A conversion rate optimization audit finds what is costing you conversions by checking six areas in order: measurement, funnel data, page content, speed and mobile UX, trust and offer, and qualitative research. You then score each finding on impact, confidence and effort, fix the clear problems and A/B test the uncertain ones.
This checklist is part of the A/B testing and conversion optimization hub. It works for a lead generation site, a SaaS trial flow or an online store, and it assumes you already know the basics of conversion rate optimization. Every step names the evidence you need, so each finding can be checked by someone else.
How do you confirm tracking before a conversion rate optimization audit?
Confirm tracking by completing every conversion yourself on desktop and mobile, then checking that each one is recorded exactly once in GA4, in the ad platforms and in the backend. A conversion that fires twice, fires on page load or misses one browser distorts every conversion rate the rest of the audit depends on.
In GA4, the actions you count are key events. Google's help page on key events defines a key event as "an event that measures an action that's particularly important to the success of your business." Check four things per key event:
- It fires once per real conversion. Reloading a thank you page must not add a second one.
- It fires on the right trigger. A submit event that fires on a click without a successful submission inflates every rate.
- It survives consent and redirects. Test with consent accepted and declined, and through any external payment domain.
- It reconciles with the backend. Compare a week of key events with orders or CRM leads.
GA4 and the ad platforms rarely match, because they attribute and date conversions differently. The guide to GA4 conversion discrepancy explains which gaps are normal.
Where in the funnel are visitors dropping off?
Funnel analysis shows where visitors drop off by comparing how many sessions move from one step to the next, split by device, channel and landing page. The largest drop on a high-traffic segment is usually the first place to look, not simply the page with the lowest overall conversion rate.
A funnel is the sequence of steps a visitor takes toward a conversion, such as product page, cart, checkout and purchase. In GA4, the Landing pages report includes a Key events column, and explorations can use the Session key event rate metric, per the key events help page.
Split every funnel at least three ways before you draw a conclusion:
- Device. Mobile and desktop often fail at different steps.
- Channel. Paid search visitors arrive with a specific query, while paid social visitors are usually colder.
- Landing page. A few pages usually carry most of the paid traffic, so a problem there costs the most.
Weight each drop by volume and value: a small drop on the page that receives most of the ad budget outweighs a large drop on a page few people see. Record each drop with its segment and date range as evidence for the page review.
What should a page level review look for on a landing page?
A page level review checks whether a landing page keeps the promise of the ad or query that brought the visitor, states a clear value proposition, offers one obvious next step and keeps forms and checkout short. Review each page the way visitors arrive: on the same device, from the same ad and with the same intent.
Message match means the headline repeats the promise of the ad or search query. Google Ads measures part of this in Quality Score, where landing page experience is defined as "how relevant and useful your landing page is to people who click your ad." The Google Ads audit checklist covers the ad side of that match.
- Value proposition. Above the fold, the page says what the offer is, who it is for and why it beats the alternative.
- Call to action. One primary action, named specifically, with a line on what happens after the click.
- Forms. Every field must earn its place. Baymard Institute's cart abandonment research (as of its October 2026 update) puts the average US checkout at 23.48 form elements by default, against 12 to 14 for an ideal flow.
- Checkout friction. On the same list, 18% of shoppers left because the site wanted them to create an account and 17% because checkout was too long or complicated.
How much do speed and mobile UX hold back conversions?
Speed and mobile UX hold back conversions when pages load slowly, shift while loading or respond late to taps, and Google's Core Web Vitals give a measurable threshold for each. Audit them with field data from real visitors on mobile, because a fast page on office wifi says little about a phone on a train.
Core Web Vitals are Google's three metrics for loading, interactivity and visual stability. According to web.dev (last updated 2024), a good experience means 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 and split by mobile and desktop. Google Search Central's page experience guide states that "Core Web Vitals are used by our ranking systems," so the same fixes help organic landing pages too.
Beyond the metrics, test the mobile page by hand:
- Interstitials and banners. Check whether a cookie banner, chat widget or pop up covers the headline or the form.
- Tap targets and sticky bars. Buttons need room, and a sticky bar must not hide the field a visitor is typing in.
- Input types. Phone and email fields should open the matching keyboard, and autofill should work.
How do you audit trust, proof and the offer?
Audit trust and the offer by checking whether a visitor can see the full price, the terms and the proof before they commit, and whether the risk of saying yes is reduced. Missing costs, missing proof and unclear return or cancellation terms stop visitors who were otherwise ready to convert.
Among US shoppers who abandoned a cart for a reason other than browsing, 40% named extra costs that were too high, 19% did not trust the site with their card details and 12% could not see the total order cost up front, according to Baymard Institute's list of abandonment reasons (October 2026 update). The same Baymard list puts the average documented cart abandonment rate at 70.22%, based on 50 studies published from 2006 through 2025.
Risk reversal is any term that lowers the cost of a wrong decision for the buyer, such as a free trial or a money back guarantee. Check per page:
- Pricing clarity. Shipping, tax, setup fees and contract length are visible before the last step.
- Proof near the decision. Reviews, named customer logos or case studies sit next to the form or the buy button, not only on a separate page.
- Risk reversal. The guarantee, trial or cancellation terms are stated in one sentence on the page itself.
Which research methods explain why visitors leave?
Qualitative research explains why visitors leave: session recordings and heatmaps show where they hesitate, surveys collect their own words, and user tests show them struggling with a task in real time. Analytics shows where a funnel breaks, but only research shows the reason, and the reason is what turns a finding into a testable hypothesis.
Nielsen Norman Group's Usability 101 (updated November 2026) defines usability as "how easily people can successfully complete a task using an interface" and states that "testing with 5 users is often sufficient to identify many of the more important usability issues in a design."
| Method | What it reveals | Typical cost | Use it when |
|---|---|---|---|
| Session recordings | Hesitation, repeated clicks, form abandonment per field | Free with Microsoft Clarity | A funnel step loses many visitors and you do not know why |
| Heatmaps | Where visitors click and how far they scroll | Free with Microsoft Clarity | You suspect key content or the CTA sits below where most visitors stop |
| On-page survey | Objections and missing information, in visitors' words | Low | A page gets traffic but few conversions |
| Moderated user test | Why a task fails, with follow-up questions | Medium, about five participants | Messaging or a complex form is in doubt |
Microsoft calls Clarity "a free service forever." Check that any recording tool fits your consent setup.
How do you score findings and decide what to ship or test?
Score every finding from 1 to 3 on impact, confidence and effort, divide impact times confidence by effort, and sort the list. Then apply the Ship or Test rule: fix bugs and research-confirmed problems directly, and A/B test only the changes whose outcome is uncertain on pages with enough traffic to reach a planned sample size.
- Write each finding as an observation. Name the page, segment, evidence and date range.
- Score impact. 3 for a step that carries most conversions or paid traffic, 1 for a minor page.
- Score confidence. 3 when two sources agree, such as a funnel drop and recordings, 1 for an opinion.
- Score effort. 1 for a copy or setting change, 3 for a rebuild or a developer sprint.
- Calculate the priority. Impact times confidence, divided by effort, gives a score from 0.33 to 9.
- Apply the Ship or Test rule. Ship broken tracking, bugs and confirmed usability problems. Test uncertain changes when the A/B test sample size calculator shows the page can reach the sample in a few weeks. Otherwise ship and compare before and after.
| Audit area | Finding and evidence | Impact | Confidence | Effort | Priority | Action |
|---|---|---|---|---|---|---|
| Measurement | Purchase key event fires again on thank you page reload | 3 | 3 | 1 | 9 | Ship fix |
| Funnel | Mobile paid social sessions drop at the form start step | 3 | 2 | 2 | 3 | Research, then test |
| Page | Headline does not repeat the ad's offer | 2 | 2 | 1 | 4 | A/B test |
| Trust | Shipping cost appears only at the last checkout step | 3 | 3 | 2 | 4.5 | Ship fix |
| Speed | LCP above 2.5 seconds on mobile field data | 2 | 3 | 3 | 2 | Ship fix, planned |
A spreadsheet with these columns is enough for most audits. A/B testing tools help once the backlog holds test ideas.
Common mistakes in a conversion rate optimization audit
The most expensive mistakes in a conversion rate optimization audit are judging pages on broken conversion data, auditing site-wide averages instead of segments, and sending every finding to an A/B test. Each one wastes weeks, because the backlog that comes out of the audit is built on the wrong evidence or the wrong method.
- Skipping the measurement check. A doubled key event makes a weak page look strong. Test every conversion before reading any rate.
- Auditing averages. A site-wide conversion rate hides a broken mobile step. Segment by device, channel and landing page first.
- Reviewing pages on a desktop screen only. The device split in GA4 shows where the traffic really comes from. Review each page on the device most of its visitors use.
- Testing what should be fixed. A broken form field does not need an A/B test. Fix it and save test traffic for uncertain changes.
- Collecting findings without evidence. "The hero looks dated" cannot be scored. Attach the data, recording or survey answer to each finding.
- Treating the audit as a one-off. Pages, ads and traffic mix change. Repeat the audit after a redesign, a tracking change or a shift in channel spend.
