Skip to content

CPA (cost per acquisition)

In Apple Ads, CPA means cost per download. It is the price of getting your app onto a device, not the price of a user who installs, starts a trial or pays.

Updated

Many ad platforms use CPA for whatever conversion you choose to optimize. Apple Ads is narrower: its acquisition is a download. That makes Apple Ads CPA a useful mid-funnel cost, and a risky number to treat as the goal for a subscription app, where the outcome that pays the bills comes several steps later.

How CPA is calculated

Adapty Ads Manager reports two versions of average CPA, one for each download definition.

  • Avg CPA (Total): total campaign spend divided by total downloads resulting from a view or a tap on your ad within the reporting period.
  • Avg CPA (Tap-Through): total campaign spend divided by the number of tap-through downloads within the reporting period.

Total CPA is always equal to or lower than Tap-Through CPA, because the same spend is divided by a larger number. Use one consistently when comparing keywords or periods.

CPA can also be read as CPT divided by download rate. A $1.50 tap that becomes a download 60% of the time costs $2.50 per download, so CPA rises when taps get more expensive or when fewer taps convert.

Per download, not per install or per paying user

An Apple Ads download is not the same event as an install, and neither is a subscriber. Adapty Ads Manager therefore reports separate cost metrics for each step:

Cost metrics at each funnel step in Adapty Ads Manager
MetricFormulaMeasures
Avg CPASpend / DownloadsCost of a download attributed by Apple
CPISpend / InstallsCost of an install recorded by the app
Cost per TrialSpend / TrialsCost of a new trial start
Cost per SubscriptionSpend / SubscriptionsCost of a new paid subscription without a trial

When someone quotes a “CPI” for Apple Ads, check which denominator they mean before comparing it with your CPA.

CPA benchmarks

Adapty’s Apple Ads for subscription apps 2026 report measured CPA as spend divided by downloads across 8,000+ apps and 1,000,000+ ad groups in 2025 campaigns. Country and category figures are separate aggregates; the report does not publish a category-by-country matrix.

Average CPA (spend per download) in Apple Ads, 2025 campaigns
ScopeAverage CPA
United States, all categories$2.51
Business apps, all countries$3.42
Health & Fitness apps, all countries$3.14
Education apps, all countries$2.09
Productivity apps, all countries$1.48
Utilities apps, all countries$0.69

Source: Adapty, Apple Ads for subscription apps 2026. Per-row app counts are not published. A benchmark helps you spot an outlier; it does not tell you what CPA you can afford.

What CPA you can afford

Your allowable CPA comes from your own funnel, not from a benchmark. Work backwards: the revenue a paying user brings in over your payback window, times the share of downloads that become paying users, is the most you can spend per download and break even. Two apps with the same CPA can have very different outcomes if one converts downloads to trials and subscribers far better than the other.

Common questions

Is Apple Ads CPA the same as CPI?

Not exactly. Apple Ads CPA is spend per download attributed by Apple. CPI divides spend by installs recorded by the app. The two are close but differ because downloads and installs are different events.

Should I use Total or Tap-Through CPA?

Optimize on Tap-Through CPA, which credits only downloads that followed a tap. Total CPA also counts view-through downloads, where the user saw the ad but never tapped; Apple does not fully explain how those are attributed, so read them separately rather than mixing them into keyword decisions.

What is a good CPA for Apple Ads?

One your revenue per user can pay back. Benchmarks vary widely by country and category, so the useful test is cost per trial and cost per subscription against what those users earn.

Look past cost per download to cost per subscriber

Adapty Ads Manager puts CPA next to installs, trials, subscriptions, revenue and cohort ROAS for every keyword.