A CRO audit is a structured review of where your site loses the people it already attracts. It starts with the numbers, follows a visitor's path from entry page to follow-up, and ends with a short, ranked list of fixes. The output is a plan you can act on this month.
A useful audit answers three questions: which step loses the most people, why, and what to change first. If your audit ends with forty observations of equal weight, it has not answered the third one.
Below is the process we use on B2B and service sites, in the order we run it. A small site takes a day or two. A larger one takes a week, mostly spent on the data step.
Step 1: Check that the numbers are true
Every finding in an audit rests on the analytics. Before you look at a single page, confirm the data reflects what happens.
- Compare form submissions in analytics with submissions in the CRM or inbox for the same two weeks. A large gap means tracking is broken, blocked or double counting.
- Check what your tracking requests are called. Ad blockers match on words in request paths. Counted on 16 November 2026 across EasyList and related lists, 1,410 rules match
/track, 318 match/events, 292 match/collect, 304 match/pixeland 68 match/popup. A form tool or popup that posts to one of these paths loses part of its data before it reaches you. - Confirm source data survives. Submit a test lead from a URL with UTM parameters and check that the CRM record shows them.
- Note what is not counted. Many sites count leads but not views or starts of the form or tool, so the drop at each step is invisible.
The form tracking guide covers the reconciliation in detail. If the data fails this step, fixing it is finding number one, and every other finding carries a caveat.
Step 2: Map the funnel and find the biggest drop
Write down every path that ends in a lead: the contact form, the demo booking, each lead magnet, each tool or calculator. For each path, record four counts for the last 30 to 90 days:
| Step | Question it answers |
|---|---|
| Views | How many people saw the offer? |
| Starts | How many began the form or tool? |
| Completions | How many finished it? |
| Leads | How many left contact details? |
The step with the steepest drop on your highest-traffic path is where the audit should spend most of its time. A ViralSales check counts views, starts, completions and leads separately for this reason: a single lead count hides which step is failing.
Two patterns appear often. A high view count with few starts points at placement or the promise. A high completion count with few leads points at the ask: the visitor finished, then saw a form they were not willing to fill in. In one client calculator funnel we audited, 59 to 63% of visitors left at a form placed before the result.
Step 3: Walk the path as a visitor
With the biggest drop identified, go through that path yourself on a phone and on a laptop, in a private window. Arrive the way real visitors do: click the actual ad, search the actual query, open the actual email.
Write down, at each screen:
- What the page promised and whether it matched what brought you there.
- What you were asked to do and whether it was clear.
- What you had to give before you got anything.
- What happened after you submitted, and how long until a person replied.
Do this for the top two or three traffic sources, because each source arrives with a different expectation. The CRO checklist is useful here as a prompt list, so you do not miss form details, proof placement or mobile issues.
Step 4: Add what visitors say and do
Your own walkthrough is one opinion. Add evidence from the people who use the page.
- Session recordings of the problem step. Watch 20 or 30 and note where people hesitate, scroll back or leave.
- Form field drop-off, if your tool reports it. One field that loses people is a clear fix.
- Sales call notes. Ask the people who take calls what prospects say they could not find on the site.
- Support and chat logs. Repeated questions are copy the page is missing.
- A one-question survey on the confirmation page: "What nearly stopped you from getting in touch?"
You are looking for agreement between the data and the qualitative evidence. If the numbers show a drop at the form and recordings show people pausing on the phone number field, you have a finding with two sources.
Step 5: Review the follow-up
Many audits stop at the form. A lead is the midpoint of the funnel.
Pull the last 20 leads and trace each one. Who owned it, how long until the first reply, what the reply said, and where it ended. If leads wait days for a reply, or the first reply ignores what the visitor told you in the form, that is a conversion problem as real as a broken button. Settling who owns a lead and when it passes to sales is covered in MQL vs SQL.
Step 6: Rank the findings
Score each finding on three things, using a simple 1 to 3 scale:
- Reach. How many visitors does it affect? A problem on the highest-traffic path scores high.
- Confidence. How sure are you it is causing losses? Two sources of evidence score high; a hunch scores low.
- Effort. How hard is the fix? A copy change is low effort; a new booking system is high.
Then sort into three groups:
- Fix now: broken tracking, broken forms, missing confirmation messages, a reply process with no owner. These are not experiments.
- Test: changes with a good reason but uncertain effect, such as a new headline, a different offer or moving the form after the result. Run them one at a time, as described in A/B testing landing pages.
- Later: small-reach or high-effort items.
Step 7: Write it up for the people who will act on it
Keep the audit report short. Lead with the single biggest drop and its fix. Then list the fix-now items with an owner and a date for each, then the test queue in order. Put screenshots and recordings in an appendix.
Assign each fix to a person by name. An audit with no owners becomes a document nobody opens again. If recurring checks keep getting dropped, a tool like SaaS BPM gives each one an owner and proof it was done. For WordPress sites, WP Commit runs scheduled checks on tracking and speed so the data problems you fixed in step 1 do not quietly return.
Where to start
- Reconcile last month's form submissions in analytics with those in your CRM, and fix any gap before you audit anything else.
- List every lead path with its views, starts, completions and leads, and circle the steepest drop on the busiest path.
- Walk that one path yourself on a phone this week, then trace your last 20 leads through follow-up.