A campaign reports a 6x ROAS, sales are arriving, and the platform suggests increasing the budget. Before you do, open the actual orders behind that number.
ROAS—return on ad spend—compares attributed revenue with advertising spend. It is useful, but it does not show whether the attributed sales were retained, which products sold, how much contribution those orders produced, or whether the customers were new. A campaign order audit connects the dashboard number to the economics of the products that shipped.
Start with one campaign and one closed period
Choose a date range old enough for routine cancellations and returns to appear. Record the platform, attribution window, spend, attributed revenue, and reported ROAS exactly as shown.
Then export the store orders that can reasonably be matched to the campaign. Use completed orders and net sales after discounts, cancellations, and refunds. Keep platform-attributed revenue and store-matched net sales in separate columns; they answer different questions and may not agree.
The plain-language guide to what ROAS measures explains the basic formula and its limits. This audit begins after that calculation.
Reconcile the headline with retained sales
Use two figures:
Platform ROAS = platform-attributed revenue ÷ ad spend
Reconciled ROAS = matched net sales ÷ ad spend
Suppose an ad platform reports $9,000 in attributed revenue from $1,500 in spend. The displayed ROAS is 6x. Your store export shows $7,600 in matched net sales after discounts, cancellations, and refunds. The reconciled ROAS is about 5.07x.
That difference is not automatic proof that either system is wrong. Attribution rules, reporting delays, view-through credit, duplicate channel claims, taxes, shipping, returns, and order matching can all change the result. The point is to make the difference visible before it becomes a budget assumption.
For a broader explanation of those credit rules, review the small-business guide to marketing attribution.
Break the orders into product and offer groups
A blended ROAS can hide very different order economics. Group matched orders by the factors that materially change contribution:
- main product or product family;
- full-price order, bundle, or discounted offer;
- new or returning customer when the classification is reliable;
- shipping promotion or fulfillment method;
- unusually high return or cancellation behavior.
Do not create so many segments that every group becomes noise. Start with the two or three product or offer groups responsible for most matched net sales.
In the example, $5,200 came from a full-price hero product and $2,400 came from a discounted bundle. Both groups helped produce the same campaign-level ROAS, but they did not leave the same amount available after variable costs.
Calculate contribution before and after advertising
For each group, subtract the variable costs that follow the order. Depending on the business, that may include materials, packaging, payment fees, picking and packing, variable shipping support, marketplace fees, and expected refunds.
Contribution before advertising = net sales − variable order costs
Contribution after advertising = contribution before advertising − allocated ad spend
The hero-product orders produced $3,120 in contribution before advertising, equal to 60% of their $5,200 in net sales. The discounted bundle produced $680, or about 28% of its $2,400 in net sales. Together, the matched orders produced $3,800 before advertising.
After subtracting the $1,500 campaign spend, $2,300 remains. That is about 30% of matched net sales and still has to support fixed costs and profit.
The contribution margin guide for product businesses provides a consistent way to define variable costs. Use the same definition across campaigns so changes in advertising performance are not confused with changes in accounting treatment.
Build a one-page campaign order audit
| Review line | Platform report | Store and cost review | Question to answer |
|---|---|---|---|
| Ad spend | $1,500 | $1,500 | Is the period identical? |
| Attributed or matched sales | $9,000 | $7,600 | What explains the gap? |
| ROAS | 6.00x | 5.07x | Which figure supports the decision? |
| Contribution before ads | Not shown | $3,800 | Which products created it? |
| Contribution after ads | Not shown | $2,300 | What remains for fixed costs and profit? |
Add the attribution window, refund cutoff date, variable-cost definition, new-customer rule, and person who prepared the audit. Those notes make the table reusable instead of turning it into a one-time screenshot.
Check customer type without overstating certainty
A campaign that mostly reaches returning customers may still be useful, but it is doing a different job from one that brings in first-time buyers. Separate new and returning customers only when your store data supports the distinction.
Then calculate acquisition cost with all included campaign spending and the same customer rule. The customer acquisition cost guide explains why total acquisition spending and a clean denominator matter.
Do not assume every platform-labeled conversion is a newly acquired customer. Likewise, do not assume a returning-customer order would have happened without the ad. Label what the data can show and keep uncertainty visible.
Decide whether the campaign is ready for more budget
A strong audit does not produce an automatic yes or no. It creates a decision with guardrails.
Increase cautiously when reconciled sales are credible, contribution after advertising is acceptable, the important product groups have enough stock and capacity, and the customer mix matches the campaign’s purpose. Choose a small budget step and a date to repeat the audit.
Hold the budget when refunds are still developing, order matching is incomplete, one low-margin offer drives most sales, or available inventory cannot support more demand.
Reduce or redesign the campaign when retained sales fall below the required threshold, contribution disappears after advertising, or the offer creates volume without enough value to support the business.
Do not let one strong week become a permanent target. Seasonality, creative fatigue, product availability, promotions, and auction costs can change the result.
Keep these marketing metrics together
ROAS should sit beside matched net sales, contribution after advertising, new-customer share, refund rate, average order value, and inventory availability. No single number explains the whole campaign.
Save the one-page audit with the campaign name and review date. Repeat it after each meaningful budget change using the same definitions. Over time, the comparisons become more useful than a dashboard screenshot because they show what sold, what remained, and what operational constraint appeared next.
Frequently asked questions
Should I use gross sales or net sales when auditing ROAS?
Use net sales after discounts, cancellations, and refunds for the store-side review. Keep the platform’s attributed figure unchanged in its own column so the difference remains visible.
How long should I wait before closing the review period?
Wait long enough to capture the business’s normal cancellation and return pattern. Record the cutoff date and reopen the audit if material late returns appear.
Is contribution after advertising the same as net profit?
No. It still needs to support fixed costs such as payroll, rent, software, insurance, and other overhead before profit is known.
What if I cannot match every attributed order?
Report the matched share and the unmatched amount. Do not quietly treat incomplete matching as complete evidence or fill the gap with assumptions.
Does a strong audited ROAS mean I should scale immediately?
Not by itself. Confirm margin, customer mix, inventory, production capacity, cash timing, and a repeat-review date before increasing spend.
Practical takeaway
Pick one recent campaign that looks ready to scale. Reconcile attributed revenue with retained store sales, group the orders by product or offer, calculate contribution before and after advertising, and write down the operational constraint that could break first.
ROAS is a useful starting signal. The orders behind it determine whether more spend is a sound business decision.
