A discount for paying before the parcel arrives is the most popular lever in cash on delivery, and the most casually set. Sellers pick 5% because it sounds right, switch it on, watch prepaid orders rise, and call it a success.
That is the wrong test. Prepaid orders rising is guaranteed, because you are paying people to prepay. The question is whether the money you gave away is less than the money you saved, and the discount is paid to everyone who prepays, including the ones who would have prepaid at no discount at all.
This article works out the ceiling, the break-even switch rate, and how to choose a number from your own figures.
What a prepaid order is actually worth to you
Start with one order at 45 in local currency. Product cost 14, shipping 3.20, return cost 2.00, ads 11 per order placed.
| Cash on delivery | Prepaid, no discount | |
|---|---|---|
| Delivery rate | 72% | 97% |
| Payment processing | none | 2.5% |
| Money held by courier | one remittance cycle | none |
| Expected value per order placed | 7.56 | 14.72 |
The gap is 7.16 per order, and almost all of it is delivery rate. A prepaid order rarely refuses the parcel, because refusing costs the customer money. Nobody refuses something they have already paid for on a whim.
That gap is your entire budget. Give away more than 7.16 per prepaid order and you have made the order worse than leaving it as cash.
The ceiling
Expressed as a percentage of the order value, the break-even discount is 16.8%. At exactly that discount a prepaid order and a cash order are worth the same to you. Above it, you are paying customers to be less profitable.
That ceiling is specific to these figures and moves sharply with one of them. The lower your cash on delivery delivery rate, the more a prepaid order is worth and the more you can afford to pay for it:
| Your COD delivery rate | COD order value | Break-even prepay discount |
|---|---|---|
| 60% | 3.60 | 26.1% |
| 68% | 6.24 | 19.9% |
| 75% | 8.55 | 14.5% |
| 85% | 11.85 | 6.7% |
| 92% | 14.16 | 1.3% |
A store delivering 92% of its cash orders has almost nothing to gain from a prepay discount and should not offer one. A store delivering 60% should be willing to pay a great deal. This is why copying another seller’s percentage is meaningless.
The part most sellers miss: you pay the ones who were already prepaying
Suppose 8% of your orders already arrive prepaid with no incentive. Switch on a discount and those customers take it too. They were never going to pay cash, so for them the discount buys nothing at all. It is pure cost.
So the real question is not whether a prepaid order is better. It is how many cash customers have to switch before the switchers pay for the free riders.
| Discount | Prepaid order value | Prepaid share needed to break even | Monthly gain if share reaches 25% |
|---|---|---|---|
| 3% | 13.44 | 10% | +898 |
| 5% | 12.59 | 11% | +685 |
| 7% | 11.74 | 14% | +472 |
| 10% | 10.46 | 20% | +153 |
Read the third column carefully. At a 3% discount you need the prepaid share to reach the figure shown before the policy pays for itself. Below it, you have handed money to customers who changed nothing.
The fourth column shows what the same policy is worth on 1,000 orders a month if a quarter of your orders end up prepaid. The smaller discounts are better per switch; the larger ones buy more switches. Where the peak lands depends entirely on how responsive your customers are, which only a test can tell you.
How to run the test
Switch it on for everyone for a fixed period and compare whole weeks. Do not try to split traffic: a discount visible on a product page leaks, and a customer who sees it once and not again reads it as a trick.
- Record four weeks without the discount: orders, prepaid share, delivery rate, revenue collected.
- Switch on one discount percentage. Leave it alone for four weeks. Changing it mid-test tells you nothing.
- Compare prepaid share and total collected revenue, not order count.
- Then compare delivered orders. The whole point was fewer returns, so if returns did not fall, the switchers were people who would have accepted their parcel anyway.
Four weeks each way is the minimum because a month contains a payday. Prepayment responds to whether people have money in their account, which moves on a monthly cycle in every market.
Three ways to make the same discount work harder
Put it where the decision is made
A prepayment option that appears after the buyer has already decided to pay cash converts far worse than one shown as an alternative beside the main button, at the moment of ordering. That is a layout question, not a pricing one, and it is covered in the comparison of a one-page form against the standard checkout.
Say what it is in money, not percent
“Save 2.25” beats “Save 5%” in most cash on delivery markets, because the buyer does not have to do arithmetic to see the benefit. If your order values vary a lot, show the amount rather than the rate.
Use it where it is worth most
Your delivery rate is not uniform. Some regions refuse far more than others, and those are the regions where a prepaid order is worth the most. Offering a larger discount into your weakest regions and none into your strongest is closer to correct than one rate everywhere.
When not to offer one at all
| Situation | Why |
|---|---|
| Delivery rate already above about 90% | There is almost nothing left to buy; the discount is a price cut |
| Card use in your market is very low | Few customers can take it, and those who can were prepaying anyway |
| Your margin is under about 20% | The break-even discount is too small to change anyone’s behaviour |
| You have not measured your delivery rate | You cannot know the ceiling, so any number is a guess |
The short version
A prepaid order is worth more because it almost never comes back. The difference between that value and your cash order value is your whole budget, and on the figures here it works out at 16.8% of the order.
Set the discount below the ceiling, remember you are also paying everyone who already prepaid, test for four weeks against four weeks, and judge it on collected revenue and returns rather than on how many people took it.
If your delivery rate is the number you do not have yet, start there: it decides everything on this page. What a cash on delivery order really earns sets out the full per-order calculation.