Product Cohort Retention

Product Cohort Retention shows how many of the customers who first bought a product came back to order again, month by month. Each row is a product. Its customers are the ones who had that product in their first order.

How It’s Calculated

A customer’s first order is their first order with you, after any Sales Channel, Segment and Filters you set. The customer is in the report only if that order was in your date range. Orders with no customer don’t count.

Each product on that first order gets the customer as part of its cohort. A customer whose first order held 3 products is in 3 rows, so the rows’ sizes add up to more than your number of new customers.

The columns count months since each customer’s first order: MO 0 is the month of the first order, MO 1 the month after, and so on. Months follow your store’s time zone and calendar months, not 30-day blocks.

With Retained customers (%) selected, each cell is the share of the row’s customers who placed any order in that month. The order can be for any product. A customer who orders in months 1 and 3 counts in both columns, and not in month 2.

MO 0 is always 100%, because the first order is in it.

Use the dropdown in the filter bar to show a different figure in each cell:

• Retained customers (%): the share of the cohort who ordered that month.
• Retained customers (#): the number of those customers.
• Orders: the number of orders they placed that month.

The Cohort size column is the number of customers in the row. Rows run from the largest cohort to the smallest, and the report shows at most 100 products.

What Counts

• Orders that were canceled, or whose payment was voided or expired, don’t count.
• Refunds aren’t taken off. An order that was refunded still counts as an order.
• A product is counted by its ID, and the row shows its title.
• The date range picks which customers are in the report by the date of their first order. Orders after the range ends aren’t counted, so the grid stops at the end of the range.
• The range is in whole months. Trailing 1 year is the default.

A blank cell means no one in the row ordered that month. That includes a month that hasn’t started yet. The report doesn’t show 0%.

A row pools customers whose first orders are anywhere in the date range, and every one of them counts in Cohort size. A customer who first ordered late in the range has had fewer months to come back. The later columns therefore read low, and the longer the range, the more they do.

Summary Cards and Averages

• First-Time Customers: the sum of every row’s Cohort size. A customer who is in several rows is counted in each.
• Month 1 Retention, Month 3 Retention and Month 6 Retention: with Retained customers (%) selected, the average of that column across the rows, weighted by Cohort size. With Retained customers (#) or Orders selected, the total of that column.

The Weighted Avg row under the grid gives an average for every column, weighted by Cohort size, in the figure you’re showing. It uses only the rows that have a value in that column.

A row left blank because no one came back is left out of the averages and of the cards. In the % view they can read higher than the true rate.

The cards show a dash when the range is too short to reach that month.

What Good Looks Like

Compare products with each other, on the same date range. A product whose row stays higher than the others from MO 1 onward brings customers who keep buying. A product whose row drops to a blank after MO 0 is bought once.

How high the rows sit depends on how often your products need replacing, so don’t compare your store with another store.

Check Cohort size first. A row with 10 customers moves 10 percentage points for every customer, so it says little. The more customers in a row, the more you can read into it.

Retention isn’t only the product’s doing. The customers in a row also bought everything else in their first order, so a product that’s always bought with a best seller looks like that best seller.

What to Do About It

• A product with high retention: use it as the first product in ads and offers aimed at new customers.
• A product with low retention: add a follow-up. Send a reason to come back, such as a related product or a reorder reminder.
• A big gap between 2 similar products: compare their first-order customers. Set Segment or Filters to see whether one product brings customers through a discount code or a different channel.
• Customers who buy the same product again: see Product Repurchase Rate.

Get AI Insight at the top of the report asks Copilot for a short read of it. Export CSV downloads the grid for the figure you’re showing.

Where to Find It

• Product Cohort Retention: open Reports and look under Products.
• Cohort Retention: the same view, with cohorts grouped by the month of the first order.
• Product Repurchase Rate: the share of each product’s customers who bought that product again.