reason: stranger eyes plus pattern-matching, without agency pricing.

My honest recommendation, and it costs me sales: run the DIY audit first. It's free, it takes an afternoon, and half of you will find your quadrant 1 problem in the Clarity recordings without paying anyone. Pay for outside eyes when you've fixed the obvious things and the number still won't move.

FAQ

What does a CRO audit include?

A CRO audit is a structured review of a page or funnel that identifies why visitors aren't converting and ranks the fixes by expected impact. It combines analytics data, session recordings, speed metrics, and a manual page review. The deliverable is a prioritized fix list, not a findings report.

How long does a CRO audit take?

A solid first pass on one page or funnel takes about 4 hours using the seven steps above. Agency audits run 2 to 6 weeks because they cover multi-page funnels, user research, and testing roadmaps. For a solo founder, the afternoon version captures most of the value.

How much does a CRO audit cost?

DIY costs nothing: GA4, Microsoft Clarity, PageSpeed Insights, and Search Console are all free. Agency audits run well into four figures. Fixed-price human teardowns sit in between; mine is $99. Start with the free version and buy outside eyes only when you're stuck.

What's a good conversion rate to audit against?

As of August 2026: the median landing page converts at 6.6% across industries, 3.8% for SaaS, and 4.2% for ecommerce, per Unbounce's benchmark report covering 41K+ pages. Sitewide ecommerce sits at 2.74% per Dynamic Yield. Compare against your subcategory and your own baseline, not a single global average.

How often should you run a CRO audit?

Quarterly for a page that matters, and immediately after any major change: new pricing, new positioning, a redesign, or a new traffic source. Between audits, keep Clarity running so the next one starts with evidence instead of guesses. The 30-check list makes repeat passes faster.

Do you need special tools for a CRO audit?

No. The four free ones, GA4, Microsoft Clarity, PageSpeed Insights, and Search Console, cover a full first audit. Paid tools add convenience, not findings, at small scale. The CRO audit tools post breaks down what each one is actually for and what you can skip.

Run it this week

The whole process: one conversion, baseline numbers, speed check, 20 recordings, one incognito walkthrough, a four-layer page review, a scored fix list. Four hours. Free tools.

Then fix quadrant 1 and nothing else for two weeks. Offer, headline, proof, friction. Not button colors.

Most founders reading this will bookmark it and redesign their hero instead. Don't be most founders. Block the afternoon, run the seven steps, and let the recordings tell you what your page is actually doing to people.

Go find your magic-link button.

– Luke