# What is CPA (cost per acquisition)?

> CPA is what you pay in ads for one customer action, such as a purchase or a sign-up. Formula, a worked example and how to set a target CPA.
- **Author**: Marxx
- **Published**: 2026-10-23
- **Category**: Metrics
- **URL**: https://marxx.ai/library-glossary/cpa

---

CPA (cost per acquisition) is the ad spend it takes to get one action you care about, such as a purchase, a sign-up or an app install. If INR 50,000 of ads brings 100 purchases, your CPA is INR 500.

## The formula

CPA = ad spend / number of conversions

Meta calls this "cost per result". The result is whatever your campaign optimises for.

## A worked example

Say a course seller spends INR 75,000 in a week. Ads Manager reports 150 purchases of a INR 2,499 course.

- CPA = 75,000 / 150 = **INR 500**
- Revenue = 150 x 2,499 = INR 3,74,850
- ROAS = 3,74,850 / 75,000 = 5.0

## How to set a target CPA

Start from what one customer is worth to you.

1. Take your profit per order. Say a INR 2,499 course leaves INR 2,000 after costs.
2. Decide how much of it you are happy to spend on ads. Say half.
3. Target CPA = INR 1,000.

Any campaign under INR 1,000 makes money. Over INR 1,000, it needs fixing or cutting.

## Why CPA rises

- Ad fatigue: the same people have seen your ad too often.
- A landing page that is slow or does not match the ad.
- A higher CPM in a busy season.
- A change that reset the learning phase.

## FAQ

### What is the difference between CPA and CAC?

CPA is ad spend per action on one platform. [CAC](/glossary/cac) is all the cost of winning a new customer, across every channel, including salaries and tools.

### Is a lower CPA always better?

Not if the customers are worth less. A INR 300 CPA for buyers who never return can be worse than a INR 600 CPA for buyers who order every month.

### Where do I see CPA in Meta?

It is the "Cost per result" column. Add "Cost per purchase" if your campaigns optimise for something else.

---
- [More Metrics articles](https://marxx.ai/library-glossary/category/metrics)
- [All articles](https://marxx.ai/library-glossary)