Online store conversion rate: measure and compare correctly

Conversion rate rises while revenue falls. Both can be true: fewer visitors with stronger purchase intent can produce a higher rate without more sales overall. We therefore interpret online store conversion rate alongside its calculation and commercial outcome.

Which conversion rate do you mean?

One possible metric divides sessions containing at least one purchase by all observed sessions, then multiplies by one hundred. Another uses orders per session. These differ when a session contains multiple orders.

In a hypothetical example, 20 purchasing sessions among 1,000 sessions give two percent. If they contain 22 orders, orders per session equal 2.2 percent. Both calculations can be useful when labelled accurately; they must not be compared as though identical.

Are the underlying events reliable?

Google Analytics documents ecommerce events for product views, carts, checkout, purchases and refunds. They need appropriate implementation; installing an analytics tool does not prove complete purchase data. Source: GA4 ecommerce documentation.

Compare a clearly defined period with the store’s back office. Measurement limitations mean exact equality is not always expected, but substantial differences should be explainable. Check whether reloading confirmation or returning from payment records a purchase more than once.

What belongs in a useful measurement record?

We establish six points before assessing a trend:

• Goal: which action counts as a conversion?
• Denominator: are sessions, users or another unit being measured?
• Period: which days and comparison periods are included?
• Quality: what known measurement gaps or changes exist?
• Segment: which devices, sources and customer groups are compared?
• Business outcome: how do orders, revenue and, where available, contribution margin develop?

These details prevent arguments over figures that measure different things. Separate German and Swiss reporting may help where payment, shipping or currency conditions differ.

Why can the overall figure mislead?

A broad information campaign may bring many new visitors. The overall rate can fall even when checkout works as before and additional orders are generated. Ending a campaign can instead raise the rate while reach and sales decline.

Compare appropriate groups. Mobile and desktop, new and returning visitors or individual sources may serve different purposes. Avoid subdividing small samples until single purchases dominate the result. Label inconclusive groups accordingly.

How does a figure become an optimisation task?

An unusual rate starts an investigation. Determine whether the issue concerns unsuitable products, selection or checkout. Abandonment after shipping details calls for different investigation from an unresponsive button.

Use our checkout guide for specific purchase-journey problems. This article establishes the measurement needed to assess change. Record implementation dates, affected pages and expected behaviour.

How do we approach this in PrestaShop?

PrestaShop lets us align individual workflows with their measurement. We inspect the actual combination of theme, modules and tracking rather than assuming a universally correct default. Technical changes are also checked through representative test orders.

As a PrestaShop agency, we want decisions grounded in traceable data. We do not promise a universal target rate. Improvements should be plausible within comparable visitor groups and support the merchant’s commercial objective.