Product Return & Kept Rate

Product Return & Kept Rate shows which products come back and which stay sold. It gives the share of units customers returned, and the share they kept.

The Product Return & Kept Rate report, with summary cards, a bar chart of the items with the most returns and a table of units sold, units returned, return rate and kept rate

How It’s Calculated

The report counts units, not money. The cards at the top show Units Sold (matured), Units Returned and Return Rate for the date range.

The chart shows the return rate of the 10 items with the most units returned. Items with too few units for a rate are left off it.

A row of tabs above the table switches what each row is: Product, Variant, SKU, Vendor, Collection or Tag. It starts on Product. Rows are listed with the most units returned first, up to 500.

The columns are:

  • Item: the product, variant, SKU, vendor, collection or tag. A variant shows as the product title and the variant title, such as Tee — Blue.
  • Units Sold (matured): the units sold on orders old enough for their returns to have settled.
  • Units Returned: the units from those orders that were refunded.
  • Return Rate: Units Returned divided by Units Sold (matured). A dash means fewer than 10 units, too few for a rate. It doesn’t mean no returns.
  • Kept Rate: 1 minus the Return Rate. It’s also a dash below 10 units.

A totals row under the table adds up the rows. Show totals row in the Report settings menu turns it on or off for every report. The cards use the same totals.

Which Orders Count

Returns take about 90 days to settle, so the report counts only orders placed more than 90 days ago. Orders from the last 90 days are left out of both sides of the rate. For a date range that ends inside the last 90 days, only its older part counts, and a range entirely inside it shows no rows.

A return is counted against the order it came from, by the order date, not the day of the refund. The data doesn’t record refund dates for individual units.

  • Canceled orders, and orders whose payment was voided or expired, are left out.
  • Units canceled from an order that wasn’t fully canceled count as returned, because the data can’t tell the two apart.
  • Order lines that don’t record whether they were returned are left out of both numbers.
  • Orders count on the day they were processed, in your store’s time zone.

What Each Row Groups

  • Product, Variant: items with no product or variant aren’t listed.
  • SKU, Vendor: lines with no SKU or no vendor aren’t listed.
  • Collection, Tag: products that have none are grouped as Unassigned. The groups are the ones your products are in now, not when the order was placed. A product in several collections or tags counts in each, so the totals row and the cards add its units more than once.

Segment, Sales Channel and Filters narrow the report. There’s no comparison period.

What Good Looks Like

Return rates differ a lot between kinds of product. Clothing and shoes come back more than a refill or a candle. Compare an item with your other items of its kind and with its own past.

  • Return Rate stays low on your best sellers. A high rate on an item that sells well costs you the most.
  • A rate that’s the same across variants. One variant far above the others, such as a size, points to a fit or quality problem with that variant.

What to Do About It

  • Start with the top of the list. It’s sorted by units returned, so the rows that cost you the most are first.
  • Switch to Variant. If one size or color is returned far more, fix its listing, its size guide or its supplier.
  • Check the cost. Refund Profit Impact shows what refunds cost your profit.
  • Watch the trend. Return Rate Trend shows whether your overall return rate is rising or falling.

Get AI Insight at the top of the report asks Copilot for a short read of it.

Where to Find It

Still need help? Contact Us Contact Us