CPT, CPA, CPI and ROAS: the Apple Search Ads metrics that matter

Apple Search Ads dashboards can feel like an acronym soup—CPT, CPI, CPA, ROAS. The trick is to treat them as different layers of the same pipeline: auction cost → tap → install → purchase/revenue. Once you read them in that order, you can decide what to fix (bids/keywords vs product page vs onboarding vs pricing).

Below is a practical mental model for each metric, what it depends on, and the fastest “next check” when it’s not behaving.

CPT (Cost per Tap): the auction’s immediate cost signal

CPT is how much you pay for each tap on your ad. In Apple Search Ads, you’re in a CPT auction: you set a max CPT bid, Apple matches your ad to eligible queries, and the final CPC is determined by the auction.

What CPT tells you

• Whether your ads are competing effectively for the keyword/query space you chose.
• Whether your bidding level is too low (high impressions, low taps) or too high (taps that cost more than your downstream value).

The key related metric: TTR

To interpret CPT, look at TTR (taps/impressions).

• High CPT + low TTR: users aren’t choosing your ad even when it shows. Likely mismatch between keyword intent and your ad/product page.
• High CPT + high TTR: users click, but the auction is expensive. Now your focus shifts to what happens after the click (install rate, conversion, revenue).

Next checks when CPT is “bad”

1. Confirm the placement scope: Your spend patterns can differ between Search Results and other placements (e.g., Today tab). Most indie budgets should start with Search Results, but still verify.
2. Split performance by match type on Search Results keywords:
   • Exact and Broad are manually configured.
   • Search Match (automatic discovery) runs in its own ad group and can broaden your audience.
3. Review max CPT bid vs realized performance: If realized CPT is consistently near your max, you may be winning expensive queries. If you’re far below max but still paying more than you can afford later, you probably have a downstream value problem—not necessarily a bidding problem.

CPI (Cost per Install): the tap-to-install efficiency test

CPI is spend ÷ installs. It answers: How expensive is it to get an install from the taps you’re buying?

What CPI is really measuring

CPI combines:

• Your CPT (auction cost per tap)
• Your install conversion rate (installs/taps)

So CPI can “look good” even with a high CPT if your product page and user intent drive excellent installs.

Next checks when CPI is too high

1. Check conversion rate: installs/taps
   • Low conversion with decent CPT: users click but don’t install. This often points to store listing mismatch (the ad promise vs the product page experience).
2. Validate your product page / custom product pages
   • Apple Search Ads doesn’t give you creative auction advantage, so the product page is one of your biggest levers.
   • Make sure the first screen answers: “What is it?” and “Why should I install now?”
3. Review keyword intent
   • Broad and discovery can inflate impressions and taps from users who don’t convert.
   • A concrete approach: if you’re using Broad heavily, compare CPI across Broad vs Exact. If Broad CPI is worse, cap or narrow the query space.

CPA (Cost per Acquisition): installs → purchase/subscription events

CPA is spend ÷ acquisition events, where “acquisition” is your chosen conversion event in the reporting setup (e.g., first purchase, subscription start, or another purchase-defined event).

The important nuance: CPA isn’t “purchase per impression”

CPA sits after the install. It depends on:

• Install quality (from CPI + conversion rate)
• Onboarding and monetization (purchase/subscription propensity)
• How well your app actually delivers what users expected from the ad query and product page

Timing/attribution reality

Apple’s attribution uses the AdServices attribution token, resolved within roughly ~24 hours. Revenue mapping is typically done by tools like RevenueCat (install → purchase chain). Practically, it means:

• Don’t overreact to single-day spikes.
• Use a window that’s long enough for attribution + purchase events to stabilize (for many teams: at least several days, longer for slow purchase behavior).

Next checks when CPA is high

1. Separate “bad installs” from “bad monetization”
   • If CPI is already high, CPA will be high too—start upstream.
   • If CPI is fine but CPA is high, focus on user value after install: onboarding flow, paywall timing, offer alignment.
2. Look for query mismatch
   • Search Match (discovery) can be productive, but it can also pull in users whose intent doesn’t match your strongest monetization segment.
   • If you see CPA regressions after enabling/expanding Search Match, narrow your app store page messaging or tighten your keyword strategy.
3. Confirm the conversion event you’re optimizing for
   • CPA only means something if the “acquisition” event corresponds to revenue reality for your business model.

ROAS (Return on Ad Spend): the revenue-based scoreboard

ROAS is revenue ÷ spend. It’s the only metric that answers: Did this spend make money?

Why ROAS can mislead if you look at it alone

ROAS is downstream of multiple things:

• CPT (how expensive taps are)
• CPI (tap-to-install efficiency)
• CPA (install-to-purchase efficiency)
• Purchase value per acquired user (average revenue from attributed users)

A high ROAS could happen with a small budget and lucky conversion rates. A low ROAS could be temporary attribution lag or a shift in query mix.

Next checks when ROAS is low

1. Decompose ROAS back into the pipeline
   • If ROAS drops: is it because spend rose (CPT/bids), because installs fell (store conversion), or because purchase rate/revenue per user fell (onboarding/offers)?
2. Verify that attributed revenue is flowing correctly
   • If RevenueCat (or your mapping) isn’t resolving the attribution token to the right revenue events, ROAS will be wrong even if the campaigns are fine.
3. Match ROAS to business reality
   • If you have subscription churn or longer-term revenue effects, your immediate “ROAS” may undercount what the campaign produces over time. In that case, track a second view for longer-term value if you can.

How to choose what to optimize (a simple decision tree)

When something is off, don’t guess—use the metric ladder.

1) CPT high (and TTR weak)

Likely issue: your ad is showing but not earning clicks for those impressions.

• Check keyword intent and placement mix.
• Consider narrowing to exact queries.

2) CPI high

Likely issue: clicks aren’t converting to installs.

• Review product page messaging and any custom product page targeting.
• Tighten keyword match types (Broad/Search Match can widen intent).

3) CPA high (but CPI is fine)

Likely issue: installs aren’t converting to purchases at the rate you need.

• Improve onboarding and paywall/offer alignment.
• Validate conversion event setup and attribution quality.

4) ROAS low

Likely issue: either the pipeline is inefficient or revenue mapping/attribution isn’t reflecting reality.

• Break down ROAS into CPT/CPI/CPA drivers.
• Confirm attribution token → revenue mapping.

Common mistakes indie teams make with these metrics

• Comparing CPT and CPI across different match types without segmenting. Broad/discovery can change intent quality.
• Ignoring TTR. CPT without TTR is like looking at speed without knowing distance.
• Overreacting to short windows. Attribution and user purchase timing can lag.
• Treating ROAS as “the only number.” ROAS tells you what, but CPT/CPI/CPA tell you why.

Practical workflow for your next Apple Search Ads review

If you want something you can do today:

1. Pick a time window (e.g., last 7 days) and segment by:
   • campaign/region
   • ad group
   • keyword match type (Exact, Broad, Search Match)
2. For each segment, record:
   • CPT
   • TTR
   • installs/taps (if you have it)
   • CPI
   • CPA (for your purchase/subscription event)
   • ROAS
3. Use the ladder: start with CPT/TTR, then CPI, then CPA, then ROAS.
4. Apply changes only to the lever that corresponds to the failing layer:
   • bidding/keywords for CPT
   • store conversion for CPI
   • onboarding/monetization and event correctness for CPA
   • confirm attribution and overall mix for ROAS

If you want this done faster, tools like AdsBuddy can read your Apple Search Ads performance plus your revenue reporting and suggest a short, prioritized set of changes to approve—one layer at a time—so you’re not guessing which lever to touch first.

Takeaway

Think of CPT, CPI, CPA, and ROAS as a chain: CPT (auction cost) → CPI (install efficiency) → CPA (purchase efficiency) → ROAS (revenue outcome).

When you align the metric you’re looking at with the pipeline stage you’re trying to improve, Apple Search Ads stops being confusing acronyms and becomes a straightforward optimization loop.