Online store conversion rate: how to calculate it and what counts as good

1. Calculate the rate on a single example.

The basic formula is: conversion rate = number of orders / number of sessions × 100. A session is a single visit to the store. If the store had 2,500 sessions and 50 orders in a completed week, the result is 2%. In other words, on average one visit in fifty ended with a purchase. This is a worked example, not the result of a real store.

Before you record the result, decide what the numerator is. Placed, paid and fulfilled orders are different sets. If you count all placed orders one time and only paid ones the next, the rate will change without any change in customer behavior. Pick one definition, write it down next to the result and use it in every later comparison. The same goes for filters: team visits, test orders and other stores shouldn't end up in any part of the formula.

2. Check what you're dividing by.

The same number of orders gives different results depending on the denominator. Analytics tools calculate the rate from sessions, from users or from views of a given step. Google Analytics 4, for example, has separate metrics for sessions (Session key event rate) and for users (User key event rate). One person may visit the store three times before buying, so a rate based on users usually comes out higher than one based on sessions.

The numerator can differ too. A rate calculated as the share of sessions with a purchase won't count a second order placed during the same visit, while a rate based on the number of orders will. When comparing two reports, check both parts of the formula first. The gap between 2% and 2.8% may come purely from a different calculation method, not from worse or better sales.

3. Read public averages with caveats.

The question "what is a good conversion rate" has a short answer: it depends on the industry, price, market and traffic sources. Even data from a single source shows this. IRP Commerce publishes monthly benchmarks for stores running on its platform in the UK and Ireland. In September 2026 the average there was 1.97% transactions per session, compared with 1.68% a year earlier. The spread between categories was more than tenfold: from 0.44% in baby products to 5.76% in crafts and art supplies.

Don't apply such an average directly to your own store. It's a different market, a single store platform and a figure that changes every month. An average across many stores also mixes cheap products bought on impulse with expensive ones that customers compare for weeks. A public benchmark is useful for one question: is my result very far from the range for a similar category? If so, check your measurement first, and only then your offer and checkout process.

4. Set your own baseline.

The best benchmark is your own store. Take the last 8 to 12 completed weeks and calculate the rate for each one separately. Mark weeks with promotions, newsletter sends, stockouts and price changes. Record the range for typical weeks and the median. If your store has clear seasonality, add the same period from the previous year.

With a small number of orders, the rate fluctuates a lot. Worked example: with 1,000 sessions and 20 orders it's 2%, and one extra order gives 2.1%. That's why you should judge weeks or months, not single days, and always record the number of orders next to the percentage. Set targets against your baseline, e.g. "get back to the typical-week range after changing the checkout form", not against someone else's average. If you use Flowpeek, you can save such a target in the eCommerce goals section, which compares it with the result for the last 30 days.

5. Break the result down by device and source.

The store-wide average can change even though no group of customers behaves differently. Worked example: in the first week, desktops have 1,000 sessions and 30 orders (3%), and phones have 1,000 sessions and 10 orders (1%). Together that's 40 orders from 2,000 sessions, or 2%. In the second week, a campaign brings an extra 1,000 sessions from phones. Phones still convert at 1% and desktops at 3%, yet the overall result drops to 50 / 3,000, or 1.67%, even though there are more orders.

That's why you should break the rate down at least by device and main traffic sources. A drop in the total while segments stay stable means a change in the traffic mix, not a problem in the store. A drop in one segment, e.g. only on phones using one browser, points to a specific place to check. In the Flowpeek funnel report, you see conversion by device in a single table together with the number of views, carts and purchases.

6. Before you start optimizing conversion, find the cause.

A sudden drop in the rate often starts outside the checkout process. First check the measurement: are orders still being recorded, and has part of the traffic disappeared from the report? Then traffic quality: a new campaign can bring in many people who are only just discovering your offer. Next, the offer itself: prices, bestseller availability and shipping costs. Only when all of these are in order should you look for a problem in the later steps of the purchase.

Also remember that conversion isn't a goal in itself. A big discount can raise it and lower the store's bottom line at the same time. Next to the rate, look at revenue per session, i.e. conversion multiplied by average order value. Worked example: 2% × PLN 200 gives PLN 4 of revenue per visit, while 2.4% × PLN 150 gives PLN 3.60. A higher conversion rate doesn't always mean a better result.

• Measurement: are orders and sessions counted the same way as before?
• Traffic: has the share of campaigns, devices or new visitors changed?
• Offer: have prices, availability or shipping costs changed?
• Purchase: at which funnel step do you see the difference?