Ecommerce KPIs: the few numbers worth watching

This guide is about shops. For websites that exist to bring enquiries, see what to measure on a business website.

Say you run an online shop selling kitchenware. All figures here are invented for the illustration. Assume there are no returns or other post-order adjustments in either month, so average order value multiplied by orders equals net sales in this example.

In March you get 10,000 visits and 200 orders, at an average order value of 60. Net sales are 12,000. Your gross margin on products is 40%, and you spend 500 on advertising.

In April you run a paid campaign. Visits rise to 14,000. Orders rise to 238, but the new visitors buy cheaper items, and the average order falls to 55. Net sales are 13,090. You spend 2,500 on advertising.

On the sales line, April looks better. In the business, it is worse.

If you had watched only sales, you would have kept spending. Conversion rate and advertising per order show the cost.

Shopify defines orders as the "number of orders placed" and net sales as "sales revenue, with discounts and sales reversals factored in". Gross sales, by contrast, are "before discounts and sales reversals". Watch net sales, because a month of heavy discounts and returns can have a good gross figure and a poor net one.

Compare each month with the same month last year, not only with the previous month, because most shops have seasons.

Shopify defines conversion rate as the "percentage of online store visits (sessions) that resulted in a sale." Other tools count differently, so keep to one source.

Conversion rate is the most quoted shop number and the easiest to misuse. Three habits help.

Read it in pieces. Phone against desktop, new visitors against returning ones, one traffic source against another. A single average hides which part is weak. A better ad that brings in more first-time browsers can lower the average while the shop is healthier.

Compare it with your own history. We give no industry benchmarks here. We could not trace a figure to a source we could stand behind, and shops differ too much: a shop selling 15 gifts and a shop selling made-to-order furniture have different buying cycles.

Remember that many visitors were never going to buy. Baymard Institute's survey found that 42% of US online shoppers had abandoned a cart because they were "just browsing / not ready to buy". Conversion rate counts them as a miss, though they might return, and they should not push you to chase a better number.

Shopify defines it as the "average value of orders placed, excluding any post-order adjustments." Shopify calculates it as gross sales minus discounts, divided by orders, excluding post-order adjustments such as returns. It matters because it limits what you can spend to win an order. A shop with an average order of 20 has little room to pay for delivery and advertising. Raising it with bundles, free delivery above a threshold or suggested accessories is one of the few levers that does not need more traffic. Be careful, though: a rising average order value can also mean you lost your small orders.

A funnel is the path from visit to purchase. Shopify reports an "added to cart rate" (visits where items were added to the cart, relative to total visits) and a "reached checkout rate" (visits that reached checkout, relative to total visits), next to the conversion rate. Google Analytics 4 records the same steps with the recommended events view_item, add_to_cart, begin_checkout and purchase.

These Shopify rates share all visits as their denominator; they are not step-to-step conversion rates. Use a consistently defined sequential funnel to investigate where people stop, and check whether each step is recorded correctly.

Few visits reach a product page. Look at your category pages and navigation. Our guide to category pages covers it.

Visits reach products but few add to cart. Look at the product page: photos, total cost, delivery date, returns. See what makes a product page useful.

Carts fill but few finish. Look at checkout: costs that appear late, account walls, long forms. See checkout and abandoned carts.

A large drop is a reason to investigate, not proof of the best change. Check the step definition, visitor intent, errors and likely revenue impact before choosing what to fix.

Shopify defines the returning customer rate as the "percentage of returning customers relative to all customers who placed orders." For a shop that sells things people reorder (food, cosmetics, supplies), it is the best sign of a business and not just a campaign. For a shop that sells something bought once a decade, it means little. Choose it only if repeat buying is possible, and then watch whether it moves after a change to delivery, packaging, or follow-up emails.

For cost per attributed order, divide the advertising and promotion spend by the orders credited to that activity. For customer acquisition cost, divide the eligible acquisition costs by new customers acquired. Name which costs, period and attribution method you use. Repeat-customer orders belong in an order count, not a new-customer count; channel attribution is imperfect. Shopify counts "new customers" as the "number of new customers who placed an order". Do it per channel if you can: a search advertisement and an influencer post can have very different costs. Without this number, a campaign can look successful while it loses money, as in the April table.

None of these numbers includes your costs of goods, shipping, payment fees, returns and refunds, or time. Analytics tools do not know them. Work out, once a month, what is left after products, delivery, payment processing, advertising and refunds. Even a rough figure from a spreadsheet beats none. Add the return and refund rate as its own line: a product with many returns can show strong sales and weak profit.

Total visits. They show attention, not trade. See the generic version of this point.

Cart abandonment rate on its own. Baymard reports a 70.22% average documented rate across 50 studies, but part of that is people browsing. It is a clue about a step to inspect, not a score to compare.

Bounce rate. In Google Analytics 4 it is the opposite of engagement rate, and it says little about whether a shopper bought.

Social followers. Useful for social media work, not a shop result.

Any "average ecommerce conversion rate". Published averages mix very different shops. We do not quote one because we cannot tie one to a source we trust, and your own trend is more useful.

Pick one source for each number and keep it. Shopify's help center notes that changes in how sessions are measured, including how bot sessions are filtered, can change the numbers without any change in shopper behavior. Differences between tools are normal. Switching tools mid-year breaks the comparison.

Record the funnel events. Check that add to cart, begin checkout and purchase fire once per action, on the phone as well as the desktop. A test order is the quickest check.

Record the order, not the click. Record the purchase once the order is successfully placed, using the platform’s order record and a unique transaction ID to prevent duplicate purchase events. Payment capture can happen later, for example for payment on delivery.

Keep customer details out of analytics. Names, email addresses and phone numbers belong in your shop records, not in analytics events or page addresses.

Write each definition down. "Conversion rate means orders divided by sessions from Shopify's report" is a sentence worth saving, so next year's number means the same thing.

Our analytics and measurement service covers the setup.

They will not tell you why. A drop in conversion could be a slow page, a stock problem, a competitor's sale or a bank holiday. They show where to look. Small shops also see noisy numbers: six orders one week and two the next may be chance, so compare three-month totals before you decide anything. And no tool sees everything: some customers see a product on your site, then buy by phone or in person.

If you want your shop set up to record the right steps and report them simply, our analytics and measurement service explains how we approach it. For the shop itself, see our ecommerce website service. When you want to talk about your own shop, you can discuss your project. More guides are in the Learn library.

1. Shopify Help Center, Analytics fields (definitions of conversion rate, sessions, average order value, returning customer rate, gross and net sales, orders, added to cart rate, reached checkout rate, new customers): https://help.shopify.com/en/manual/reports-and-analytics/shopify-reports/report-types/analytics-fields

2. Shopify Help Center, Session measurement update (how session changes can alter numbers): https://help.shopify.com/en/manual/reports-and-analytics/discrepancies/session-measurement-update

3. Google Analytics Help, Recommended events (ecommerce events view_item, add_to_cart, begin_checkout, purchase): https://support.google.com/analytics/answer/9268036

4. Google Analytics Help, Engagement rate and bounce rate (bounce rate is the opposite of engagement rate): https://support.google.com/analytics/answer/12195621

5. Baymard Institute, 50 Cart Abandonment Rate Statistics (70.22% across 50 studies, 42% just browsing): https://baymard.com/lists/cart-abandonment-rate