Cost per acquisition is the number that decides whether paid media is a growth engine or a slow leak. It is also the number most teams track at the account level and almost never where it matters.
An account-wide CPA hides everything useful. The work happens when you break it down by campaign, by product, and by customer type.
What CPA actually is
The formula is straightforward:
Total spend ÷ Number of acquisitions
The complications sit in both terms. Total spend should include management fees and creative production, not just media, if you want a number that reflects reality.
And “acquisitions” has to mean the same thing every month. A CPA that counts leads one month and customers the next is not a metric, it is a moving target.
Decide early whether you are measuring cost per lead or cost per paying customer. Both are valid; mixing them is not.
Setting your target CPA
The target does not come from benchmarks. It comes from your own margin.
Start with contribution margin per sale: price minus product, shipping, processing, and any other variable cost.
If your contribution margin is 40 dollars, any CPA below 40 is technically profitable on the first sale. But first-sale breakeven leaves nothing for fixed costs, so the working target sits lower.
A common approach is targeting a CPA of 50% to 70% of contribution margin, leaving room for overhead and profit. In the example, 20 to 28 dollars.
Then adjust for repeat purchase. If customers typically buy three times, your allowable acquisition cost rises substantially, which is what lets well-run subscription and consumable businesses outbid competitors.
The levers that actually move it
Conversion rate on the landing page
The fastest lever available and the most ignored. Doubling conversion rate halves CPA without touching a single campaign setting.
Most paid traffic lands on pages built for browsing rather than for the specific promise the ad made.
Offer strength
CPA is partly a function of how compelling the offer is. A free trial, a guarantee, or a clearer value proposition moves CPA more than most bid adjustments.
Audience and keyword quality
Cutting the lowest-intent traffic raises CPA efficiency immediately. Negative keywords on search, exclusions on social.
Creative
On social especially, creative is the primary variable. A worn-out asset raises costs silently over weeks, and most accounts run assets long past their useful life.
Bidding strategy and data quality
Automated bidding works well when conversion data is clean and plentiful. It works badly when events are misconfigured or volume is too low to learn from.
Funnel stage mix
Cold-audience campaigns will always show higher CPA than retargeting. Judging them against the same target leads to switching off the campaigns that feed the funnel.
Where to look first when CPA rises
Work through these in order rather than changing several things at once.
Check measurement before anything else. A duplicated conversion event or a broken tag produces CPA movement that has nothing to do with performance.
Compare against the same period last year, not last month. Seasonality explains a large share of apparent CPA problems.
Look at creative fatigue: frequency climbing while click-through falls is the signature pattern.
Check whether competitors entered the auction. Rising CPC with stable conversion rate points outward rather than inward.
And review whether the offer or price changed. A price increase raises CPA mechanically, and teams frequently forget to account for it.
Attribution makes CPA look worse than it is
Platform-reported CPA and real CPA diverge, and the gap has widened with tracking restrictions.
Part of your conversions go unattributed entirely. Those sales happened, but no channel claims them, which inflates reported CPA across the board.
The practical check is blended CPA: total spend across all channels divided by total new customers. If blended CPA is healthy while platform CPA looks alarming, the problem is measurement, not performance.
Run that comparison monthly. It is the single best defense against switching off campaigns that are quietly working.
The mistakes that raise CPA quietly
Optimizing toward the wrong event. Campaigns set to maximize clicks or page views will cheerfully deliver traffic that never converts.
Scaling too fast. Large budget jumps push the system into lower-intent audiences before it has learned.
Running too many campaigns on too little budget. Each one stays below the data threshold needed to optimize.
Ignoring the post-click experience entirely, which is where most of the loss occurs.
And treating CPA as a single number rather than a distribution. The average hides the campaigns that are profitable and the ones subsidizing them.
A 30-day plan to bring it down
If you want a sequence rather than a list of levers, this one produces measurable movement within a month.
Week one: verify the measurement. Confirm each conversion event fires once, carries a value, and reconciles with your back office. Establish baseline CPA by campaign in writing.
Week two: cut waste. Review search terms and add negatives, exclude geographies you do not serve, and pause campaigns below the data threshold, consolidating their budget.
Week three: fix the landing page. Match the headline to the ad promise, remove fields nobody uses, and test on a mid-range phone. This is usually the largest single gain.
Week four: refresh creative. New assets for anything with climbing frequency and falling click-through, testing a different angle rather than a different color.
Then hold everything steady for two weeks and read the result. Accounts that change five variables at once learn nothing from any of them.
What CPA will not tell you
Two blind spots are worth naming, because both lead to confident wrong decisions.
CPA says nothing about customer quality. Two channels at identical CPA can deliver customers with very different retention and order values, and the cheaper one is frequently worse.
And CPA ignores everything that happens before the click. Brand searches, direct traffic, and organic visits often rise because of paid activity that never gets credited.
Track customer quality by cohort and watch total new customers against total spend. Those two habits catch most of what CPA alone misses.
Building a CPA you can act on
Break it down three ways and it becomes a management tool rather than a scorecard.
By campaign and channel, to decide where budget belongs. By product or line, since margins differ and a single target rarely fits all of them.
And by new versus returning customer, because acquisition and retention economics are different and should never share a target. We cover the profitability calculation in what ROAS is.
At Conqr we set CPA targets per line from contribution margin, then work the landing page before the bidding, because that is where the leverage usually is.
Frequently asked questions
What is a good cost per acquisition?
One comfortably below your contribution margin, with room for fixed costs. There is no cross-industry benchmark worth applying to your business.
Should CPA include agency fees?
For business decisions, yes. For campaign optimization, media-only CPA is the working number. Track both and be clear which one you are quoting.
Why is my CPA higher on cold audiences?
Because you are paying to reach people with no prior interest. That is the cost of filling the funnel, and it should be budgeted separately from retargeting.
How quickly should I react to a CPA increase?
Give it at least a week unless the spike is severe. Daily CPA is noisy, and reacting to noise restarts the learning phase repeatedly.
Can CPA be too low?
Yes. A very low CPA often means you are only harvesting existing demand and leaving growth unclaimed. Check whether volume is also where you want it.
Want to see where your acquisition cost is actually going? Let’s talk.