Cost Per Acquisition (CPA)
Cost per acquisition is advertising spend divided by the number of conversions it produced over the same period, expressed as a dollar figure per outcome such as per booked job, per sale, or per signed customer.
Why it matters
It is the most direct measure of whether advertising is working - provided the conversions counted are the ones that make you money. A CPA calculated from the wrong conversion action tells a confident, precise, completely misleading story about performance.
What people get wrong
Comparing it to an industry benchmark instead of to the same account's own history. A roofing account with a $955 cost per acquisition reads as broken against a generic benchmark and as entirely normal against its own trailing average. The useful question is almost never "is this above average" - it is "has this changed", and changed against what baseline.
How it actually works
CPA is a simple division: total spend over a period divided by total conversions in that same period, using whichever conversion action or actions the account is reporting against. That simplicity is also its weak point - the number is only as meaningful as the conversion definition feeding it.
Platforms calculate a version of CPA automatically using whichever actions are marked to count in their own reporting, which is why the CPA shown in an ad platform's dashboard and the CPA a business calculates from its own booked-job records frequently disagree, sometimes by a wide margin. Both can be correct answers to slightly different questions.
Because CPA is an average over a period, a single expensive week or a single cheap week can move it substantially in a small account, which is why CPA read over a short window is a noisier signal than it looks.
What goes wrong with it in real accounts
The most damaging version is a CPA calculated from a conversion action that does not represent a real customer - a page view, a form load, or a duplicate submission counted as a separate job. The reported CPA looks excellent, spend keeps flowing toward whatever produced those cheap fake conversions, and the business sees no matching increase in actual work, usually attributed at first to a seasonal slowdown rather than a tracking problem.
The second is comparing CPA across channels or campaigns that use different attribution windows or counting settings, which produces an apples-to-oranges comparison that looks like a performance gap when it is a measurement settings gap. Reallocating budget based on that comparison moves money toward whichever channel happens to count more generously, not whichever channel actually produces more customers.
The third is treating a rising CPA as automatically bad without checking what changed upstream - a seasonal shift in competition, a broadened service area, or a legitimate move into more competitive but higher-value keywords can all raise CPA while improving the business, and cutting spend reflexively when CPA rises can eliminate exactly the growth an account was trying to produce.
How it relates to the other terms
CPA is calculated only from actions marked as primary conversions, which is why the same account can report two different CPA figures depending on which actions are included in that setting.
It sits alongside cost per lead as a related but distinct measure - cost per lead counts every enquiry regardless of quality, while CPA is typically applied to a further-down-funnel outcome, and confusing the two produces comparisons that look inconsistent for no real reason.
It is also meaningless without reliable conversion tracking behind it, and it says nothing about margin on its own, which is the gap that return on ad spend is meant to fill.
Frequently asked
What counts as a good cost per acquisition?
There is no universal figure, because it depends on the value of the job being acquired and the margin on it. A CPA that is profitable for a business selling large installations can be ruinous for a business selling small repairs. Judge it against the account's own history and against the value of the average customer, not against an industry number.
Why does my platform's reported CPA not match what my business is actually spending per customer?
The platform calculates CPA using whichever conversion actions it has been configured to count, which may include duplicate submissions, funnel steps, or leads that never became customers. Reconcile the platform's conversion count against your own booked-job or sales records to find the real figure.
Should I pause a campaign the moment its CPA rises?
Not automatically. Check whether the rise reflects a real efficiency problem or a legitimate shift, such as expanded targeting, higher-value keywords, or seasonal competition. A rising CPA alongside rising job value or volume can still represent an improving campaign.