What is CPA in marketing?
How is CPA calculated?
Divide what you spent by the conversions counted.
The word that matters is "counted". A conversion can be a purchase, a sign-up or a lead. The platform and your settings decide which. A CPA built on sign-ups says nothing about sales.
What does CPA hide?
Four things worth knowing:
- What counts. If the platform counts several event types as one conversion, CPA looks lower than your cost per order.
- Who they are. A returning customer who would have bought anyway still counts. Your cost per new customer is a different number.
- Whether the ads caused it. CPA counts credited conversions, not conversions the ads caused. Those can be very different numbers.
- What the order is worth. The same CPA is good on a high-margin order and bad on a low-margin one. CPA has no margin in it.
That last point is why CPA alone cannot say whether ads pay. Compare it with the margin on the order. Our view is that the margin to use is contribution margin before marketing: what is left per order after variable costs.
Worked example
Easily confused with
CPA vs CAC (customer acquisition cost)
What it costs to win one new customer, usually acquisition spend divided by new customers in the same period. CPA counts conversions. CAC counts customers.
CPA vs nCAC (new-customer CAC)
Acquisition spend divided by first-time customers only, leaving out repeat buyers.
CPA vs ROAS
Attributed revenue divided by ad spend. CPA is cost per conversion. ROAS is revenue per euro. Both come from the same platform credit.
CPA vs CTR
Clicks divided by impressions. It is an upstream measure, before any purchase.
Which of these numbers fits which decision is covered in ROAS vs MER vs POAS.
FAQ

Is a lower CPA always better?
No. A low CPA can come from cheap conversions that are worth little, or from sales that would have happened anyway. Check it against margin.
Is CPA the same as CAC?
No. CPA divides spend by counted conversions. CAC divides acquisition spend by new customers.
Let’s find what’s actually working.
Next step. Check what your ROAS looks like on explicit assumptions. Use the Is your ROAS real? calculator.