AdPilot MCP
Sign in
Browse documentation
Last reviewed

Apple Ads metrics and formulas

Apple Ads metrics describe different stages of the acquisition funnel. A reliable review names the denominator, time range, attribution scope, and revenue definition before interpreting a result. CPI explains install efficiency; ROAS asks whether observed proceeds are covering ad spend. Neither metric is meaningful without context.

Delivery and acquisition metrics

Metric Formula What it answers
Impressions Count of ad appearances How much eligible visibility did the ad receive?
Taps Count of ad taps How many people opened the destination?
Tap-through rate (TTR) taps / impressions How often did an impression earn a tap?
Installs Attributed downloads or redownloads, according to report scope How many conversions did Apple attribute?
Tap-through conversion rate tap-through installs / taps How often did a tap lead to an install?
Spend Sum of Apple Ads cost in the report currency What did the selected traffic cost?
Average CPT spend / taps What was the average cost of a tap?
Install CPA or practical CPI spend / installs What was the average advertising cost per attributed install?

Apple uses CPI specifically for Apple Ads Basic pricing, while Apple Ads Advanced reporting commonly uses average cost-per-acquisition for installs. Growth teams often say CPI for spend / installs; state the formula so there is no ambiguity.

TTR and conversion rate may be displayed as percentages. A stored decimal of 0.08 equals 8%.

Business outcome metrics

Apple Ads ends at attributed acquisition. Your business funnel continues:

Metric Formula Important definition
Activation rate activated users / installs Define the meaningful product event
Trial start rate trial starters / eligible installs State eligibility and measurement window
Trial-to-paid rate new payers / trial starters Account for trials that have not finished
Payer CAC ad spend / acquired payers Define payer attribution and cohort
ARPU revenue or proceeds / users State which users and revenue basis
ARPPU revenue or proceeds / paying users Excludes non-paying users from denominator
Observed ROAS observed revenue or proceeds / ad spend State source, window, and attribution scope
LTV Cumulative value per acquired user or payer State cohort, horizon, and forecast method
Payback period Time until cumulative value covers acquisition cost State whether value means revenue or proceeds

CAC is not CPI. CPI uses installs as the denominator; payer CAC uses paying customers. If 100 installs produce 10 payers, payer CAC will be ten times CPI when all spend and users share the same scope.

Read ROAS as a ratio or percentage

ROAS is calculated as:

ROAS = attributed or observed value / ad spend

The same result can be expressed two ways:

  • 1.5x ROAS means value is 1.5 times spend.
  • 150% ROAS means the same thing as 1.5x.
  • 1.0x equals 100% and indicates value equals ad spend before other business costs.
  • 0.6x equals 60% and indicates 60 cents of observed value per dollar spent.

ROAS is not profit margin. Even 1.0x proceeds ROAS may be insufficient after operating costs, refunds not yet reflected, support, infrastructure, and the required return on working capital.

AdPilot reports observed ROAS as a ratio. Multiply by 100 to display the equivalent percentage.

Choose gross revenue or proceeds deliberately

Revenue definitions change the numerator:

  • Gross revenue usually starts from what the customer paid, subject to the reporting source’s exact definition.
  • Revenue net of taxes subtracts estimated taxes.
  • Proceeds subtract estimated taxes and store commissions or fees.

AdPilot’s profitability service uses synchronized RevenueCat proceeds because proceeds are closer to what the developer retains. RevenueCat values remain estimates and may change as receipts, refunds, taxes, commissions, or exchange rates are updated.

Never compare spend in one currency with proceeds in another without conversion. Record the conversion source and date when conversion is necessary.

Apply denominator and zero rules consistently

Rates and unit costs are undefined when their denominator is zero:

  • If impressions are zero, TTR is not available, not 0%.
  • If taps are zero, average CPT and tap-through conversion rate are not available.
  • If installs are zero, CPI or install CPA is not available; report spend and zero installs separately.
  • If spend is zero, ROAS is not available, even when revenue exists.
  • If revenue data is unavailable, ROAS is unknown, not 0x.

Using zero for an undefined ratio hides the difference between “measured and none occurred” and “cannot be calculated.” AdPilot returns null for unavailable derived metrics where applicable and supplies data-quality context.

Worked example: acquisition and profitability

For 30 complete days, an app records:

  • 100,000 impressions
  • 5,000 taps
  • 1,000 installs
  • $4,000 Apple Ads spend
  • 80 acquired payers under the team’s attribution model
  • $6,000 in synchronized RevenueCat proceeds for the app

Calculations:

  • TTR = 5,000 / 100,000 = 5%
  • Conversion rate = 1,000 / 5,000 = 20%
  • Average CPT = $4,000 / 5,000 = $0.80
  • CPI = $4,000 / 1,000 = $4.00
  • Payer CAC = $4,000 / 80 = $50.00
  • Observed app-level ROAS = $6,000 / $4,000 = 1.5x, or 150%

The first five calculations can share Apple Ads and an agreed payer attribution model. The final calculation is only campaign-level ROAS if the $6,000 numerator is truly attributed to those campaigns. In AdPilot today, RevenueCat proceeds are app-level and blended, so the 1.5x result is observed app-level ROAS across mapped Apple spend, not proof that each campaign generated a particular share of proceeds.

Understand attribution windows and cohort maturity

Apple Ads reporting can credit installs after an ad interaction within Apple’s applicable attribution rules. Revenue and subscriptions happen later. A seven-day acquisition cohort with annual trials cannot have mature payer or LTV evidence yet.

When comparing periods:

  • use complete days in the workspace timezone;
  • compare cohorts at the same age;
  • allow time for trials, renewals, and refunds;
  • avoid mixing acquisition-date and transaction-date analysis without labeling it;
  • note view-through versus tap-through scope;
  • check whether mappings and synchronizations cover the whole period.

Early ROAS is an observation, not a final lifetime result. Projected LTV should be labeled as a forecast and kept separate from observed proceeds.

How AdPilot calculates and qualifies metrics

AdPilot synchronizes Apple impressions, taps, installs, and spend for complete days. It can calculate CPI, TTR, and conversion rate for evaluations and return performance totals through MCP tools.

When RevenueCat is connected and mapped, AdPilot divides app-level proceeds by mapped Apple spend for the same period. The response explicitly warns that proceeds are blended across mapped campaigns and are not deterministic campaign-level attribution. Freshness, missing mappings, incomplete RevenueCat coverage, and cohort maturity appear in the data-quality envelope.

Try: Calculate TTR, conversion rate, CPT, CPI, and observed ROAS for the last 30 complete days. Show every formula, use null for undefined ratios, and explain attribution and cohort caveats.

Review checklist

  • Use one date range and workspace timezone.
  • Confirm spend and value use compatible currencies.
  • Name the install, payer, revenue, and attribution definitions.
  • Keep observed and projected value separate.
  • Treat zero-denominator results as unavailable.
  • Compare cohorts at equivalent maturity.
  • Review absolute volume alongside rates.
  • Do not turn app-level proceeds into campaign-level claims.

Next steps

Official sources

Need help with AdPilot MCP?

Contact support at support@kovacssoftware.com

AdPilot is an independent service and is not affiliated with, endorsed by, sponsored by, or operated by Apple Inc. Apple and Apple Ads are trademarks of Apple Inc.