CPI Calculator

Cost per install is the toll a user-acquisition team pays at the app-store door. Enter what you spent and how many installs it bought — then add a budget to see how many more that rate would deliver.

CPI (cost per install) = app ad spend ÷ installs. It is the headline price of a new app user, the metric every user-acquisition campaign reports first. But CPI is only an entry fee: a cheap install that never opens the app is worse than a pricier one that retains and pays. iOS installs typically cost more than Android, and incentivised traffic can slash CPI while delivering users who churn. Read CPI alongside retention and post-install value, never alone.

The calculator

CPI Calculator inputs and result

User-acquisition media for the period.
Installs attributed in the same period.
Optional — estimates installs delivered.
Cost per install (CPI)
$0.00
0installs
0installs per budget
Export

Walkthrough

How to use this calculator

  1. Pull spend and installs from the same windowTake both numbers from one date range and the same set of campaigns. A CPI that pairs this month’s spend with last month’s installs is meaningless.
  2. Use attributed installs, not store downloadsCount installs your measurement partner credits to paid media. Raw store downloads include organic and inflate your apparent efficiency.
  3. Read the cost per installThe headline is what one new install costs. Compare it across campaigns and platforms — but never forget iOS and Android live in different price worlds.
  4. Add a budget to project volumeEnter a planned spend and the tool estimates the installs it buys at this CPI — the input for a user-acquisition forecast.
  5. Export for the UA reviewCopy a share link, drop the CSV into your growth model, or print a one-page summary for the acquisition stand-up.

From the desk

RGM Expert Says

Real Growth Matters — Mobile growth practiceHow we use this tool with clients

When we audit a mobile user-acquisition program, CPI is the first number on the whiteboard and the first one we argue about. A team will celebrate a falling CPI without noticing the cohort behind it stopped opening the app. We treat CPI as the price of a ticket, not the value of the trip — the honest read only arrives once you trace those installs to day-1, day-7 and day-30 retention.

The classic trap is incentivised and low-quality traffic. Offer-wall and bot installs can drop a CPI to pennies and gut your retention at the same time. We watch CPI and retention as a pair: if CPI falls while retention falls faster, the channel is buying you a vanity metric. The goal is the lowest cost per retained, paying user, which often means accepting a higher CPI from a cleaner source.

Platform context decides everything. Apple Search Ads installs behave nothing like Google App campaign installs, and iOS post-ATT measurement makes raw CPI comparisons across platforms genuinely unsafe. We benchmark each source against itself over time and judge new spend on blended payback, not on whoever is reporting the cheapest install this week.

The math

How it works

CPI is the bluntest user-acquisition metric: divide what you spent on installs by the installs you got.

CPI = App ad spend ÷ Installs
Installs per budget = Planned budget ÷ CPI
  • App ad spend — paid UA media driving installs in the period.
  • Installs — attributed installs in the same period; use MMP data, not raw downloads.
  • Planned budget — optional future spend, converted to a forecast install count.

CPI says nothing about quality. Pair it with retention and post-install ROAS; see RGM’s cost per install deep dive.

Why it matters

Why a cheap install can be the expensive one

CPI is seductive because it is simple, and dangerous for the same reason. The number that actually decides whether user acquisition pays is not the cost of an install but the cost of a retained, monetising install. A $0.40 install that churns by day three is pure loss; a $1.50 install that stays for months and subscribes is a bargain. Optimising CPI in isolation quietly selects for the worst users you can buy.

The second issue is comparability. iOS installs cost more than Android, premium-network installs cost more than incentivised ones, and post-ATT measurement on iOS makes raw cross-platform CPI comparisons unreliable. Treat CPI as a within-channel trend line, not a leaderboard across channels with different economics.

Used properly, CPI is a planning input, not a trophy. Once you know your CPI and your downstream value — retention curves, in-app ARPU, install-to-paid rate — you can set an allowable CPI that still clears payback, then hold every source to it. That turns user acquisition from a race for cheap installs into a disciplined hunt for profitable ones.

Benchmarks

How to read cost per install

There is no single ‘good’ CPI — it swings by platform, geography, vertical and source quality. Use public ranges to orient, then judge against retention and post-install value.

SignalWhat to watchWhy it matters
CPI by platformiOS usually well above AndroidDifferent audiences and auction dynamics; never blend blindly
Incentivised trafficVery low CPI, poor retentionCheap installs that churn destroy payback
Geo and verticalTier-1 geos and gaming run higherBenchmark within the same segment, not across
The real KPICost per retained / paying userCPI is the entry fee, not the return
General patterns from mobile UA reporting (e.g. AppsFlyer cost benchmarks and Liftoff mobile reports); always verify against live, segment-specific data.

Voices worth trusting

What growth leaders say about installs

The vanity metric in mobile is the cheap install. The number that compounds is the retained user, and retention is a feature of the product long before it is a feature of the media plan.
Author, The Cold Start Problem (paraphrase)
Buy users you can keep. A lower cost per install that hides a worse cohort is not efficiency — it is a slower way to lose money.
Founder, Reforge (paraphrase)

Go deeper

Reading on growth and metrics

Related on RGM

Keep learning

FAQ

Common questions

How do you calculate CPI?
CPI = app ad spend ÷ installs in the same period. If you spent $5,000 and got 4,000 installs, your cost per install is $1.25. Use attributed installs from your measurement partner, not raw store downloads.
What is a good cost per install?
There is no universal figure — it depends on platform, geography, vertical and source quality. iOS runs higher than Android, gaming higher than utilities. Benchmark each source against its own trend, and judge it on the cost per retained user, not the raw install price.
Why is iOS CPI higher than Android?
iOS audiences tend to monetise better and the inventory is more contested, so the auction clears higher. Post-ATT measurement changes on iOS also make raw cross-platform CPI comparisons unreliable; treat them as separate worlds.
Should I just minimise CPI?
No. Minimising CPI alone selects for the cheapest, least valuable users — often incentivised or bot installs that churn immediately. Optimise for cost per retained, paying user, which sometimes means paying a higher CPI from a cleaner source.
What is the difference between CPI and CPA for apps?
CPI pays for the install itself; CPA (cost per action) pays for a deeper event such as a registration or first purchase. CPA is closer to value but harder to optimise early; many teams buy on CPI and grade on CPA or post-install ROAS.
How can I lower CPI without hurting quality?
Improve creative and store-listing conversion, tighten targeting, cut incentivised and low-retention sources, and lean on lookalikes built from your best cohorts. The aim is a lower cost per good install, not just a lower number.

Related tools

Related tools