At a glance
The share of units sold that come back as returns, for the selected period and the prior one. Return rate is a margin killer and a product-quality early warning rolled into one: every return costs the referral-fee admin charge, the return shipping, often the unit itself if it cannot be resold, and it drags on account health. A return rate climbing past roughly 8% is the trigger to dig into reasons by ASIN before it eats the category.
| What it counts | Returned units ÷ units sold over the same window, expressed as a percentage. Read against the prior period so a spike stands out from the baseline. |
| Numerator | Units returned (customer returns) in the window. Both FBA returns (handled by Amazon) and FBM returns (handled by you) count, where data is available. |
| Denominator | Units sold (ordered units) in the window. Using units, not orders, keeps multi-unit orders weighted correctly. |
| What “good” looks like | Highly category-dependent. Apparel and shoes run structurally high (size and fit); consumables and many home goods run low. Read the trend against your own baseline, not a universal number. |
| Cost of a return | A refund usually returns most of the referral fee but keeps a refund administration fee, plus return shipping and the unit value if it is not resellable. Returns also feed Order Defect Rate and account health. |
| FBA vs FBM | FBA returns route through Amazon’s returns centres and may be graded as sellable or unsellable (the latter becomes stranded/unfulfillable). FBM returns come to you and you decide disposition. |
| Reasons live elsewhere | This card is the rate; the why is in Return Reason Clusters by ASIN. |
| Time window | 30D vsP (last 30D vs the prior 30D) |
| Alert trigger | >8%, driven by the return-rate sentiment key |
| Roles | owner, operations |
Calculation
Calculated automatically from your Amazon Seller Central data. See the At a glance summary above for what the metric tracks and the worked example below for a typical reading.Worked example
A home-and-kitchen FBA seller. Period: 02 Apr 26 to 01 May 26 (30D), compared against the prior 30D. Figures are illustrative.| ASIN group | Units sold | Units returned | Return rate |
|---|---|---|---|
| Cookware set (new listing) | 1,100 | 132 | 12.0% |
| Knife block | 1,800 | 90 | 5.0% |
| Storage containers | 2,400 | 96 | 4.0% |
| Utensil set | 900 | 36 | 4.0% |
| Return Rate (this card) | 6,200 | 354 | 5.7% |
- The portfolio rate hides a problem. At 5.7% the headline sits below the 8% alert, so Nerve Centre stays quiet at the portfolio level. But the cookware set alone is returning at 12%, well into problem territory. Always drill into per-ASIN reasons when the rate moves, not just the blended number.
- A new listing is the culprit. The cookware set is recently launched and returning twice the catalogue average. New-listing return spikes usually mean a listing-accuracy issue (photos, dimensions, what is in the box) or a genuine quality defect, both fixable, both urgent before reviews tank.
- The cost is bigger than the refund. Each of the 354 returns kept a refund admin fee, paid return shipping, and some units came back unsellable. The true margin hit is well above the refunded sale value. Pair with Net Revenue (after fees + refunds).
- Account health is downstream. A rising return rate, especially if returns are reason-coded as defective or not-as-described, feeds Order Defect Rate and ultimately Account Health Status. Catch it here before it becomes a health flag.
Sibling cards merchants should reference together
The rate tells you something is wrong; these tell you what and what it costs:| Card | Why pair it with Return Rate |
|---|---|
| Return Reason Clusters by ASIN | The diagnostic partner. The rate says “returns are up”; the clusters say “sizing on ASIN X, quality on ASIN Y”. |
| Order Defect Rate | Defect-coded returns feed ODR, the metric Amazon polices. A return-rate spike often precedes an ODR problem. |
| Star Rating Drift (top-50 revenue) | Returns and falling stars usually share a root cause. If both move on the same ASIN, the product or listing is the issue. |
| Net Revenue (after fees + refunds) | Quantifies the margin damage. Refunds plus return costs are subtracted here. |
| Negative Feedback (30d) | Returns and negative seller feedback often spike together when fulfilment or product quality slips. |
| Account Health Status | The end of the chain. Sustained high returns can pull account health down. |
Reconciling against Amazon Seller Central
Where to look in Seller Central: The closest Amazon-native views are:Reports → Fulfilment → Customer Returns (FBA returns, reason-coded) and the Return reports for the per-order return detail, plus Reports → Business Reports for units ordered, which is the denominator.For FBM, returns are managed under Orders → Manage Returns. Combining FBA and FBM returns against total units ordered gives the blended rate this card shows. Timing, settlement, and reporting-lag table:
| Topic | Detail |
|---|---|
| Return-window lag | A unit sold today can be returned weeks later. A 30D return-rate window counts returns that happen in the window against units sold in the window, so a late-arriving return is attributed to its return date, not the original sale date. This is why return rate lags a quality problem by the return window. |
| Reason coding lag | The customer-selected return reason is captured when the return is initiated; Amazon’s grading (sellable vs unsellable) can come days later. The rate is available before disposition is final. |
| FBA vs FBM timing | FBA returns appear quickly in Amazon’s returns report. FBM returns depend on you logging them, so FBM return data can lag if processing is manual. |
| Refund vs return | A refund without a physical return (a goodwill refund) is not the same as a return. The card focuses on returned units; refund-only events are a separate signal. |
| Reason | Direction | Why |
|---|---|---|
| Numerator vs denominator window | Either direction | If returns are counted by return date and sales by sale date, a 30D window can show a return against a unit sold just before the window opened. Amazon’s own reports may align dates differently. |
| FBM return completeness | Ours may be lower | FBM returns depend on consistent logging. If FBM returns are processed offline, they can be under-counted until reconciled. |
| Goodwill refunds | Ours may be lower | A refund with no physical return is not a return. If you count goodwill refunds as returns elsewhere, the rates differ. |
| Multi-unit orders | Ours uses units | The card uses units returned over units sold. An orders-based rate (returns ÷ orders) reads differently for multi-unit baskets. |
| Card | Expected relationship | What causes legitimate divergence |
|---|---|---|
amazon.order_defect_rate | Returns feed ODR when defect-coded. A rise in defect-reason returns precedes an ODR rise. | Not every return is a defect. A high return rate from sizing or buyer’s remorse may not move ODR at all. |
shopify return / refund cards | Channel-specific behaviour. The same product can return at very different rates on Amazon vs a DTC store because of audience and policy differences. | Amazon’s generous returns policy typically drives a higher return rate than a stricter DTC policy on the identical product. |