Blog · Finance metrics and formulas · Sales teams
Average ROAS is only comparable when revenue, spending and attribution definitions match. Calculate a weighted campaign average, test the contribution behind it and evaluate outside benchmarks without assuming a universal target.
Average ROAS is total attributed revenue divided by total advertising spending for a comparable set of campaigns. It is a useful reference only when the underlying sales, cost scope and attribution rules match. Whether that return is good for your business depends on the contribution those sales leave after variable costs.
A ratio of totals answers how much attributed revenue the combined advertising spending produced. An arithmetic mean answers what the average campaign ratio was when every campaign received equal weight. A median describes the middle observation in a distribution. These are different statistics even when a chart labels each one an average.
For your own portfolio, use:
Combined ROAS = Total comparable attributed revenue / Total comparable advertising spending
Google Ads describes ROAS in terms of conversion value per advertising dollar. Check what the exported value represents: sales revenue, an assigned lead value or another supplied business value. Only the first is directly comparable with a revenue-based benchmark.
The general revenue-versus-profit distinction belongs in marketing ROI vs ROAS. Here the task is to decide whether a claimed average is comparable and financially useful.
Before using an outside number in a budget review, keep a short comparison record. An industry name and publication date alone do not establish comparability.
| Field | What the reference needs to disclose | What to match in your file |
|---|---|---|
| Outcome value | Revenue or assigned conversion value | The same economic measure |
| Revenue treatment | Gross sales or net of returns and discounts | Matching adjustments |
| Spend scope | Media only or additional campaign costs | The same included expenses |
| Audience | Existing customers, new customers or a mixture | A comparable customer mix |
| Attribution | Eligible interactions, credit rules and window | An equivalent comparison basis |
| Population | Channel, geography, period and reporting sample | Relevant campaigns and dates |
| Statistic | Ratio of totals, mean, median or another measure | The matching calculation |
If a reference omits these details, describe the uncertainty rather than applying a numerical correction with no evidence. A published multiple may be useful background while still being unsuitable as a target for your business.
The window matters because later sales change the numerator. Google Ads defines a conversion window as the time after an ad interaction during which a conversion is recorded. A newly launched campaign with incomplete outcomes should not be judged against a fully observed historical cohort.
An illustrative US replacement-parts seller reviews completed campaign cohorts. The revenue below is net matched sales after agreed return and discount adjustments. Media costs reconcile to the advertising spending file. The campaigns use a common attribution policy, and the example is not an industry survey.
| Campaign group | Advertising spending (USD) | Net attributed revenue (USD) | ROAS |
|---|---|---|---|
| Brand search | 10,000 | 80,000 | 8.00x |
| New-customer search | 30,000 | 90,000 | 3.00x |
| Paid social | 20,000 | 80,000 | 4.00x |
| Total | 60,000 | 250,000 | 4.17x |
Each ratio divides revenue by advertising spending. Brand search is 80,000 / 10,000 = 8.00x; the combined result is 250,000 / 60,000 = 4.17x, rounded to two decimal places.
With spending in B and revenue in C, a spreadsheet calculates the campaign row as:
=IF(B2=0,"No spending",C2/B2)
The total is:
=SUM(C2:C4)/SUM(B2:B4)
Do not replace that total with =AVERAGE(D2:D4). Here the simple average is (8 + 3 + 4) / 3 = 5.00x. The smallest campaign has the largest ratio, so giving every campaign equal weight overstates the revenue produced per dollar across the portfolio.
Multiply each campaign's unrounded ROAS by its spending. The result must reproduce attributed revenue:
| Campaign group | Spending × ROAS | Reconstructed revenue (USD) |
|---|---|---|
| Brand search | 10,000 × 8 | 80,000 |
| New-customer search | 30,000 × 3 | 90,000 |
| Paid social | 20,000 × 4 | 80,000 |
| Total | 250,000 |
Dividing that reconstructed total by $60,000 of spending reproduces the combined 4.17x. The count-weighted and value-weighted comparison explains why the choice of weight changes a reported average.
Also reconcile the underlying sales to unique eligible transactions. Matching the arithmetic does not prove attribution is correct. A sale counted in separate channel reports can produce perfectly calculated ratios and still inflate a combined result.
Finance supplies the campaign-specific pre-ad contribution margin: net revenue less the agreed variable product, fulfillment and payment costs, before advertising. These margins are illustrative inputs from a cost schedule, not public sector averages.
| Campaign group | Net revenue (USD) | Pre-ad contribution margin | Pre-ad contribution (USD) | Advertising (USD) | Contribution after advertising (USD) |
|---|---|---|---|---|---|
| Brand search | 80,000 | 35% | 28,000 | 10,000 | 18,000 |
| New-customer search | 90,000 | 50% | 45,000 | 30,000 | 15,000 |
| Paid social | 80,000 | 30% | 24,000 | 20,000 | 4,000 |
| Total | 250,000 | 38.80% | 97,000 | 60,000 | 37,000 |
Pre-ad contribution equals revenue times the supplied margin. After-ad contribution subtracts advertising. Total pre-ad contribution is 28,000 + 45,000 + 24,000 = $97,000; the combined margin is 97,000 / 250,000 = 38.80%. The remaining contribution is 97,000 − 60,000 = $37,000, before fixed and other excluded costs.
Paid social has a higher ROAS than new-customer search but leaves less contribution in dollars. Neither the ranking nor the combined average is enough to allocate the next advertising dollar. Review incremental demand, capacity and the reliability of the margin assumptions before changing spending.
Use target ROAS to translate an approved margin and retained-contribution requirement into a business threshold. That target answers a different question from what other advertisers reportedly achieved.
Marketing operations owns the spending export, channel labels and attribution documentation. Sales operations confirms eligible customers and transactions. Finance reconciles net revenue, approves the contribution cost boundary and signs off on the target used in budget decisions.
Keep the original benchmark source, its publication period and the comparison notes with the review. Repeat the internal calculation using the same definitions. If a new return policy, measurement window or revenue rule changes the basis, show the break in comparability beside the result.
Bring your campaign spending export, matched sales file and finance-approved margin assumptions, plus any benchmark and its definitions. Covirage maps the columns, and its tools compute and check the weighted returns and supplied contribution comparison; the external AI model explains the result with every figure traced to its rows. Create a free Covirage account in about a minute, set up your business and bring a sample of your own data.
See customer profitability for the financial context around campaign returns. Covirage is not affiliated with Google.
No single revenue multiple is suitable for every business. A company with a lower contribution margin needs more revenue per advertising dollar to cover the same spending. Compare matched campaign cohorts and test the contribution remaining after advertising before adopting a target.
Use total comparable revenue divided by total advertising spending for the portfolio result. A simple average gives a small campaign the same influence as a large one. Keep campaign-level ratios for diagnosis, and reconcile the combined figure to the underlying totals.
Only after resolving overlapping attribution. Several platforms can claim credit for the same sale under their own rules. Reconcile transaction IDs and an agreed credit method against sales records, and label platform-reported figures separately when they cannot be combined without duplication.