How duplicate customers distort your repeat purchase rate

Most repeat-customer numbers start with a count of customers. In Shopify, that count is a count of profiles, and a person with two profiles counts twice. This page shows the arithmetic, works through a small example, shows when duplicates push the rate the other way, and sets out how to measure the effect on your own store.

The formulas

All three use the same period, for example the last 12 months.

• Repeat purchase rate = customers with 2 or more orders ÷ customers with at least 1 order
• Orders per customer = orders ÷ customers with at least 1 order
• Revenue per customer (a simple, historical CLV) = revenue ÷ customers with at least 1 order

Orders and revenue don’t change when you count people instead of profiles. The number of customers does, and so does who counts as a repeat buyer. Shopify’s own reports define their customer metrics in their own way, so check the definition before comparing a report with your own sums.

Where to find it in Shopify

Shopify’s customer reports use their own names. Its Returning customers report “displays data about all your customers whose order history includes two or more orders”, and its customer cohort report lists “customer retention rate” as a metric without defining the formula (Customers reports). The page gives no “repeat purchase rate” formula, so write down which formula you use before you compare a report with your own sums.

A worked example

When counting people lowers the rate

Counting people doesn’t always raise the rate. It falls when every profile of a split person is already a repeat buyer. Here is a second made-up example, separate from the first one.

Joining two repeat profiles removes one repeat customer and one customer, so the rate falls whenever it was below 100%. Orders per customer still rise in this example, from 8 ÷ 6 = 1.33 to 8 ÷ 5 = 1.6.

Wrong matches skew it too. Group two different people who live together and you invent a repeat buyer. That is why the steps above keep profiles that share only an address apart until someone has checked them, and why the result is a range.

What MergeGuard’s demo store shows

MergeGuard’s demo store (Northsail Supply) shows the same effect on lists that depend on order history:

• One customer: Amelia Hughes: 3 profiles with 4, 3 and 2 orders → one customer with 9 orders Demo store (Northsail Supply)
• Win-back, for customers who haven’t ordered in a while: 905 customers look like they stopped ordering → 842 really did; 63 still buy on another profile Demo store (Northsail Supply)
• VIPs: 18 VIPs on single profiles → 27 VIPs as people Demo store (Northsail Supply)

The win-back line is the costly one. Those 63 customers still buy, under a newer profile, yet counted by profile they look lapsed, as if they had stopped ordering, and would be offered a discount to come back.

Reading your result

A large gap between the two views tells you two things. First, your reports undercount loyalty, so decisions based on them, like how much to spend on winning new customers, rest on a low number. Second, your customer lists are wrong in the same direction: win-back lists include active buyers, and VIP lists miss real VIPs.

A small gap in the repeat rate does not show that duplicates are few. If both profiles of a split person already repeat, the rate stays put while their orders and spend sit on two profiles. Count the surplus profiles too (profiles minus people), and read orders per customer and revenue per customer beside the repeat rate. Cleanup is housekeeping only when the surplus is small.

Either way, merging is a separate decision. You can count people for reporting without merging anyone. Merging in Shopify is permanent, and if any profile in a group is subscribed to marketing, the merged customer stays subscribed. Read the guest checkout guide before you merge.