Resources / Measurement & strategy
Influencer campaign CPM, ROAS and ROI: the useful calculations
The same figures can tell opposite stories depending on the costs included and the share of sales actually caused by the campaign.
Four metrics, four different questions
| Metric | Calculation | What it tells you |
|---|---|---|
| Media CPM | Media spend ÷ ad impressions × 1,000 | Distribution cost per 1,000 impressions; not the full campaign cost |
| Cost per 1,000 observed impressions | Full campaign cost ÷ observed impressions × 1,000 | Descriptive benchmark, especially when organic and paid are combined |
| Media ROAS | Attributed revenue ÷ media spend | Attributed revenue per 1 unit of ad budget; not a profit |
| Estimated ROI | (Margin on incremental sales − full cost) ÷ full cost | Estimated profitability, if incrementality and margin are known or reasonably estimated |
Standard CPM is calculated on ad impressions, as explained by Google Ads. Standard ROAS is based on an attributed conversion value divided by ad cost; it shouldn’t be renamed “ROI”. Detailed definitions and the distinction with incremental ROAS are in Google Ads help on Conversion Lift.
Fictional example: a creator and paid social campaign
To illustrate the method, assume creator fees of €2,400, production of €600 and €2,000 of paid distribution, a total of €5,000 in campaign cost. The creator’s post generates 140,000 reported impressions and the ads 110,000 impressions. Tracking attributes 180 orders to the campaign, with an average basket of €80. The assumed contribution margin after variable costs is 40% of revenue, or €32 per order. All these amounts are fictional, excluding tax and before fixed costs; refunds, shipping costs, discounts or other items must be added if they aren’t already included in the margin.
- Media CPM: 2,000 ÷ 110,000 × 1,000 = €18.18.
- Full cost per 1,000 observed impressions: 5,000 ÷ (140,000 + 110,000) × 1,000 = €20. This total adds two types of impressions, may count the same person several times and doesn’t measure unique reach.
- Cost per attributed order: 5,000 ÷ 180 = €27.78.
- Media ROAS: (180 × 80) ÷ 2,000 = 7.2. This ratio ignores the €3,000 of creative and doesn’t prove the 180 orders are incremental.
Links tagged with utm_source, utm_medium and utm_campaign can help identify visits; Google Analytics describes how to use them. On their own, they don’t establish that a sale was caused by the campaign.
Break-even: how many extra orders?
With €32 of contribution margin per order, the threshold is 5,000 ÷ 32 = 156.25, i.e. 157 extra orders to cover the campaign cost. This threshold depends on the real margin, the costs included and the period analysed. An awareness or lead-generation project needs other criteria before judging its value.
Swipe the chart to read it on mobile.
The real sticking point: are attributed sales incremental?
If all 180 attributed orders were incremental, the margin would be 180 × 32 = €5,760. The balance after campaign cost would be €760, an estimated ROI of 15.2%: (5,760 − 5,000) ÷ 5,000. But if only 60% were truly incremental, you’d count 108 orders, a margin of €3,456 and a balance of −€1,544, i.e. −30,9 %. The 60% rate is just a sensitivity assumption, not a measurement.
To estimate the share caused by the campaign, you need a suitable method: control group, lift study, geographic comparison or another approach whose biases are examined. Google Ads presents the principle of Conversion Lift. Without reliable incrementality measurement, report attributedorders and revenue, with a range of scenarios, instead of announcing a certain ROI.
What if the campaign targets leads rather than sales?
Follow the chain in order: impressions, clicks, observed visits, enquiries, qualified leads, signed customers. The CTR is the number of clicks divided by ad impressions; Google Ads gives the official definition. A page’s conversion rate can be calculated as qualified enquiries ÷ measured visits, keeping the same period and a stable definition of “qualified”.
For a cost per qualified lead, divide the full campaign cost by the number of qualified leads, and state that this figure differs from the CPA shown by a platform that only counts its media spend. To estimate profitability, then start from the observed close rate and the contribution margin per customer, with a range rather than a close rate assumed to be certain. Without enough history, the journey steps are signals, not a reliable ROI forecast.
Questions to ask before any comparison
- Are the amounts in the same currency, over the same period, and all excluding or all including tax?
- Does the cost include creative, rights, products sent, agency, media and measurement fees?
- Do organic and paid impressions use a comparable definition? What is the unique reach?
- Does revenue include returns, cancellations and customers who would have bought without the campaign?
- What contribution margin remains after variable costs, and which method estimates the incremental effect?
A good wrap-up sets out the answers, even when they’re incomplete. To choose metrics before distribution, also see our guide to measuring an influencer campaign.