ROI vs. ROAS vs. CPA: Formulas and a Worked Example
March 17, 2026
Author: Shusaku Yosa
ROAS measures attributed revenue per unit of advertising spend. CPA measures the cost of a defined acquisition. ROI measures profit relative to the investment being evaluated. They describe different parts of performance and should be used together.
A campaign can have a low CPA and still lose money. It can also have a high ROAS while contributing little profit. The difference becomes clearer when all three metrics use the same example.
The formulas and what they mean
Metric | Formula used here | Question answered |
|---|---|---|
ROAS | Attributed revenue ÷ ad spend | How efficiently did ad spend generate credited revenue? |
CPA | Ad spend ÷ defined acquisitions | What did each acquisition cost? |
Marketing ROI | Profit after included costs and ad spend ÷ ad spend | How much profit remained per unit of media investment? |
Multiply ROAS or ROI by 100 to express a percentage. CPA is a currency amount. ROI has several conventions: some organizations include production, agency fees, or other investment in its denominator. Label the convention instead of comparing incompatible figures.
One hypothetical campaign, three results
Assume a campaign has $2,000 in media spend, 100 purchases, $10,000 in attributed revenue, $6,000 in product costs, and $1,000 in other included variable costs.
Calculation | Result |
|---|---|
ROAS: $10,000 ÷ $2,000 | 5x / 500% |
CPA: $2,000 ÷ 100 | $20 per purchase |
Profit after included costs: $10,000 − $6,000 − $1,000 − $2,000 | $1,000 |
Marketing ROI: $1,000 ÷ $2,000 | 50% |
These figures are illustrative, not a business case or industry benchmark. Fixed overhead is excluded. The 500% ROAS does not mean a 500% profit return: most revenue pays for the product and other costs.
Why identical ROAS can hide different profitability
Suppose another campaign has the same $10,000 revenue and $2,000 spend but $8,500 in non-ad costs. Its ROAS remains 500%, while the contribution after advertising is negative $500. Under the same ROI definition, its return is −25%.
This is why a single ROAS target may be misleading across products with different margins. Work out the break-even and target ROAS for the relevant cost structure.
Why the lowest CPA may not be the best option
An acquisition must be defined. A download, free trial, qualified lead, sale, and retained subscriber are not interchangeable.
For B2B campaigns, track the progression from inquiry to qualified opportunity and closed business. A cheap inquiry that rarely qualifies can cost more per opportunity than an expensive inquiry from a suitable buyer. Compare groups after similar time has elapsed so recent leads are not penalized for an unfinished sales cycle.
Choose the metric for the decision
Decision | Start with | Also check |
|---|---|---|
Compare purchase acquisition costs | Purchase CPA | Order value, margin, returns |
Monitor revenue efficiency | ROAS | Total revenue, spend, profit |
Evaluate a broader investment | Clearly defined ROI | Cost scope, timing, cash recovery |
Increase media budget | Marginal expected return | Saturation, capacity, uncertainty |
Historical average performance does not guarantee the same return on the next dollar. Increasing spend may reach different audiences or more expensive inventory.
Align the data before presenting a comparison
- Use the same reporting period, currency, and tax basis.
- Specify the conversion event and treatment of cancellations.
- Document whether costs mean media only or a wider investment.
- Use a consistent attribution source and avoid double counting platforms.
- Separate observed repeat revenue from forecast lifetime value.
Attribution is an accounting rule for assigning credit, not proof of causation. See the attribution guide before treating every credited sale as an incremental sale.
Put the metrics into an operating report
Show spend, acquisitions, revenue, included costs, and profit alongside the ratios. This lets readers distinguish a rate improvement from a volume decline or a change in product mix.
If performance deteriorates, use the ad cost-effectiveness diagnostic to separate measurement problems from traffic, conversion, and margin problems. The right response depends on the cause, not on which ratio looks worst.
