Three numbers that answer different questions
Paid advertising reports can be crowded with figures. Three of them do most of the work in judging whether spending makes sense:
- Cost per click (CPC): what you pay, on average, for one click.
- Cost per acquisition (CPA): what you pay, on average, for one result, such as a sale or a lead. It is also called cost per conversion or cost per action.
- Return on ad spend (ROAS): how much revenue you get back for each unit of money spent on ads.
All the examples below use made-up, round numbers for illustration. They are not benchmarks, and real campaigns will differ.
Cost per click
CPC is total ad cost divided by the number of clicks.
If a campaign spends 200 and receives 100 clicks, the CPC is 200 / 100 = 2 per click.
CPC tells you how expensive attention is, but nothing about whether that attention is useful. A cheap click that never converts is worse than an expensive click that often does. Treat CPC as an input to the other two figures, not a target on its own.
Conversion rate
Between a click and an acquisition sits the conversion rate: the share of clicks that complete the action you care about.
If those 100 clicks produce 5 sales, the conversion rate is 5 / 100 = 5 percent.
Conversion rate depends on the ad, the offer, the price, the audience and especially the page people land on. It is the number most within your control after the campaign is live, and it is where changes to a landing page or call to action show up.
Cost per acquisition
CPA is total ad cost divided by the number of conversions.
Using the same example: 200 spent, 5 sales, so CPA = 200 / 5 = 40 per sale.
You can also reach it from the other two figures: CPA = CPC / conversion rate. Here that is 2 / 0.05 = 40. The two routes agree, which is a useful check when your data comes from different reports.
This relationship explains why a rise in click price is not always bad news. If CPC rises to 3 but the conversion rate rises to 10 percent, CPA becomes 3 / 0.10 = 30, which is lower than before.
Is 40 a good CPA?
It depends entirely on what a result is worth. Suppose each sale gives a gross profit of 60 after product costs. A CPA of 40 would leave 20 per sale before other expenses such as staff time, shipping and software. If the gross profit were 35, a CPA of 40 would lose money on each sale, at least on the first purchase.
That is why it helps to decide a maximum acceptable CPA before launching. Customer lifetime value is relevant here: a business with repeat buyers may reasonably accept a higher first-purchase CPA, but only if the repeat behavior is real and measured, not assumed.
Return on ad spend
ROAS is revenue from ads divided by ad cost.
Say the 5 sales each bring in 100 of revenue, so 500 in total. With 200 of ad spend, ROAS = 500 / 200 = 2.5. This is sometimes written as 250 percent, or “2.5 to 1”.
ROAS looks at revenue, not profit, so it can mislead on its own. The break-even ROAS depends on your margin:
break-even ROAS = 1 / gross margin
If the gross margin on those sales is 40 percent (0.40), break-even ROAS is 1 / 0.40 = 2.5. In this illustration the campaign would roughly break even on gross profit: 500 in revenue gives 200 in gross profit, which equals the ad spend of 200. Any costs beyond that would put it in the red.
ROAS versus ROI
The two terms are often confused. ROAS compares revenue with ad spend. Return on investment compares profit with total cost. As an illustration, if a campaign costs 100 and earns 250 in profit attributable to it, the ROI is (250 - 100) / 100 = 150 percent. ROAS ignores product costs, staff time and other overheads, so a healthy-looking ROAS can still hide a loss.
Practical cautions
- Attribution is imperfect. Platforms and analytics tools may credit conversions differently, and some sales happen days later or through a different channel. Compare reports with your own sales records where you can.
- Small samples mislead. A CPA built on three conversions can swing sharply with one more. Wait for a reasonable amount of data before acting.
- Cost of the whole funnel counts. Time spent on follow-up and refunds belong in your thinking, even if they are absent from the ad report.
- Tracking rules differ. Conversion tracking relies on tags or tracking pixels, and consent rules vary by country. Check what applies to you; this is not legal advice.
Outcomes vary, and these formulas cannot predict results. Their use is to show, before and after spending, whether the sums can work at all.