What is a good ecommerce conversion rate?

Take one store and split its sessions by where they came from. Someone typing the brand name into a search engine has already chosen; someone who tapped an advertisement while scrolling has not. Those two populations can differ by an order of magnitude in how often they buy, and a store-wide figure sits somewhere between them, describing neither.

The same is true across templates. A product page carries a buying decision; a category page carries a browsing decision; an article carries a reading decision. Reporting them together produces a number that moves whenever the traffic mix moves, which happens every time you change a campaign, publish content or gain a ranking.

That is the real problem with benchmarks. A published figure encodes somebody else's traffic mix, price band, category, country and device split. Comparing your number to it tells you how different your mix is, not how well your site is doing.

So the useful question is not whether your rate is good. It is which segment of your traffic is underperforming its own history, and what changed.

Four cuts do most of the work. Source and medium, because intent varies more by channel than by anything else. Device, because the mobile experience is a different product from the desktop one. Landing page template, because the job of the page differs. And new against returning, because returning visitors already trust you.

Build those into a saved report so you look at the same view each time, rather than re-slicing the data whenever a number looks odd. Consistency is what turns a metric into a signal.

Then add price band if your catalogue is wide. A store selling both a small accessory and a large item will see the blended rate move purely with the mix of what people happened to be looking at that week, and reading that as a site problem leads to changes that fix nothing.

Category and season sit underneath all of it. A product bought once a decade converts differently from a consumable, and a category with a concentrated buying season has weeks that look poor and weeks that look excellent for reasons unrelated to the site.

Two systems will give you two different rates for the same store, and both can be right. Your ecommerce platform typically divides orders by sessions using its own definition of a session. An analytics tool divides purchase events by its own sessions, which start and end on different rules and can miss events that consent or a blocker prevented.

Decide which one governs and use it consistently. For most stores the platform figure is the better one for trend reporting because it is closer to the orders, and the analytics figure is the better one for segmentation because it can slice by source, device and page.

Then watch the definition of the denominator. Sessions and users produce very different rates, and a store that quietly switches from one to the other will appear to have improved or collapsed overnight.

Write the definition down alongside the number in whatever report the team reads. Most arguments about conversion rate turn out to be arguments about which denominator someone used.

Conversion rate can be improved by doing things that lose money. Deep discounting raises it. Removing higher-priced products raises it. Aggressive popups can raise it while damaging repeat purchase. In each case the ratio improves and the business does not.

Revenue per session multiplies conversion rate by average order value, so it catches the trade-off. A change that lifts conversion while reducing basket size shows up as flat or worse, which is the honest reading.

Use it as the primary measure and keep conversion rate as a diagnostic underneath it. When revenue per session moves, the segmented conversion view tells you where, and average order value tells you whether the movement came from more buyers or bigger baskets.

For the whole picture, add contribution per session once you know your margin by product. That is the number a business decision should be made on, and it is the one that stops a successful test from being an expensive one.