At a glance
The total dollar amount refunded to customers in the period. Absolute view (not ratio). The number to subtract from gross volume for a net-of-refunds figure.
| What it counts | SUM(refund.settleAmount) where transactionType = refundTransaction and transactionStatus = settledSuccessfully. |
| API endpoint | getTransactionListRequest. |
| Currency | USD-dominant; multi-currency stacked. |
| Refunds counted | Full and partial. |
| Refund timing | Issue date (when the refund was processed). |
| Disputes / chargebacks | NOT counted (separate metric). |
| eCheck.Net (ACH) refunds | Counted; ACH refunds settle 3 to 5 business days like the original. |
| Voided pre-settlement | Excluded (no money moved). |
| Time window | 30D vsP. |
| Alert trigger | +25% vsP relative spike. |
| Roles | owner, finance, operations |
Calculation
Calculated automatically from your Authorize.net data. See the At a glance summary above for what the metric tracks and the worked example below for a typical reading.Worked example
“Heartland Hardware Co.”, 30 days ending 02 May 26.| Channel | Refund volume | Count | Avg refund |
|---|---|---|---|
| Web checkout | USD 12,400 | 38 | USD 326 |
| B2B portal | USD 4,800 | 3 | USD 1,600 |
| MOTO phone | USD 1,700 | 1 | USD 1,700 |
| ARB recurring | USD 0 | 0 | - |
| Total | USD 18,900 | 42 | USD 450 |
- Total USD 18,900 against USD 1.87M revenue is 1.0%. See
aut_refund_ratefor the rate view. - B2B average refund USD 1,600 is 5x web average. Reflects B2B AOV; a single B2B return is materially worse for revenue than a web return. Investigate the 3 B2B returns individually; each is a strategic loss.
- Web channel 38 refunds is volume-noisy. Top 1 to 3 SKUs typically drive 50 to 70% of count. Drilldown by SKU.
- MOTO single-event USD 1,700 is suspicious. A high-value phone-keyed-in MOTO refund warrants review for fraud (a sales-rep collusion pattern), legitimate large customer return, or refund-keyed-in error.
- ARB at zero is correct. Subscription cancellations stop future rebills; refunding past rebills is rare and typically only for billing errors.
Sibling cards merchants should reference together
| Card | Why pair it with Refund Volume |
|---|---|
aut_refund_rate | The rate; volume divided by revenue. |
aut_total_volume | Gross volume; subtract this card for net. |
aut_chargeback_rate | Buyer-initiated cousin. |
Stripe stripe_refund_value / PayPal pp_refund_volume | Cross-PSP comparison. |
Reconciling against the vendor’s own dashboard
Where to look in the Authorize.Net Dashboard: account.authorize.net → Reports → Transaction Statistics withtransactionType = refundTransaction filter. Sum the amount column.
Why our number may differ:
| Reason | Direction | Why |
|---|---|---|
| Time zone | Boundary days off | Pacific batch cut-off vs UTC. |
| Refund settlement timing | Theirs slightly later | A refund issued today settles overnight; we count by issue date. |
| Comparison | Expected | Why |
|---|---|---|
aut_refund_volume ↔ commerce-platform refund total | Approximately equal for Authorize.Net-paid orders | Other gateways’ refunds excluded. |
Known limitations / merchant FAQs
A single large refund event spiked my volume, normal? Possible causes: (1) batch refund of a defective-product recall, (2) accidental double-charge corrected, (3) wholesale-customer return, (4) sales-rep error on a B2B order. Drill down byreferenceTransactionId to identify the original sale.
Refund volume is higher than gross sale, possible?
In a single day yes (refunds of older sales exceed today’s new sales). Over a 30-day window it shouldn’t happen and indicates a data sync issue or aggressive product-recall cycle.
Are unlinked refunds (credits) counted?
Yes. Authorize.Net allows “unlinked refunds” via transactionType = refundTransaction without a refTransId; these go through if the acquirer permits. Counted here. Most acquirers restrict unlinked refunds to chargeback-prevention scenarios.
Refund window expiration?
Authorize.Net allows 120 days from the original transaction. Beyond that, only unlinked credits work.
Refund vs void, financial difference?
A void cancels a captured-but-not-yet-batched transaction; the customer was never billed. A refund returns funds to a customer who was already billed (post-batch). This card counts only refunds.
ACH refunds, special handling?
Same refundTransaction type. Settlement takes 3 to 5 business days, same as the original.
Visa ownership of Authorize.Net, refund processing change?
No. Refund mechanics unchanged.