Delivery and Return Rates: What a Failed Order Really Costs
Most stores watch their delivery rate as a percentage and stop there. "80% delivered" sounds fine. The question that changes decisions is different: what does each of the other 20% cost you, and how many good orders does it take to cover one?
Why the delivery rate alone is not enough
A rate treats every failed order as one unit of failure. In money they are not equal. A parcel refused at the door after outbound shipping costs more than an order cancelled before dispatch. A returned item that cannot be resold costs more than one that goes back on the shelf. And whichever it is, the advertising money spent to win that order is gone.
So two stores with the same 80% delivery rate can have very different economics. Percentages describe frequency. Only a cost per failed order describes damage.
The cost stack of one failed order
Build it line by line from your own invoices and contracts. Do not use anyone else's numbers.
Outbound shipping
Paid even if the parcel is refused
25Return shipping
If your courier charges for the way back
20Packaging
Usually not reusable
5Handling
Picking, packing, restocking time
5Advertising
Spent to win an order that did not complete
20
Total 75. Check each line against your courier terms; some couriers price returns differently.
Add it up: 25 + 20 + 5 + 5 + 20 = 75. Two lines are easy to miss. The first is damaged or unsellable stock, which depends on your product and is a separate loss on top of the 75. The second is the cost of the money and time tied up while the parcel travels both ways.
Shares are rounded. Your own mix will differ by courier and product.
How many good orders pay for one failure
Suppose a delivered order leaves a contribution of 90 before advertising, as in the method in Contribution Margin per Order. This is an illustrative example, not real results.
- A failed order loses 75 in the stack above, including its ad cost.
- 75 ÷ 90 ≈ 0.83, so about five good orders in six are needed just to cancel one failure.
Now see the effect at scale. Ship 100 orders and compare two outcomes, counting only costs before advertising (so a failure costs 55, the stack without the ad line):
- 80 delivered, 20 failed: 80 × 90 = 7,200 minus 20 × 55 = 1,100 gives 6,100, or 61 per shipped order.
- 90 delivered, 10 failed: 90 × 90 = 8,100 minus 10 × 55 = 550 gives 7,550, or 75.5 per shipped order.
Ten points of delivery rate moved the result by 1,450 on 100 orders, about 24% more, and no ad was changed. This is why failed orders deserve the same attention as ad cost.
Trade-offs and when the recommendation changes
- Cheaper shipping can raise failures. A courier that costs less but delivers worse can cost more overall. Compare couriers on cost per delivered order, not cost per parcel.
- Stricter confirmation can reduce volume. Screening out risky orders lowers failures but also lowers orders. Test it and compare total contribution, not the rate.
- High-ticket or fragile items shift the stack toward product loss; low-ticket items toward shipping.
- If your return policy is generous by design, failures belong in your pricing, not in a one-off fix.
One more habit helps: review failed orders by cause every month. If most failures are incomplete addresses, the fix sits in checkout and confirmation. If most are refusals at the door, look at expectations set by the ad and the page, and at the time between ordering and arrival. The same 20 failures can have very different remedies, and only the cause tells you which one is worth paying for.
Implementation steps
- Separate failed orders into types: cancelled before dispatch, refused at the door, undeliverable address, returned after delivery.
- Price each type from your courier terms and your own handling costs.
- Record the resale status of returned stock.
- Compute cost per failed order and failures per 100 shipped, by courier, city and product.
- Put the figure beside your break-even limit so the ad budget reflects it.
What this cannot tell you
Reason codes from couriers and customers are often vague, so causes remain partly guesses. Failures also lag orders: this week's returns belong to last week's ads, so reading rates on a short window distorts the picture. And the platform never sees failures unless you send that outcome back to it.
Next step
If you do not know your cost per failed order by courier and source, a Full Ads & Tracking Audit can start from your order data and show where the loss sits. First, make sure confirmations are measured properly in Cash-on-Delivery Confirmation Rate, then compare channels in COD vs Prepaid.