A conversion rate optimization audit is a structured review of where and why a website loses visitors before they convert, ending in a prioritised list of fixes and tests. For a B2B site, a useful CRO audit takes about a week of focused work: one day on goals and tracking, one on segmented data, two on page, form and speed review, one on recordings and buyer feedback, and the last on scoring and the plan.
This post is that process, step by step, with the outputs each day should produce, a worked prioritisation example, and a findings template you can copy. It is not a list of tactics. If you want ideas to fix what the audit finds, our post on B2B conversion rate optimization tactics covers those. If you are auditing an existing personalization programme specifically, use the website personalization audit checklist instead.
Our Position: Most CRO Audits Start in the Wrong Place
Most CRO audits we read start with the homepage and a list of design opinions: the hero is too busy, the CTA should be orange, add social proof. That is a heuristic review, and it is the least reliable part of an audit. It reflects the reviewer's taste more than the visitor's behaviour.
We think a B2B conversion audit should start with two unglamorous questions: is the tracking right, and which visitors are we actually trying to convert? A surprising share of "conversion problems" turn out to be a demo form that fires its event twice, a thank-you page that is indexed and gets organic traffic, or a blended conversion rate dragged down by job seekers reading the careers page. Fix the measurement and segment the traffic first. Opinions about button colours come last, if at all.
What You Need Before You Start
- Admin or editor access to your analytics (usually GA4) and tag manager
- A behaviour analytics tool with at least two weeks of recordings and heatmaps (Microsoft Clarity or similar)
- Read access to the CRM, so you can follow conversions to opportunities
- A list of your three to five most important conversion events
- Thirty minutes each with one salesperson and two recent customers
If any of these are missing, our guide to conversion rate optimization tools covers what each job needs and which tools are free.
The One-Week CRO Audit Process
Day 1, Step 1: Define Goals and the Funnel
Write down the conversion events that matter and rank them. For most B2B sites the list looks like: demo or contact request, pricing page view, signup (if you have self-serve), content download, newsletter signup. Then draw the path for the top one, for example: landing page, product or solution page, pricing page, demo form, thank-you page, meeting booked.
Output: a one-page funnel map with one primary conversion and the pages that lead to it. Everything later in the audit is judged against this.
Day 1, Step 2: Tracking QA
Before trusting a single number, check that it is real.
- Submit each key form yourself, in a private window, and confirm the event fires exactly once in GA4's real-time or DebugView.
- Compare one month of form submissions in analytics with the number of new leads in the CRM from the same form. A gap of more than about 10 to 15% deserves an explanation (consent rejection, spam filtering, duplicate events).
- Check that thank-you pages are not indexed and are not the only signal of a conversion. If someone bookmarks the thank-you page, every return visit counts as a lead.
- Check UTM consistency. "linkedin", "LinkedIn" and "li" as three sources will split your channel data.
- Confirm internal traffic (your own team, agencies) is filtered out.
Output: a list of tracking fixes. Do these first. Some audits end here, with the discovery that the conversion rate everyone worried about was never accurate.
Day 2, Step 3: Segment the Data
A blended B2B conversion rate is close to useless, because the traffic is so mixed. Pull the last 90 days and split the primary conversion rate by:
- Source and medium: paid search, organic, paid social, email, direct, referral
- Landing page: the top 10 entry pages by sessions
- Device: desktop vs mobile. Many B2B demo requests happen on desktop, so a weak mobile rate may be normal rather than a bug See mobile B2B personalization for what mobile visitors need.
- New vs returning visitors
- Company or segment: if you have visitor identification, split by industry, company size or target account list
Look for big gaps between segments with meaningful volume, not small ones. A paid search campaign converting at a third of the rate of organic, on the same landing page, is a finding. A 0.2 percentage point difference between two blog posts is noise.
Output: a table of segments with sessions, conversions and rate, with the three to five biggest gaps highlighted. The B2B marketing KPIs post covers which pipeline metrics to put next to these rates.
Days 3 and 4, Step 4: Page-Level Review
Review two lists of pages: the top 10 by traffic, and the top 5 by intent (pricing, demo, contact, comparison and integration pages). The second list matters more in B2B because these pages see fewer visitors who are much closer to buying.
For each page, answer the same five questions and write the answer down, so the review stays consistent:
- Message match: does the headline match the ad, email or search query that brought most visitors here?
- Clarity: can a visitor tell within five seconds what the product does and who it is for?
- Next step: is there one obvious next action, and does it fit the visitor's stage? A cold visitor from a blog post is rarely ready for "Book a demo".
- Proof: is there evidence relevant to this visitor (their industry, their company size, their use case)?
- Objections: are the common sales objections (price, integration, security, implementation time) answered on the page?
For landing pages specifically, our B2B landing page optimization post goes deeper on what to look for.
Day 4, Step 5: Form Audit
Forms are where B2B conversion most often leaks, and they are the easiest step to measure.
- Track form starts as well as submits, and calculate abandonment per form
- Count the fields. For each one, ask: does sales use this before the first call? If not, remove it or enrich it later
- Check error handling: does the form reject personal email addresses without explanation, or clear the fields on an error?
- Check what happens after submit. Is there a scheduler, a clear message about when sales will reply, or a dead-end thank-you page?
- Submit the form on mobile and time how long it takes
Day 4, Step 6: Speed and Core Web Vitals
Run your top-intent pages through PageSpeed Insights and check field data in Search Console. Google's published thresholds for a good experience are a Largest Contentful Paint within 2.5 seconds, Interaction to Next Paint of 200 milliseconds or less, and Cumulative Layout Shift of 0.1 or less, measured at the 75th percentile of page loads.
In B2B audits the usual culprits are third-party scripts: chat widgets, several analytics tags, testing tools and cookie banners loading before the main content. List every third-party script on the page and who owns it. Scripts nobody claims are candidates for removal.
Day 5, Step 7: Session Recordings
Filter recordings to the pages and segments flagged in steps 3 and 4. Watch 20 to 30 sessions of visitors who reached a high-intent page and did not convert. Note patterns, not individual sessions: the same dead click, the same scroll-back to the pricing table, the same exit after opening the form.
Day 5, Step 8: Voice of the Customer
Talk to one salesperson and two customers who bought in the last six months. Ask the customers what nearly stopped them, what they needed to know before speaking to sales, and which page they remember. Ask the salesperson which questions come up in every first call. Questions that come up in every call are content missing from your site.
Day 6, Step 9: Prioritise With a Scoring Model
By now you will have 20 to 40 findings. Score each one so the backlog is ranked by evidence, not by who argued loudest. Two common models:
- PIE: Potential (how much room for improvement), Importance (how valuable the traffic on that page is), Ease (how simple the fix is). Each scored 1 to 10, then averaged.
- ICE: Impact, Confidence (how strong the evidence is), Ease. Same scoring.
We prefer ICE for B2B audits because the Confidence score forces you to separate findings backed by data from findings backed by opinion. Here is a worked example (illustrative findings and scores):
| Finding | Impact | Confidence | Ease | ICE score |
|---|---|---|---|---|
| Demo form fires the conversion event twice | 8 | 10 | 9 | 9.0 |
| Demo form asks 9 fields; 4 unused by sales | 7 | 7 | 8 | 7.3 |
| Pricing page has no answer to "does it integrate with our CRM?" | 6 | 8 | 7 | 7.0 |
| Paid search landing page headline does not match ad copy | 6 | 6 | 8 | 6.7 |
| Homepage hero "feels busy" | 4 | 3 | 5 | 4.0 |
The tracking fix wins because confidence is total and effort is small. The homepage opinion lands at the bottom, which is where opinions without evidence belong.
Day 7, Step 10: Decide Fix, Test or Research
Split the ranked backlog into three buckets:
- Just fix: broken tracking, bugs, missing answers to common objections, obviously unused form fields. No test needed.
- Test: changes where you genuinely do not know the outcome, on pages with enough traffic. Check the numbers with our A/B test sample size guide before committing.
- Research more: findings with low confidence. Get more recordings, ask more customers, or watch the metric for another month.
The Audit Deliverable: A Findings Template
Keep the deliverable to one spreadsheet and a one-page summary. A 60-slide deck is how audits get admired and then ignored. Use these columns:
| Column | What goes in it |
|---|---|
| ID | F-01, F-02 and so on |
| Page or step | URL or funnel step |
| Segment | All traffic, or the segment where the issue shows |
| Finding | One sentence describing what is wrong |
| Evidence | Analytics number, recording links, customer quote, or "heuristic" |
| Recommendation | The specific change |
| Impact / Confidence / Ease | 1 to 10 each |
| Score | Average of the three |
| Bucket | Fix, test or research |
| Owner and due date | A named person, not a team |
| Result | Filled in later: what happened after the change |
The Result column is the one most teams skip, and it is the one that makes the next audit faster. After two or three audits you will know which kinds of findings actually moved conversions on your site.
After the Audit
- Ship the tracking fixes in the first week, then wait two weeks before reading any conversion rate again.
- Ship the "just fix" bucket and annotate the release dates in analytics, so before-and-after comparisons are possible.
- Start the highest-scoring test. Run one test per page at a time; for when to test several elements at once, see A/B testing vs multivariate testing.
- Re-run steps 2 and 3 every quarter. A full audit once or twice a year is usually enough.
Two places where Markettailor helps in this process. The segment step is much stronger when you can split conversion rates by company, industry or target account rather than only by channel; Markettailor's visitor identification adds that company layer to anonymous traffic. And when the audit shows one segment converting far below the rest, its segmentation lets you build that segment and serve it a different message, measured against a holdout. For broader context on B2B personalization programmes, see our State of B2B Personalization 2026 report, and for a broader B2B testing plan read our B2B A/B testing strategy.
Frequently Asked Questions
What is a conversion rate optimization audit?
A conversion rate optimization audit is a structured review of a website's tracking, traffic segments, key pages, forms, speed and visitor behaviour to find where and why visitors fail to convert. It ends with a prioritised list of fixes, tests and open questions, usually scored with a model such as ICE or PIE.
How long does a CRO audit take?
For a typical B2B site, about a week of focused work: one day on goals and tracking, one on segmented data, two on page, form and speed review, one on recordings and customer conversations, and one on prioritisation. Very large sites with many funnels take longer.
What should a CRO audit checklist include?
At minimum: conversion goals and funnel map, tracking QA, conversion rates by source, landing page, device and company segment, a review of top-traffic and top-intent pages, a form audit, Core Web Vitals, session recordings, customer and sales feedback, and a scored backlog with owners.
What is the difference between ICE and PIE scoring?
Both score findings from 1 to 10 on three factors and average them. PIE uses Potential, Importance and Ease. ICE uses Impact, Confidence and Ease. ICE's Confidence factor rewards findings backed by data over opinions, which is why many teams prefer it for audits.
How often should you run a CRO audit?
Run a full audit once or twice a year, or after a redesign, a major change in traffic mix, or a drop in conversions you cannot explain. Re-check tracking and segmented conversion rates every quarter in between.