E-commerce CRO audit: the method in 7 steps

A CRO (conversion rate optimization) audit examines a store's pages to find what holds back the purchase. It is not an SEO audit, which is about bringing visitors in, nor a general opinion on design: it is a list of obstacles.

A good audit shows three things. Each obstacle is located: you know which page, which device and where on the screen. Each obstacle is evidenced: a screenshot, an exact sentence, a measurement in the code that you can check. And the whole is prioritised: what costs the most sales comes first.

If one of the three is missing, you have a list of opinions, not an action plan.

A visitor doesn't browse your whole site. They come in through a few pages and decide there. Start with the ones that carry the purchase:

• the homepage, which must say in one screen what you sell and to whom;
• a collection page, where the choice between products happens;
• your best-selling product page, where the add-to-cart decision is made;
• the cart and checkout, where buyers who had decided drop off.

Four well-chosen pages of one site say more than a skim of fifty. The cart and checkout usually can't be reached by an address: audit them from screenshots.

The same page is not the same experience on a computer and on a phone. On a 1440 × 900 screen the buy button is often visible on arrival; on a 390 × 844 phone it can end up two screens down, under an image or a banner.

So you need two full captures per page, one per device, and to read each on its own terms. What sits above the fold (what the visitor sees without scrolling) deserves particular attention: that is where the decision to stay is made.

To judge the whole, weight by your real traffic. If 70% of your visits come from mobile, a mobile obstacle counts for more than a desktop one of the same severity.

The first screen must answer three questions without scrolling: what is sold, for whom, and why buy here rather than elsewhere. Check:

• that the main headline states a benefit, not just the brand name ("Welcome to …" says nothing);
• that the promise couldn't be signed by any competitor;
• that the visual shows the product or its use rather than a mood;
• that there is a single main headline (one H1 tag) and that it carries that promise.

A buy button must be reachable without hunting for it. On every page, look at:

• whether it is visible on the first screen, especially on mobile;
• whether it stands out clearly from the rest of the page by colour and contrast;
• whether its label states the action ("Add to cart") rather than something vague;
• whether it stands alone: three buttons of equal weight side by side compete with each other;
• on a long product page, whether a button stays within reach at the bottom of the screen while scrolling the gallery.

A visitor who doesn't know you looks for reasons to trust you before paying. Check that these exist and are visible at the moment of decision, not only in the footer:

• reviews and the rating, near the price rather than at the bottom of the page;
• shipping and returns terms, with their amounts and times;
• the accepted payment methods and the mention of secure payment;
• a way to contact you;
• real photos of the product.

A signal present in the code but invisible on screen (a review badge loaded at the bottom of the page, for instance) reassures nobody at the moment of clicking.

Friction is anything that slows a buyer down or makes them hesitate: pop-ups on arrival, confusing navigation, search hidden behind an icon, forms that are too long, missing information (size, delivery time, shipping cost) at the moment of choosing.

Visual hierarchy is the path the eye follows. It should lead from the promise to the button. Look for what pulls attention away: an image filling the screen and saying nothing, a promotional banner more visible than the product, elements of equal weight competing.

On mobile, also check tap-target size, legibility without zooming and the absence of horizontal scrolling.

A screenshot shows what the visitor sees; the code shows what the page really is. The two complement each other. A few simple measurements, and why they matter for conversion:

• server response time and page weight: a slow page loses visitors before it even displays;
• forms: number of fields, fields with no label, no autofill, email and phone fields typed as plain text on mobile;
• images with no alt text and buttons with no name, which get in the way of screen readers;
• structured product data (price, availability, rating), which feeds rich results;
• the third-party scripts the page loads (chat, reviews, tracking) and their weight.

For the speed your real visitors experience, Google publishes thresholds for the Core Web Vitals: the main content displayed (LCP) within 2.5 seconds, a response to interactions (INP) within 200 milliseconds, and a layout shift (CLS) of 0.1 or less, measured at the 75th percentile.

An audit listing forty points in no order is no use on Monday morning. Rank each obstacle by three criteria: its severity (does it prevent the purchase or only slow it down?), the confidence you have in the finding, and the traffic share of the device concerned.

Then separate quick fixes, done in a few days, from deeper work. A 30, 60 and 90-day plan gives each its place.

Finally, don't add up the gains each fix promises: they overlap. Several improvements on the same page don't multiply sales.

An amount only makes sense from your own figures: visits, conversion rate, average order value. Without them, speak in percentages and say so.

If you want to verify a fix with an A/B test, work out its duration first: at 95% confidence and 80% power, a store with modest traffic can need several months to reach a conclusion. Beyond 90 days, you are better off fixing and measuring before/after.

Everything above can be done by hand, with a phone, a computer and some method. Allow several hours per page for serious work.

OrderLeak applies exactly this method to your pages in 15 to 20 minutes: desktop and mobile captures, six axes scored out of 100, measurements from the code, every obstacle located with its evidence, a 30, 60 and 90-day plan. You can read a complete report before buying.