Your ecommerce site has four conversion rates, not one

Every ecommerce founder we talk to knows their conversion rate. It is usually a number between 1.5% and 3.5%, and it is usually the thing they are trying to fix.

Here is the problem. That number is a lie.

The 2.5% to 3% figure Shopify cites as the ecommerce average is the answer to a single question: what percentage of your site visitors buy something (Shopify). It is a useful headline. It is nearly useless as a target. Because there are actually four conversion rates hiding inside it, each with its own leak, its own fix and its own economics. When brands say "we are going to work on CRO," they usually mean "we are going to redesign something and hope." The brands that consistently improve conversion don't work on CRO. They work on one of the four rates, and they know which one.

Unbundling the conversion rate

The most useful reframe in ecommerce website work comes from Michael Steele, quoted inside Shopify's own CRO guide (Shopify). Split your conversion rate into four sequential rates and treat each one separately.

View product rate. What percentage of your site visitors ever look at a product page. This is largely a homepage, category page and traffic-quality question.

Add-to-cart rate. Of the visitors who see a product page, how many click Add to Cart. This is a product page question. Almost entirely.

ATC-to-checkout rate. Of the visitors who add to cart, how many reach the checkout. This is a cart question, plus a shipping and pricing transparency question.

Checkout-to-purchase rate. Of the visitors who reach checkout, how many actually complete their purchase. This is a checkout flow question, plus a payment options question.

Multiplied together, these four rates give you your headline conversion rate. Improve one of them by 20% without touching the others and your headline number moves. Fix the wrong one, and everything looks the same in the dashboard while the sales team wonders why the redesign did not help.

What breaks at each stage

Each rate leaks for different reasons, and diagnosing the right one is where the actual CRO work lives.

When view product rate is low, the problem is almost always upstream of the product itself. The homepage does not push visitors to specific products fast enough. The category pages are cluttered. The site search does not work well. Or the traffic is wrong for the product mix, which is a media buying issue, not a site issue. Redesigning the product page will not fix this. You have not got the visitor there in the first place.

When ATC rate is low, the product page is the problem. This is where most of the actual sale happens, and it is the page most agencies quietly ignore. The photography is not good enough. The description does not answer the question in the buyer's head. The price feels wrong for the perceived value. There are not enough reviews. The trust signals are weak. This is where the largest reasonable proportion of ecommerce CRO work should live, and where it least often does.

When ATC-to-checkout rate is low, the friction is between the cart and the checkout. The most common cause is not a UX problem. It is a pricing surprise. The shipping cost appears at the last moment. A discount code the buyer expected to work does not. The delivery date is longer than the ad implied. This is the leakiest stage in most ecommerce sites, and the fix is usually cost transparency further up the funnel, not a cart redesign.

When checkout-to-purchase rate is low, the checkout itself is the problem. Too many form fields. Forced account creation. Not enough payment options. A password reset that fails. Any one of these can quietly cost you 10% of the buyers who were literally clicking "buy." Pura, a scent brand cited in Shopify's own guide, lifted its checkout conversion rate by 15% simply by switching to a unified checkout experience (Shopify). That is 15% at the very bottom of the funnel, on buyers who were already committed.

The two things that leak across all four stages

Two issues will quietly drag every one of the four rates down at the same time. They do not get their own stage, but they show up at every stage. And they are almost always underinvested in.

Speed. Google's own guidance is that a page should load in under 2.5 seconds (Google). Very few ecommerce sites in the wild hit this. Every second above the threshold drops conversion at every stage. The visitor bounces from the homepage, does not wait for the product page, abandons the cart, times out at checkout. Speed is the CRO work most brands are unwilling to do because it involves developers, not designers.

Mobile. In the second quarter of 2025, mobile devices accounted for 62.5% of all website traffic (Shopify). Yet most ecommerce sites are still designed on desktop first. The homepage looks brilliant on a MacBook and passable on an iPhone. The product page has a beautiful image gallery on desktop and a broken zoom on mobile. The checkout works flawlessly on Chrome and prompts a keyboard glitch on Safari mobile. If your site is not mobile-first in design and mobile-tested every release cycle, you are ignoring the majority of your buyers.

What people consistently get wrong

The same patterns show up across nearly every ecommerce site we audit.

They treat CRO as a redesign project. Someone briefs an agency for a homepage refresh. Six months and $50k later, the site is prettier and the conversion rate is the same. That is because CRO is never a project. It is an ongoing diagnosis-and-fix loop, applied to specific stages, informed by data. The redesign might make the brand look better. It rarely moves the number.

They optimise the wrong page. The homepage gets 90% of the design attention and drives about 20% of the conversion. The product page gets a template treatment and drives about 60%. The ratio of attention is upside down.

They ignore the checkout because it feels technical. The checkout is where the biggest CRO gains almost always live, because the visitor there is the highest-intent visitor on the entire site. A 1% lift in checkout completion is worth ten times a 1% lift in homepage engagement. But the checkout is boring, is often locked down by the platform, and does not win design awards.

They chase conversion rate instead of revenue per visitor. A 4% conversion rate at a $40 AOV is worse than a 2% conversion rate at $120 AOV. Conversion rate is a component of revenue per visitor. It is not, by itself, the thing to maximise. Bundling, cross-sell, upsell and pricing changes can move revenue per visitor without touching the conversion rate at all.

They forget that ad platform ROAS is downstream of all this. A leaky product page or checkout does not just cost you conversion. It costs you campaign efficiency. Meta and Google are optimising on the buyers they can find. If most of the buyers you send them are dropping off at the ATC-to-checkout stage, the platforms see fewer purchases per click, cost per acquisition rises, and it looks like the ads stopped working. The ads did not stop working. The site did.

Fix the stage, not the site

Ecommerce website optimisation has been sold for a decade as a design and copywriting exercise. It is not. It is a diagnostic exercise followed by a targeted fix. Unbundle the conversion rate. Find the stage that is leaking. Fix that stage. Then look at the next one. The brands that scale profitably in ecommerce are the ones who stopped trying to improve "the site" and started fixing specific rates.

This is the gap we work in. At Dadek Digital we treat the ecommerce site as four sequential conversion problems, not one, and we work on the leakiest stage first. If your headline conversion rate is stuck and the last redesign did not move it, that is almost always because the wrong stage got the attention, and it is fixable.