Model refunds and customer acquisition before you buy ads
Make the cost of refunded orders and paid traffic visible in your product economics.
An order is not the same as a retained customer payment
If you receive 100 orders and expect five to be refunded, begin with 100 initial orders and a 5% refund assumption. Do not enter 95 orders and also apply 5%; that would count the expected loss twice.
The model treats the refund percentage as the expected share of fully refunded, equal-value orders. Mixed baskets and partial refunds require a more detailed ledger. A forecast can use fractional expected refunds even though actual customer orders are whole transactions.
The fee does not always come back
Some providers return certain fee components while retaining others. That also depends on whether you use your own connected payment account. For this reason the planner provides a separate fee-retention assumption: 100% means the modeled transaction fees are retained on refunded orders; 0% means they are fully returned.
The default is a conservative planning assumption, not a claim that every platform retains every fee. Check the provider’s current refund policy and your own account before changing it. A single percentage can approximate a blended outcome but cannot reproduce individual settlement entries.
Calculate acquisition cost from initial orders
If an ad campaign costs $300 and generates 60 initial orders, acquisition cost is $5 per order. Enter $5, not the cost per click. If the traffic produces no orders, there is no meaningful per-order average; put the campaign’s committed spend in fixed overhead for the loss scenario.
The model charges acquisition cost even on refunded orders because an ad network typically does not reimburse the click that led to them. Avoid counting the same campaign in both overhead and acquisition cost.
A worked downside scenario
Consider 100 orders at $25, with $2 in hypothetical transaction fees, $5 acquisition cost and $1 delivery cost per order. At a 10% refund rate and full fee retention, retained revenue is $2,250. Fees are $200 and variable operating costs are $600. Contribution is $1,450 before fixed expenses.
With no refunds, contribution would be $1,700. The $250 difference is real even though the initial order count and ad budget were unchanged. Save both cases to see how much room your offer has for uncertainty.
Use a range until you have evidence
Try low, base and high acquisition assumptions. A fee comparison may save cents while a weak acquisition channel loses dollars on every order. Improving the offer, customer fit or repeat purchases can matter more than changing platforms.
Do not treat the resulting margin as a guarantee that ads will work. Track actual orders, refunds and advertising spend after launch, then replace assumptions with observed numbers.