Growth Marketing Glossary

Cost Per Install (CPI)

C·P·Inoun

The price of a download — the app marketer's steering metric, and a trap for anyone who forgets installs are not users.

install÷spend=CPIwhat each app install costs
Schematic — spend divided by installs
Term
Cost Per Install (CPI)
Formula
Spend ÷ installs generated
Domain
Mobile app user acquisition
Trap
Cheap installs that never open the app

Forms & parts of speech

CPI · noun
Spend per app install.
"Rewarded inventory halved our CPI - and day-7 retention told us we'd bought downloads, not users."

Definition in plain terms

Cost per install (CPI) is your paid-media spend divided by the app installs it produced. Spend $10,000 and drive 4,000 installs, and your CPI is $2.50. It is the headline efficiency metric of mobile USER ACQUISITION — the number app marketers quote, benchmark, and bid against — and it is also where mobile marketing's most reliable self-deception lives, because an install is a download event, not a customer.

The mechanics

CPI is shaped by platform (iOS installs typically price above Android), geography, category (games, finance, and dating all run different markets), creative quality, and crucially the inventory type. Incentivized or rewarded placements — where users install to earn in-game currency — produce dramatically cheaper installs with dramatically worse intent, which is how a falling CPI can mean a deteriorating business. Measurement runs through MOBILE MEASUREMENT PARTNERS and the app stores' own attribution, with iOS privacy rules (App Tracking Transparency, SKAdNetwork) blurring install-source precision since 2021. The mature pattern is to treat CPI as the top of a chain, not the goal: track cost per activated user, cost per retained user (day 7, day 30), and cost per paying user beside it, and let RETENTION and revenue per install decide which sources deserve budget. The classic failure is optimizing the auction toward whoever installs cheapest — bots, reward-chasers, accidental taps — while the cohorts that monetize quietly concentrate in the 'expensive' channels. App-install fraud (device farms, click injection claiming organic installs) makes the cheap end cheaper and faker still, which is why fraud filtering and cohort-quality reporting are standard kit.

When it matters

CPI matters to anyone buying app growth — games, fintech, subscription apps — because install pricing decides how far an acquisition budget stretches. But it earns its place only when read against downstream quality. The discipline is to set an allowable CPI per channel from what its installs actually return — retention-adjusted, revenue-adjusted — rather than chasing the lowest sticker price, to separate incentivized from intent-driven inventory in every report, and to validate sources against fraud. The app marketers who win treat CPI as the price of an audition, with day-7 retention as the callback decision.

Worked example. A puzzle-game studio celebrates cutting CPI from $3.20 to $1.40 by shifting budget into rewarded-video networks. Thirty days later the finance view tells the real story — day-7 retention on the cheap cohorts is 4% against 19% from the old channels, and revenue per install collapsed faster than CPI fell, so the cost per retained player nearly doubled. The studio rebuilds reporting around the chain: CPI, cost per day-7-retained user, and 90-day revenue per install by source, with incentivized inventory broken out and an MMP fraud filter on. Budget flows back toward intent-driven placements whose $3 installs become $0.92 retained players, and rewarded inventory survives only for a launch-week chart push, priced as what it is. The dashboard metric had improved while the business got worse — the chain caught it.
Failure modes to watch. Chasing the lowest CPI into incentivized and fraudulent inventory; reporting installs without activation, retention, and revenue beside them; ignoring iOS measurement blur and crediting sources SKAdNetwork cannot verify; benchmarking CPI across categories and platforms as if one market existed; and buying downloads when the business needs users.

Synonyms & antonyms

Synonyms

cost per installCPIinstall cost

Antonyms

organic installcost per acquisition

Origin & history

CPI emerged with the app-store economy after 2008, as mobile games needed a unit price for growth and ad networks productized install campaigns. The metric's vocabulary — installs, cohorts, day-7 retention, rewarded inventory — consolidated through the 2010s mobile-gaming boom, and Apple's 2021 privacy changes reshaped how installs are attributed without dethroning CPI as the channel's steering number.

Etymology: source.

Usage trends

Search interest for this term over the last five years:

View interest-over-time on Google Trends →

Common questions

What is cost per install (CPI)?
Paid-media spend divided by the app installs it generated — spend $10,000 for 4,000 installs and your CPI is $2.50. It prices a download, not a customer.
What is a good CPI?
One that produces retained, monetizing users at sustainable cost — judge each source by cost per retained user and revenue per install, since cheap incentivized installs often retain near zero.
Why did iOS make CPI harder to measure?
App Tracking Transparency and SKAdNetwork limit install-level attribution, blurring which source drove which install and pushing measurement toward aggregated and modeled reporting.

Related tools & calculators

Resources & people to follow

Curated, non-competitor resources verified per term.

Related training

Disciplines

Areas of marketing where cost per install (cpi) is a core concern:

Sources

  1. trendsGoogle Trends — "cost per install"