A refund is not the opposite of a sale.
A sale debits Cash (or AR) and credits Sales, Sales Tax Payable, and reduces Inventory while debiting COGS. A refund reverses all four — but most cheap POS systems only reverse the cash and revenue side, not the inventory and COGS side. That gap is where your books drift from reality.
A correct refund posts: DR Sales Returns (contra revenue), DR GST Payable, CR Cash, then DR Inventory, CR COGS at the original cost basis. Five lines, automatic. If your POS does fewer than that, your COGS is overstated and your inventory is understated by the cost of every refund.
Original cost vs replacement cost.
When a returned item goes back into inventory, what cost basis does it carry? If you bought it for A$50 last month and replacement cost is now A$55, returning at A$55 inflates inventory and understates COGS. Always return at original cost basis.
This is where weighted-average cost (WAC) helps you. The original cost is the WAC at the moment of the original sale, which the POS already recorded on the original transaction. The refund just looks up that historical WAC and posts the inventory-back at that value. No spreadsheet math.
Refund with vs without receipt.
A refund with the original receipt is straightforward — POS finds the transaction, reverses it. A refund without the receipt is harder because the original cost basis is unknown. Two options: refund at the lowest sale price of that SKU in the last 90 days, or require a manager override with a fixed cost.
Many chain retailers refuse no-receipt refunds and offer store credit only. That side-steps the cost basis problem (store credit is a liability, not a refund) and limits abuse. Set the policy, train the floor, and post the policy at the counter.
- With receipt: POS-driven refund at original price and cost basis
- Without receipt: store credit at last-sold price, manager approval
- Damaged item: write to Damaged Goods, refund at original price
- Outside warranty window: store credit only, manager approval
Exchanges are two transactions.
A customer brings back a size-medium and wants a size-large. Most cashiers think of this as one transaction — the swap. Accountingly it is two: a refund on the medium and a sale of the large. Treat it that way and the books are clean. Treat it as a single swap and your inventory math gets weird.
Even-exchange (same price) is the easy case. Uneven exchange (returned item A$40, new item A$55) is the common case where most POS systems fumble. The customer pays the difference of A$15 in cash, but the POS has to record both halves: an A$40 refund and an A$55 sale.
Tax treatment on refunds.
Sales tax follows the refund. If the original sale charged 10% GST on A$85 (A$8.50 tax), the refund reverses the tax — DR GST Payable A$8.50. This affects your BAS reporting period because your tax liability decreased by A$8.50.
Under ATO rules (and most VAT/GST regimes globally), the refund must reference the original invoice and be transmitted as a credit note. If the original sale was in January and the refund is in March, the January period is closed and the refund affects your March return. Document carefully so the variance is explainable.
The refund abuse pattern.
A cashier issues refunds to a personal card or a friend's card after closing time. The transaction looks legitimate in the daily report. The next day no customer disputes it because no real customer was involved. Over six months this can add up to A$10,000-20,000.
Defense: every refund over a small threshold requires a manager PIN, refunds to card require the original card to be present and the last 4 digits to match the original sale, refunds after closing time are blocked, and a refund report goes to the owner weekly. Each layer reduces the attack surface.
The store credit option.
Store credit is a liability on your balance sheet (CR Customer Credit Liability) instead of a cash outflow. Customers who want flexibility get a credit voucher with an expiry, you keep the cash, and the liability is written off after the expiry if unused.
Many retailers under-use store credit because it feels like friction with the customer. The data says otherwise: 30-40% of issued store credit goes unused at retailers with 6-month expiry. That is essentially free working capital. Offer it as the first option, refund-to-original-payment as the second.
Reporting and trend monitoring.
Pull a refund report monthly: total refunds as percentage of sales, refunds per cashier, refunds without receipt, refunds after hours, refunds over A$50. Each number tells you something about your floor and your customer mix.
A healthy retailer runs 1-3% refunds-to-sales. Above 5% is a customer-satisfaction or product-quality problem. In Nonari the refund dashboard surfaces these trends per branch with month-over-month comparisons. The story is in the trend, not the absolute number.