Revenue per customer: formula, Shopify reports and how to raise it
Revenue per customer is revenue divided by unique buyers. The formula, where to find each number in Shopify, why it drops, and why tests use RPV instead.

Will O'Brien
The revenue per customer formula
Revenue per customer = revenue ÷ unique customers who bought
Say you run a bamboo sock and underwear brand out of Brisbane. Most people come in through a three-pack, and the good ones come back when the elastic goes. Here's last quarter (1 July to 30 September):
- Net sales: $96,000.
- Unique customers who ordered: 1,600.
- Revenue per customer: $96,000 ÷ 1,600 = $60.
The same quarter also gives you the two numbers underneath it:
- Orders: 2,000.
- Average order value: $96,000 ÷ 2,000 = $48.
- Orders per customer: 2,000 ÷ 1,600 = 1.25.
- Check: $48 × 1.25 = $60.
That last step is the useful part. Revenue per customer always splits into order size and order count, so when it moves you can see which one moved it.
The denominator is people, not orders. A customer who ordered three times in the quarter counts once. That's the difference from average order value, which divides by orders.
Where to find each number in Shopify
Shopify doesn't show revenue per customer as one figure, so you pull two numbers and divide.
Revenue. In Analytics > Reports, the Sales over time report gives you net sales and orders for any date range. Shopify defines net sales as gross sales minus discounts minus sales reversals (returns, cancellations and edits). Total sales adds tax, shipping, duties and fees on top. Either works, as long as you use the same one every time. We use net sales, because tax and shipping aren't money you get to keep.
Customers. Filter the Reports list by the Customers category and open New vs returning customers. It counts first-time and returning customers for each time unit you choose. Set Group by to the same span as your window (a quarter, in the example above) and use the customer count for the period. If you group by month instead, someone who ordered in July and again in September shows up in two months.
Shopify calls a customer "returning" if their order history already holds at least one order, at any point. So a buyer from two years ago who orders this quarter is returning, even if they're the only order you've had from them in ages.
For the long view, open Customer cohort analysis and set the metric to Amount spent per customer. Each row is a group of customers who first bought in the same month, and the columns show what they've spent per head since. That's revenue per customer tracked over time for each starting month. If you have 24 months of data, Shopify can also project future spend per customer for each cohort.
Why revenue per customer drops when nothing is wrong
Two things move this number that have nothing to do with how good your store is.
The first is the window. The same sock brand's figure for the full year was $372,000 from 4,650 customers, or $80 per customer. That's higher than any single quarter because a year gives people time to come back. A monthly number will be lower again. Compare a month with the same month last year, never with a quarter.
The second is who bought. In October, a creator posted a "socks that don't fall down in your gym shoes" video and it took off. Here's what happened:
- September: $40,600 from 700 customers, so $58 per customer.
- October regulars: 600 customers spent $36,000, so $60 each.
- October new buyers from the video: 900 customers spent $34,500, about $38 each. Nearly all bought one three-pack to try.
- October overall: $70,500 from 1,500 customers, so $47 per customer.
Revenue per customer fell 19%. Revenue rose by $29,900 and the regulars spent more than in September. The drop came from 900 people placing a first, cheap trial order, which is exactly what you'd hope a viral video does.
So before you react to a lower number, split it by new and returning customers. If the returning figure held, you're fine. You've just got more first orders in the mix, and the question becomes how many of those 900 come back.
Returns pull it down too. Sales reversals land on the date the return is processed, so a wave of wrong-size returns can dent next month's number. For apparel, a clear size chart on the product page is one of the cheaper fixes for that.
How it fits with AOV, revenue per visitor and lifetime value
These four get mixed up because they share a numerator. The difference is what you divide by.
- Average order value: revenue ÷ orders. How big each basket is.
- Revenue per customer: revenue ÷ unique customers. How much each buyer spent in the window.
- Revenue per visitor: revenue ÷ visitors. What each person who landed on the store was worth, buyer or not.
- Lifetime value: revenue per customer with the window stretched over the whole time someone stays a customer. Shopify's cohort report gets you closest to it.
Revenue per customer tells you about the people who already bought. It says nothing about the people who left without buying. That's why it's a good number for retention and acquisition decisions, like how much you can pay to win a customer, and a poor one for judging changes to the store.
Why A/B tests shouldn't be judged on revenue per customer
In an A/B test, visitors are split at random, so both versions get the same kind of traffic. Customers aren't split at random. Who becomes a customer is the outcome of the test. A variant that scares off small buyers shrinks its own denominator and looks better per customer, even if it makes less money.
Say the sock brand tests a free pair of no-show socks with any order over $60, aimed at pushing people from one three-pack to two. Each version gets 20,000 visitors:
- Control: 500 customers spent $30,000. Revenue per customer is $60, and revenue per visitor is $30,000 ÷ 20,000 = $1.50.
- Variant: 450 customers spent $29,250. Revenue per customer is $65, and revenue per visitor is $29,250 ÷ 20,000 = $1.4625.
- Revenue per customer change: (65 - 60) ÷ 60 = +8.3%.
- Revenue per visitor change: (1.4625 - 1.50) ÷ 1.50 = -2.5%.
The free gift got buyers to spend more, and it lost 50 buyers who only wanted one pack. Judged per customer, the variant wins. Judged per visitor, the store took $750 less from the same traffic. Revenue per visitor is the number that matches your bank account, because its denominator is fixed by the random split.
There's a second problem. Most of what lifts revenue per customer over time is repeat orders, and those arrive weeks or months after a two-week test ends. So run the test properly on revenue per visitor: one change at a time, a sample size set before launch, at least two full weeks, a winner called at 95% significance, and no peeking before the sample is in. If the variant loses, take it off the site and log what you learned.
If your testing setup writes the variant onto each order as a tag or note attribute, you can add a slower second check. Pull the customers from each version 90 days later and compare their revenue per customer. That tells you whether the winner also brought back buyers, or just bigger first baskets.
You can run all of this yourself. Tools like Intelligems and Shoplift handle the split and the maths, and the research, design, build and reading of results stay with you. An agency on a monthly retainer takes those hours off your plate, and you pay the retainer whether the tests win or not.
Two ways to raise it
Since revenue per customer is order value times orders per customer, every idea you have pulls one of those two levers. It helps to know which one you're pulling.
Bigger orders. Bundles, a free shipping threshold, a "complete the set" add-on in the cart. These show up inside a normal test window, so you can test them on revenue per visitor and know within a few weeks. There's a longer walkthrough in how to increase AOV.
More orders from the same people. This is where the sock brand's real upside is. Open the cohort report, find the month most customers place their second order, and time a reminder just before it. Subscriptions, refill reminders and post-purchase emails all live here. They take longer to read, so judge the on-site parts (a subscribe option on the product page, say) on revenue per visitor, and watch the cohort report for the repeat effect.
Neither lever is better. A store selling something people rebuy, like socks or coffee, usually has more room in order count. A store selling a once-every-five-years mattress has to make the first order count.
Your next step
Pull last quarter's net sales, orders and unique customers, and write down three numbers: revenue per customer, AOV and orders per customer. Then do the same for new and returning customers separately. Whichever of those has moved least since last year is where to look first.
That's what we do: our team researches your store, designs tests for order size and repeat buying, and judges each one on revenue per visitor from real Shopify orders. AI speeds up the build, and you approve every test in Slack. We only charge when you make money. If you're curious what your store's customers could be worth, book a call.