Cost per acquisition:
how CPA differs from CAC, with a break-even test

CPA: what you paid for each conversion

Cost per acquisition (also called cost per action or cost per conversion) tells you how much a campaign spent to produce one defined result. That result might be a purchase, a sign-up or an app install, so the first step is to decide, and keep fixed, which event counts.

Always use the same attribution window and time period for both parts. A campaign that spent money in September but is credited with October sales will look better or worse than it truly is.

CPA versus CAC

Customer acquisition cost (CAC) looks at the whole business and counts only new customers. It includes everything it takes to win them: ad spend, but also sales salaries, agency fees, tools and creative production. CPA is usually narrower and platform-level.

The three figures describe the same campaign. None is wrong, but they answer different questions, and mixing them in one report is how teams talk past each other.

The blended number hides the channels

A single CPA for the whole account averages away large differences. Here is the ₹1,20,000 and 150 conversions split by channel.

Against the ₹840 break-even from the margin test below, search earns about ₹284 per order while social and display lose money on a first purchase. That does not mean cutting them automatically, since they may be starting journeys that search then closes. It does mean you should ask the question channel by channel, and test before shifting budget.

Is your CPA affordable? The break-even test

A CPA means nothing without margin. The ceiling you can pay on first purchase is the gross profit that purchase produces, before the marketing cost.

A ₹40 cushion is thin. Returns, payment fees or a modest rise in CPA can erase it, which is the reason to track the break-even figure beside the CPA at all times.

Repeat purchases change the answer

If customers buy again, the first-order break-even is too strict. Compare CAC with the lifetime gross profit of a customer, usually shortened to LTV. A CAC of ₹1,222 is acceptable if a typical customer generates ₹3,000 of gross profit over their lifetime, but dangerous if most never come back.

• Use realistic repeat rates from your own cohorts, not an optimistic average.
• Watch how long payback takes, as cash is tied up in the meantime.
• Exclude one-off promotional spikes when estimating typical LTV.

Reading the number and avoiding mistakes

• Do not compare CPA across different goals: a lead and a purchase are not the same thing.
• Check conversion quality. A cheap CPA made of low-value or fraudulent sign-ups is worse than a dearer one.
• Mind the volume. A CPA calculated from 8 conversions swings wildly; wait for a bigger sample.
• Look at marginal CPA. The next rupee often buys conversions at a higher cost than the campaign average.

What counts as a conversion

Definition changes move the number as much as performance does. Counting sign-ups instead of paid orders makes CPA fall without any business benefit, and counting an order that is later returned flatters the figure.

Adjust for reality. If 10 percent of the 150 orders are returned or cancelled, net conversions are 135 and the effective CPA is 1,20,000 ÷ 135 = ₹889, not ₹800. Report the net figure next to the gross one.

• Agree on one primary conversion event per campaign and write it down.
• Keep micro-conversions such as add-to-cart as diagnostics, not as the target.
• Use cost per lead for early funnel goals and reserve CPA for paid outcomes.

Linking CPA to clicks and conversion rate

CPA is the cost per click divided by the share of clicks that convert. That gives a handy way to see what is driving a poor figure, and what a bid ceiling should be.

Run it backwards for bidding: a maximum affordable cost per click equals target CPA multiplied by conversion rate. A target of ₹700 at 2.5 percent supports a ceiling of ₹17.50 per click.

It also tells you where to work. If CPA is too high, either clicks cost too much or too few visitors convert. Improving a landing page from 2.5 to 3 percent conversion lowers CPA from ₹800 to about ₹667 with no change in bids, since 20 ÷ 0.03 = 666.7. Compare that with the effort of winning cheaper clicks and decide which lever is cheaper to pull.

Finally, put the numbers in a rhythm. Review CPA weekly at campaign level for quick corrections, and CAC and payback monthly at business level, where noise averages out and the decisions are about budgets and channels. Keep a short log of changes to targeting, creative and landing pages beside the figures, because without it a movement in CPA a month later is almost impossible to explain, and the same experiment ends up being repeated.

Keep one more distinction in view when you report upwards: efficiency is not the same as growth. A campaign with an excellent CPA may be reaching only the few people who were already close to buying, and pushing its budget up can make the cost per customer climb quickly. Report CPA together with the volume of conversions so that the reader sees both how cheap each customer was and how many customers there were.