How is CPA different from CPI?

CPI measures the cost of one install. CPA measures the cost of one acquired customer who performed a monetization event (purchase, subscription, signup). CPA is always higher than CPI because only a fraction of installs convert to the target event.

What event qualifies as an acquisition?

Whatever your business defines as one. Common choices: first-purchase event (e-commerce), trial-start (subscription apps), signup with email verification (B2B/utility), level 5 completion (mobile games). The event should correlate strongly with LTV.

How do I measure CPA when SKAdNetwork blocks user-level data?

Use SKAN conversion values to encode the target event. SKAN 4.0 has three postback windows (0-2, 3-7, 8-35 days) and fine/coarse values; map your CPA event into one of those values. Aggregate-level CPA is then computed from SKAN postbacks per campaign.

What is CPA in Apple Search Ads? Cost Per Acquisition Explained

What is CPA?

Cost Per Acquisition (CPA) is the cost to acquire one converting customer — defined by the advertiser, not the platform. CPA captures only the installs that progress to a meaningful business outcome.

CPA is the most important efficiency metric for subscription apps, e-commerce apps, and any app where install-to-monetization conversion is below 50%. For free apps monetized by ads, CPI may be closer to the right metric.

How to compute CPA

The formula:

CPA = Total Spend ÷ Acquired Customers
    = CPI ÷ (Install → Acquisition Conversion Rate)

If CPI is $5 and 25% of installs become paying users within 30 days, CPA = $5 ÷ 0.25 = $20.

The “acquired customer” definition is the advertiser’s choice:

E-commerce: first purchase
Subscription: trial start (less strict) or paid conversion (stricter)
Free with IAP: any purchase event
B2B / utility: verified signup, profile completion
Mobile games: level X completion, first IAP, day 7 retention

LTV-to-CPA ratios

The conventional ratios:

Ratio	What it means
LTV ≥ 5× CPA	Aggressive growth opportunity — scale spend
LTV ≥ 3× CPA	Healthy efficiency — sustainable growth
LTV ≥ 2× CPA	Breakeven plus operating cost
LTV ≤ CPA	Unsustainable — fix before scaling

The right ratio depends on payback window tolerance and capital cost. Venture-backed apps typically tolerate longer payback (12-18 months); bootstrapped apps need faster (3-6 months).

CPA in Apple Search Ads specifically

Apple Search Ads bills CPT, not CPA. Apple does not offer a built-in target-CPA bidding mode the way Google does. CPA is computed post-hoc by:

Pulling tap and install data from the Ad Services API (deterministic, near-real-time).
Joining with downstream conversion data from your own backend (deterministic) or from SKAN conversion values (aggregate, privacy-preserving).
Aggregating spend ÷ acquisitions per campaign / ad group / keyword.

For most operators, the simplest setup:

Use own backend events for deterministic CPA in your ASA-only dashboard.
Use SKAN conversion values for cross-network attribution via your MMP.

CPA traps

Over-attributing to the last touch. If a user installs via ASA and converts only after seeing 4 retargeting ads on other channels, ASA gets credit for the acquisition but the full cost picture is higher.
Short-window CPA bias. A 7-day CPA looks great for subscription apps where most users trial and churn; 30-day or post-trial-conversion CPA tells a different story.
Selection bias from Brand vs Discovery. Brand campaigns show low CPA because Brand intent already had high purchase probability before the ad — they would have likely converted organically.

How ASAPilot helps

ASAPilot’s analytics layer computes CPA per campaign / ad group / keyword by joining ASA spend with a conversion event you define (via app backend or SKAN). Findings appear alongside CPI in the dashboard.

Read the audit guide for the workflow that surfaces high-CPA keywords, or pricing for plan tiers.

CPI — the install-cost upstream metric
ROAS — the return-on-spend efficiency metric
SKAdNetwork — Apple’s privacy-preserving attribution