If you run TikTok Shop affiliate campaigns, you send creators free product to review. In your settlement export, those samples frequently appear as orders with a 100% discount — real order rows, zero revenue. Left unhandled, they quietly corrupt three numbers you rely on. Here's what breaks and how to account for it properly.
What breaks
- Order count — every sample inflates your total orders, so your conversion and volume metrics look better than reality.
- Average order value — a pile of $0 "orders" drags AOV down, making pricing and cohort analysis wrong.
- Profit — the sample cost you real COGS and shipping, but it shows zero revenue. If you don't record the cost, profit is overstated; if you count the row as a normal order, revenue is understated. Either way the books are wrong.
The right way to account for a sample
A creator sample is not a sale — it's a marketing expense. The clean treatment:
- Exclude the row from revenue. It isn't income; don't let it sit in your sales figure at $0 dragging AOV.
- Record the product's cost as marketing spend. The COGS of the unit plus any shipping you paid is a real cost of the campaign — book it as marketing, not as cost of goods sold against a sale that never happened.
- Attribute it to the campaign if you can, so you can later judge whether that creator's samples produced enough affiliate sales to justify the spend.
How to spot them in the export
Samples show the tell-tale pattern of an order line whose discount equals the item price — net revenue zero. The trap is that a naive spreadsheet SUM of "orders" or "revenue" treats them as ordinary zero-value sales and silently distorts every derived metric. You have to identify and separate them before any of your numbers mean anything.
This is one of the specific things a purpose-built analyzer earns its keep on. When Sumhound reads a TikTok Shop settlement, 100%-discount sample rows are recognized and kept out of revenue, so your reconciled numbers reflect actual sales — not sales plus giveaways.
Why it matters beyond bookkeeping
Get this right and you can finally answer the question that justifies the whole affiliate motion: did the samples pay for themselves? Samples are a cost; affiliate-driven sales are the return. You can only compute that ROI once the samples are out of your revenue and sitting in marketing where they belong.