What is cost per acquisition? CPA is the average amount you spend to win one conversion: a sale, a signup, a booked call, whatever counts as a win. Divide total ad spend by the number of conversions it produced. Spend $2,000, get 40 conversions, your CPA is $50. It is the single fastest way to tell whether a campaign is making money or quietly bleeding it.
The CPA formula (it really is this simple)
CPA = Total ad spend / Number of conversions
Run $5,000 of Google Ads in a month and book 100 demos? Your CPA is $50 per demo. That is the whole calculation. The hard part is not the math, it is being honest about what counts as a "conversion" and making sure your tracking captures it.
CPA goes by a few names depending on the room you are standing in. Paid search people say "CPA." Performance marketers say "cost per action." Finance might call it "customer acquisition cost," though strictly speaking CAC includes more than just ad spend (sales salaries, tooling, overhead). On this page we mean the ad-channel number: spend in, conversions out.
What is cost per acquisition, and why does it matter?
Clicks feel like progress. Impressions feel like reach. Neither pays your rent. CPA connects the money you put in to the outcome you care about, which is why it is the metric senior marketers watch first and the vanity metrics second.
Here is the logic that makes it powerful. If you know what a customer is worth to you, CPA tells you instantly whether a channel is profitable. Customer worth $400 and your CPA is $90? Scale it. CPA creeping toward $400? You are running to stand still. This is exactly the math that separates "we got a lot of traffic" from "we made money," and it is the conversation we have with every client before we touch a budget. If you do not know roughly what a customer is worth to you yet, that is the first thing to figure out, because every CPA target downstream depends on it.
What counts as a "good" CPA?
There is no universal good CPA, and anyone who quotes you one without asking about your business is selling, not advising. A good CPA is any CPA comfortably below what a customer is worth to you.
The number that governs your CPA is your target CPA: the most you are willing to pay for a conversion while still making money. You set it by working backward from value.
- A customer is worth, say, $600 in profit over their lifetime.
- You want at least a 3:1 return on acquisition spend.
- Your target CPA is roughly $200.
Those numbers are an illustration, not a benchmark. Plug in your own. The point is that "good" is relative to your economics, not to an industry chart. A $300 CPA is a disaster for a $40 product and a steal for a $50,000 enterprise contract. It is also why we never quote a flat "good CPA" on a sales call before we have seen your margins; for how we think about budget and fees generally, our pricing lays it out without the runaround.
Target CPA bidding
Most ad platforms let you hand them that target directly. Target CPA bidding (in Google Ads, part of the Smart Bidding family) tells the algorithm: get me as many conversions as you can at roughly this cost. The machine then adjusts bids in real time per auction to chase that goal. It works well once a campaign has enough conversion history to learn from, and poorly when it is starved for data or pointed at a target that is unrealistically low. Set the target too aggressively and the system simply stops spending. It is a powerful lever, not a magic one.
CPA vs. the metrics people confuse it with
CPA gets mixed up with a few neighbors. Here is the clean breakdown.
| Metric | Measures | When you reach for it |
|---|---|---|
| CPA (cost per acquisition) | Spend per completed conversion (sale, signup, booking) | The bottom-line efficiency check on any channel |
| CPL (cost per lead) | Spend per lead captured, before they buy | Top of a longer funnel, where a "lead" is not yet a customer |
| CPC (cost per click) | Spend per click on the ad | Auction-level cost; an input to CPA, not an outcome |
| CPM (cost per mille) | Spend per 1,000 impressions | Awareness and reach buys, where the goal is eyeballs, not action |
| ROAS (return on ad spend) | Revenue earned per dollar spent | The revenue-side companion: CPA tells you cost, ROAS tells you return |
CPA vs. CPL specifically
This is the one that trips people up most. A lead is not an acquisition. Cost per lead measures the price of a hand raised: a form fill, a download, an email. Cost per acquisition measures the price of the thing that pays you. If 4 leads close out of every 10, your CPA is more than double your CPL. Optimize to CPL alone and you can rack up cheap leads that never buy, congratulating yourself while the revenue line stays flat. We have seen plenty of accounts "win" on CPL and lose on the only number that matters. Always tie your funnel back to CPA.
CPA, CPC, and CPM
These three are not competitors; they are different floors of the same building. CPC is what you pay to get someone onto your page. CPM is what you pay for the impression that earned the click. CPA is what you pay once that visitor does the thing you wanted. A cheap CPC means nothing if those clicks never convert, which is why CPA sits at the top of the stack: it absorbs every cost below it and tells you whether the whole chain paid off.
CPA and ROAS together
CPA tells you what a conversion cost. ROAS tells you what it returned. You want both. A low CPA on tiny-order-value sales can still lose money, and a healthy ROAS can hide a CPA that will not scale. Read them as a pair.
How attribution quietly changes your CPA
Here is the part most "what is CPA" explainers skip, and it is the part that bites. Your CPA is only as honest as your tracking. The same campaign can show wildly different CPAs depending on what you let count as a conversion and how long after the click you keep counting it.
That window of time is the attribution window. A 1-day window credits only conversions that happen within a day of the click. A 30-day window credits conversions up to a month later. Widen the window and your reported CPA drops, because the same spend now gets credit for more conversions. Nothing about the underlying performance changed. Only the accounting did.
This matters enormously when you compare channels or compare your numbers to an old agency's. If one report uses a 7-day click window and another uses 30-day click-plus-view, you are not comparing the same metric, and the "better" CPA might just be the more generous ruler. Before you trust any CPA number, ask what attribution window and model produced it. We put that on the table up front, because a CPA you cannot reproduce is a CPA you cannot manage.
There is a deeper version of this question too: even a perfectly tracked conversion might have happened without the ad. That is the gap between attributed CPA and true cost, and it is why incrementality testing exists. You do not need it on day one, but once you are spending at scale it is the difference between a CPA that looks good and a CPA that is real.
How to lower your CPA (the levers that move it)
Lowering CPA is mostly about wasting less and converting more. In rough order of impact:
- Tighten targeting. Cut the audiences, placements, search terms, and geographies that spend without converting. Most accounts have roughly a fifth of their spend doing none of the work. Find it and kill it; our walkthrough on how to reduce Google Ads wasted spend shows exactly where it usually hides.
- Fix the landing page, not just the ad. Half of a bad CPA lives after the click. A faster page, a clearer offer, and a shorter form often beat any bidding tweak. This is its own discipline; pushing the post-click conversion rate up even a point or two drops CPA without spending an extra dollar, which is the whole premise of conversion optimization.
- Improve relevance and quality. On search, higher ad relevance lowers your cost per click, which flows straight down into a lower CPA. Better creative does the same on social and programmatic. Relevance is not cosmetic: the platforms reward it with cheaper auctions, so the gain compounds.
- Feed the algorithm clean conversion data. Smart bidding is only as good as the signals it gets. Server-side tracking through a conversion API and accurate conversion values let Target CPA bidding hit the target instead of optimizing toward noise. Garbage in, expensive out.
- Match the conversion to the funnel stage. Optimizing toward a deep, rare action when you have little data can starve the system. Sometimes the right move is a mid-funnel signal that gives the algorithm enough volume to learn, then tightening later.
- Exclude what already converted (and what never will). Frequency caps, negative keywords, and audience exclusions stop you paying twice for the same person or paying at all for the wrong one.
None of these is a silver bullet. CPA comes down through disciplined, unglamorous optimization, week over week, not one heroic setting change. That is the work, and it is the PPC management work we do.
Stop guessing at your CPA. Start managing it.
If your current reporting shows a CPA you cannot explain, reproduce, or trust, that is not a metric problem. It is an agency problem. We run Google Ads, social, and programmatic with CPA front and center, the attribution spelled out, and the optimization shown to you in plain English, not buried in a dashboard you never log into.
Want a straight read on whether your paid channels are profitable? Book 30 minutes or email admin@moonsauceagency.com. No pitch, no pressure, just the real numbers.
Keep going: ROAS (the revenue-side companion to CPA), CPC and CPM (the costs that feed it), and attribution windows (the setting that secretly changes your CPA). Or browse the full marketing glossary.
