The true cost of returns: one return erases several orders
A return costs more than lost revenue: two-way shipping, packing, ad spend and sometimes the goods. Formulas, worked numbers and tracking by channel.
By The DataxLand team· · 5 min read
When an order is returned, most shop owners think: "I just lost that order." In reality, a return usually takes more than its own profit: the outbound shipping the shop already paid, the return shipping, packing labour, the ad money spent to win the order, and, if the item cannot be resold, the cost of the goods too. This post puts a price on that with a simple formula, shows why the same product can be profitable in one channel and loss-making in another purely because of the return rate, and how to track it so you see it early.
The post uses two kinds of examples: (1) the DataxLand sample data (a simulated scented-candle shop, three channels, 1 Aug – 29 Oct 2026) and (2) illustrative examples with stated assumptions. None of the numbers belong to a real customer.
Returns, cancellations and failed deliveries: three different things
Before measuring, separate three states that many reports lump together:
| State | Happens when | Main cost | Counted as a valid order? |
|---|---|---|---|
| Cancelled before shipping | Buyer or shop cancels before handing over to the carrier | Almost none, apart from handling | No |
| Failed delivery | The buyer refuses or misses the parcel and it comes back | Two-way shipping, packing | Usually not |
| Return / refund | The buyer received the item, then asks to return it | Refund, return shipping, item may be damaged | Yes (a successful order that was then returned) |
In DataxLand, the cancellation rate (cancel_rate) is calculated over all orders, while the return rate (return_rate) is
calculated over valid orders. Each marketplace's raw statuses (every marketplace names them differently) are mapped to one standard
set of statuses; Si proposes the mapping and you approve it. If an unknown status appears that has not been mapped yet, the data
checks flag it instead of quietly miscounting (see Verified numbers).
The cost of one returned order
Illustrative example, assumptions: selling price 300,000 ₫, cost of goods 150,000 ₫, marketplace and transaction fees 18% (54,000 ₫) when the order succeeds, the shop pays 15,000 ₫ outbound shipping, 5,000 ₫ packing and 25,000 ₫ return shipping. We assume the marketplace does not charge fees on the refunded amount (some fee types are calculated on the amount after refunds — check with your marketplace).
| Successful order | Returned, resellable | Returned, damaged | |
|---|---|---|---|
| Net revenue | 300,000 ₫ | 0 ₫ | 0 ₫ |
| Cost of goods | −150,000 ₫ | 0 ₫ (back in stock) | −150,000 ₫ |
| Marketplace and transaction fees | −54,000 ₫ | 0 ₫ | 0 ₫ |
| Outbound shipping, packing | −20,000 ₫ | −20,000 ₫ | −20,000 ₫ |
| Return shipping | — | −25,000 ₫ | −25,000 ₫ |
| Contribution margin | +76,000 ₫ | −45,000 ₫ | −195,000 ₫ |
Another way to read the table:
- One resellable return wipes out about 0.6 profitable orders.
- One return that cannot be resold wipes out about 2.6 profitable orders.
That is before ads. If each order costs 40,000 ₫ in advertising, that money is spent whether or not the order is returned.
The formula: profit per shipped order
Let r be the return rate, L the contribution margin of a successful order and H the cost of a returned order:
Expected contribution margin per shipped order = (1 − r) × L − r × H
Break-even return rate = L ÷ (L + H)
True ad cost per kept order = Ad cost per order ÷ (1 − r)Applied to the example above with goods that cannot be resold (H = 195,000 ₫):
| Return rate | Expected margin per shipped order | Versus a 3% return rate |
|---|---|---|
| 3% | 67,870 ₫ | — |
| 8% | 54,320 ₫ | −20% |
| 12% | 43,480 ₫ | −36% |
| 20% | 21,800 ₫ | −68% |
| 28% | 120 ₫ | break-even |
Going from 3% to 12% — the gap between two channels in the sample data below — cuts the margin per shipped order by more than a third, before ads.
Sample data: one channel returns four times as much
Over three months of sample data:
| Channel | Return rate | Contribution margin / net revenue | Profit after ads |
|---|---|---|---|
| Shopee | 2.69% | 32.2% | 89,927,600 ₫ |
| Website | 3.21% | 41.8% | 13,175,700 ₫ |
| TikTok Shop | 12.41% | 18.5% | −13,802,100 ₫ |
TikTok Shop's return rate is more than 4 times Shopee's. TikTok Shop's contribution margin is lower for several reasons (fees, channel pricing, product mix), but returns are a big part: refunds come straight off net revenue, return shipping goes into costs, and the sample data assumes returned goods are not restocked, so the cost of goods is not recovered. The result: the channel ranked second by revenue is the only one losing money after ads.
Across the whole shop, return fees (return shipping) total 3,119,000 ₫ — small on their own. The expensive part is elsewhere: the refunds, the unrecoverable cost of goods, and the ads already spent on those orders. So do not judge returns by the "return fee" line on your settlement statement.
What return rate is too high?
There is no reliable universal threshold across industries; fashion and cosmetics usually see more returns than household goods, and impulse buys from livestreams usually see more returns than deliberate searches. Instead of comparing with an "industry average", compare three things of your own:
- Against the product's break-even return rate (formula above). This is the absolute ceiling.
- Across channels selling the same product. Big differences usually come from how it is sold (livestreams, deep vouchers), not from the product.
- Over time. A sudden jump in returns after a new batch of stock, a new carrier or a new campaign is a signal to act now.
Reducing returns: a checklist by cause
- Listing does not match reality: real photos, dimensions, scent (for candles and essential oils), materials. "Not as described" is usually the shop's fault and the most expensive reason.
- Damaged in transit: repack fragile products; track return rate by carrier.
- Impulse buying: in livestreams, repeat the key product information before closing; avoid very deep vouchers just to pull orders.
- Failed deliveries: confirm high-value cash-on-delivery orders, especially from new accounts.
- By SKU: filter SKUs with twice the shop's average return rate; usually a few SKUs cause most returns.
- Stop advertising SKUs whose return rate exceeds the break-even rate until the cause is fixed.
Tracking returns without merging files
Measuring this correctly means joining orders (statuses), settlements (refunds, return fees) and cost of goods by date — from every channel. DataxLand does that join and shows return rate, return fees and profit after fees by channel and product on the retail operations dashboard and the finance dashboard. Set an alert, for example "weekly return rate above 8%", to get a message on Telegram or Slack as soon as it crosses the threshold (see Reports and alerts).
Read next: How to calculate true profit on Shopee, TikTok Shop and Lazada · What ROAS is profitable? · Automated reports on Telegram