How to improve your e-commerce conversion rate: a 5-step process

Conversion rate optimization (CRO) is the work of increasing the share of visits that end in an order without buying more traffic and without giving away margin. Your conversion rate is orders divided by sessions: 120 orders from 6,000 sessions is 2%. If you are not yet sure what a reasonable number looks like for your kind of store, start with our guide to conversion rate benchmarks.

CRO is not a list of tricks. Countdown timers, pop-ups and discount banners can move a number for a week and cost you trust for a year. The durable gains come from removing the reasons a willing shopper fails to buy: missing information, surprise costs, doubts about the seller, and friction in the checkout. That is why this guide is a process rather than a list. The right fix depends on where your store loses people, and that is different for every store.

A single conversion rate tells you that something is wrong, never what. Break the journey into steps and measure the share of shoppers who reach each one. Every analytics tool built for e-commerce, including GA4's e-commerce reports and the reports in Shopify and WooCommerce, can show these steps if purchase events are set up correctly.

Then split every step by device (phone, desktop, tablet), traffic source (paid search, paid social, organic, email, direct) and new versus returning visitors. A store with a healthy desktop checkout and a broken mobile one will show an average that looks merely mediocre, and the fix will never be found by looking at the average.

Multiply each step's drop-off by the number of shoppers who reach it. A 30% drop on a step 10,000 people reach is a bigger opportunity than a 60% drop on a step 500 people reach. This is the single most useful calculation in CRO, because it tells you where to look first.

Numbers show where shoppers leave; they never show why. Pair the funnel with qualitative evidence from the same step:

Almost every lost sale traces back to one of four causes. Naming the cause before you choose a fix stops you from adding trust badges to a page whose real problem is a missing size chart.

Write every idea down with three scores from 1 to 5: reach (how many sessions it affects), evidence (how strongly the data from step 2 points to it) and effort (inverted, so easy work scores high). Work from the top. The point is not precision; it is to stop the loudest opinion in the room from deciding the roadmap.

Two rules override the scores. First, fix broken things immediately. A payment method that fails on one browser or a discount code that errors is not something to test. Second, do the obvious, low-risk improvements that research has already settled, such as showing the total price before checkout, without waiting for your own experiment to prove it again.

An A/B test shows two versions of a page to random halves of your traffic and compares the results. It is the most reliable way to know a change worked, but it needs traffic. To detect a 10% relative improvement on a 2% conversion rate (from 2.0% to 2.2%) with the usual 95% confidence and 80% power, you need roughly 80,000 sessions in each version. A store with 20,000 sessions a month would need eight months for one test.

If you do not have that traffic, test bigger changes (a redesigned product page, not a button colour), test on higher-traffic steps such as add-to-cart rather than completed orders, or ship the change and compare the same weeks before and after while watching for seasonality, campaigns and stock changes. Whatever you do, decide the success metric and the duration before you start, run whole weeks, and do not stop a test the day it looks good.

Every store is different, which is the point of the process above. But independent research is consistent about where online stores tend to lose the most willing buyers:

Fake urgency (timers that reset, “only 2 left” that never changes), inflated “was” prices and pre-ticked add-ons can raise a conversion rate briefly and lower repeat purchases permanently. In the EU, a price reduction must also be shown against the lowest price of the previous 30 days, so an invented reference price is a legal risk as well as a trust one.

Do not copy a large retailer's design because it works for them. Large brands convert on recognition and habit that a newer store does not have yet, so a smaller store usually needs more explanation and more proof on the page, not less.