What your CPA number is not telling you
Cost per acquisition is the most quoted metric in paid media and one of the least useful on its own. Here is what to look at beside it.
Every account review starts the same way. Someone opens a dashboard, points at cost per acquisition, and says it is either good or bad. Almost nobody says what it is good or bad relative to.
CPA on its own is a ratio with a hidden denominator. It tells you what you paid for a conversion event. It does not tell you whether that event was worth paying for.
The three questions CPA cannot answer
Was the conversion real? A form fill is not a customer. If your lead-to-close rate moves from 20 percent to 12 percent while CPA holds flat, your true cost per customer just went up 66 percent and your dashboard looks unchanged.
Was it incremental? Branded search converts beautifully and reports a wonderful CPA. A meaningful share of those people were going to find you anyway. That is not acquisition, it is attribution.
Was it the right customer? Two campaigns can post identical CPAs while one brings customers who reorder for three years and the other brings one-time discount hunters.
What to put next to it
The fix is not to abandon CPA. It is to stop letting it travel alone.
- Cost per qualified lead, not cost per form fill. This requires CRM data flowing back into the ad platform, which is a week of setup and pays for itself immediately.
- Contribution margin per acquisition, so the number accounts for what you actually keep.
- New-to-file rate, which separates growth from harvesting.
- Payback period, because a higher CPA that pays back in 40 days beats a lower one that takes nine months.
A practical starting point
Pick one campaign. Export 90 days of conversions. Match them against closed business in your CRM. You will almost always find that the campaign your dashboard ranked third is the one producing the revenue.
That exercise takes an afternoon and it changes budget allocation more than any bidding strategy change we have ever made.