Discount Ad Campaigns: Check the Margin Before You Launch

In the fictional example below, taking 20% off a $100 product cuts the advertising allowance from $27.70 to $8.30 per order. The offer looks modest to a shopper, but it changes which campaign the business can afford to approve.
We would reject that offer under the store’s current contribution target and test a smaller saving instead. Here is the calculation, the approval record and the evidence we would need before spending again.
Set the decision before choosing the saving
A promotion needs a purpose specific enough to challenge. Introducing a product to first-time customers, encouraging a two-item purchase and clearing an older range are different decisions. They should not inherit the same discount simply because a sale is approaching.
Our fictional store wants to test whether a selected desk organiser can attract first-time buyers while leaving at least $20 per eligible order after advertising towards overheads and profit. That $20 is the owner’s chosen contribution requirement, not a universal margin benchmark or net profit figure.
For reporting, a first-time customer means someone with no previous completed purchase in the store’s customer records. A new email address alone is insufficient evidence. Uncertain identities stay unclassified, and returning customers remain visible separately.
The offer will be available to all buyers of the selected item. Its commercial purpose is acquisition, but we will not pretend that restricting an advertising audience makes every order a new customer.
Calculate what the discounted order leaves for advertising
All figures are fictional USD planning assumptions per placed order. The organiser normally sells for $100 with delivery included. Product cost is $34, picking and packaging $4, outbound delivery $6, and the assumed payment fee is 3% of the amount collected plus $0.30. These are example costs, not provider rates. Sales tax is outside this model.
We also allow $5 per order for the expected net effect of returns, refunds and recoveries, separate from the listed costs. Keep that reserve at $5 for each offer while comparing them; replace it with actual net costs later rather than deducting both.
Full price: $100 − $34 − $4 − $6 − $3.30 − $5 = $47.70 before advertising. Retaining $20 leaves a $27.70 ad-cost allowance.
20% off: $80 − $34 − $4 − $6 − $2.70 − $5 = $28.30 before advertising. Retaining $20 leaves only $8.30 for advertising.
10% off: $90 − $34 − $4 − $6 − $3 − $5 = $38 before advertising. Retaining $20 leaves an $18 ad-cost allowance.
The 20% saving removes $19.40 of contribution before ads. Most fulfilment costs remain, so the amount available for advertising falls by about 70%, rather than 20%.
Revenue-based targets must change too. At full price, $100 ÷ $27.70 gives a contribution-target ROAS of approximately 3.61. With 20% off, $80 ÷ $8.30 gives approximately 9.64. For 10% off, $90 ÷ $18 gives 5.00. These ratios preserve the chosen $20 contribution; they are higher than thresholds that merely cover the listed order costs and ads. Use our break-even ROAS guide to separate those questions, and use the same revenue basis when comparing reports.
Choose the offer by contribution, not order count
Suppose the store’s comparable full-price advertising has cost $15 per placed order. That fictional reference gives us a starting assumption, not evidence that a promotion will repeat it.
At $15 per order, the 20% offer would leave $28.30 − $15 = $13.30, missing the target by $6.70. The 10% offer would leave $38 − $15 = $23, giving a $3 buffer. We can justify testing the smaller offer within a cap; the deeper discount needs a different case.
Consider the volume claim directly. Ten orders at 10% off would leave $230 after $150 in advertising. At 20% off and the same $15 acquisition cost, it would take 18 orders to exceed that contribution: 18 × $13.30 = $239.40. Seventeen would leave only $226.10. Those 18 orders would also require $270 in advertising, beyond the test budget below.
This does not predict either offer’s demand. It shows what “we will make it up in volume” requires before we accept that argument.
A bundle could be worth modelling, but add the second item’s cost, packaging and delivery changes first. Free shipping provides no new saving here because delivery is already included. Neither alternative earns approval merely by increasing basket value.
Write an approval record that can say no
Our fictional owner and campaign manager agree the following record before producing the ads:
20% offer: declined. Its $8.30 advertising allowance is below the $15 planning assumption. Reconsider only with credible evidence of lower acquisition cost, lower order costs or an explicitly revised contribution requirement.
10% offer: limited test approved. Selected organiser at $90, delivery included, for three days; other products excluded. Retain the $20 contribution requirement.
Exposure: $240 maximum media spend, with an $80 daily planning cap and no automatic extension. The owner accepts that the authorised amount could be spent without enough orders.
Capacity: 60 eligible units allocated, with dispatch capacity confirmed by the store manager.
Ownership: the campaign manager checks spend and order costs each morning and afternoon; the store manager reports stock, fulfilment problems and refunds. The owner authorises any further allocation.
A clearance campaign could deliberately accept less than $20 to release cash or avoid a documented disposal cost. That would need its own approved contribution floor and comparison with keeping or disposing of the stock. We would not quietly use clearance economics to justify this acquisition campaign.
Prove the promise works at checkout
The customer-facing offer now needs to match the approved one: “10% off the selected desk organiser. Delivery included. Other products excluded.” Add the actual start and end dates, eligible variants and any code conditions consistently across the ad, product page and checkout.
Before launch, the store manager completes one controlled eligible test order and checks an excluded basket. Confirm the final amount is $90, delivery treatment is correct, the intended variant is available and incompatible discounts cannot create an unplanned saving.
The campaign manager then checks that purchase records contain actual discounted receipts and unique order IDs. Name who removes the offer and stops spending when it expires. Our product page readiness checklist helps resolve delivery, variant and mobile checkout problems before traffic arrives.
Read the test without inventing additional sales
Extend the fictional scenario: the test reaches its $240 cap and records 16 eligible campaign-attributed orders. The provisional advertising cost is $240 ÷ 16 = $15 per order. Contribution after media is (16 × $38) − $240 = $368, or $23 per order, against a $320 total requirement.
Order records identify 12 first-time customers and four returning customers. Spending per first-time customer is therefore $240 ÷ 12 = $20. Do not compare that different denominator directly with the $18 allowance calculated across all eligible orders.
The contribution calculation passes provisionally. The acquisition argument still needs scrutiny. Attribution connects orders with advertising under a reporting rule; it does not establish that all 16 orders happened because of the campaign. First-time buyers are not automatically incremental either.
Compare total store contribution, full-price sales and customer mix with a suitable reference period. Investigate other promotions, traffic changes and timing differences. A planned comparison group may strengthen a future test; a simple before-and-after comparison cannot isolate the discount’s effect. Record uncertainty rather than upgrading attributed sales into proven additional revenue.
Reconcile the costs before approving a repeat
Stop immediately for an incorrect offer, failed checkout or unavailable advertised stock. Stop new spending at the cap even when results look encouraging. If acquisition costs exceed the allowance, review the completed and pending orders before authorising more; unresolved outcomes cannot become assumed profitable sales.
After the store’s relevant returns handling period, reconcile refunds, recoveries and extra costs against the $5 reserve. Keep an allowance for unresolved cases. In this example, the apparent $3 buffer disappears if net returns impact rises from $5 to $8 per order. At $9, contribution falls to $19 per order, or $304 across 16 orders, below the $320 requirement.
The repeat decision therefore needs reconciled contribution, reliable customer classifications and a credible explanation of what the offer added. Passing a provisional ROAS target alone is insufficient.
Our Meta Ads management work connects the offer, creative and purchasing economics before a promotion goes live. Book a strategy call with your selling price, order costs and contribution requirement so we can discuss which offer is worth testing.
Comments