Cost per install benchmarks for 2026, and why CPI shouldn't be your main metric

Recent published CPI figures by vertical, region and country, why they vary so much, why CPI misleads, and how to set a target from your own unit economics.

Every founder planning a paid launch asks the same question: what should an install cost? Published cost per install benchmarks give a useful starting range, but they vary by a factor of thirty or more between countries and categories, and a low CPI can hide a campaign that loses money. This guide pulls together recent published figures for the US and India, explains what moves CPI, and sets out what to track instead.

Every figure below comes from a named, linked report. Benchmarks are medians or averages across many apps; your own numbers will differ.

Recent cost per install benchmarks

Mobile measurement company Adjust publishes CPI data across its client base. Its Mobile app trends 2026 report, based on data from January 2024 to January 2026, reported these global CPIs for 2025:

The same report shows how much region matters within a vertical:

For a country-level view, Adjust's shopping app insights report for 2026, as reported by PPC Land, put the global median shopping app CPI at $1.08 in the first half of 2026. The United States was $3.28 (up 22% on 2025), the UK and Ireland $3.85 (up 80%, the highest in the dataset), North America overall $3.12 and India $0.10 (down 57%).

Two things stand out. A shopping install in the US costs over thirty times one in India. And finance installs in North America cost several times the global finance figure. A single "average CPI" is close to meaningless for planning.

How to read any CPI benchmark

Before you put a published figure into a plan or a pitch deck, check five things:

• Whose data it is. An MMP's benchmark reflects its own client base, which may skew toward certain categories, company sizes or regions.
• Median or average. A median is less distorted by a few very expensive campaigns. Mixing medians from one report with averages from another gives misleading comparisons.
• The period. CPIs move quickly. Adjust's data shows UK and Ireland shopping CPI up 80% in a single comparison period, so a two-year-old figure can be badly out of date.
• The vertical definition. "Finance" can mean banking, payments, trading, crypto or lending, each with different economics.
• What counts as an install. Paid-only, or paid plus organic; all platforms, or one. Blended figures look lower than pure paid ones.

Use benchmarks to sanity-check a plan, not to set targets. After a few weeks of your own spend, your data is the only benchmark that matters.

Why CPI varies so much

• Country. CPI tracks how much advertisers can earn per user in that market. US and UK users tend to spend more, so more advertisers compete for them and auction prices rise.
• Platform. iOS users are often more expensive to acquire than Android users in the same country, but the gap depends on category and market. In India, Android dominates: AppsFlyer's State of App Marketing in India 2025 report puts Android at 94% of installs in 2025.
• Category. Finance, subscription and other high-value categories pay more because each user is worth more. Casual games buy huge volumes of cheap installs and monetize through ads.
• Optimization goal. A campaign optimizing for installs will find cheap installers. The same campaign optimizing for a purchase or subscription will show a higher CPI and, usually, better users.
• Creative and store page. Better creatives and higher store conversion lower your CPI without changing your bids.
• Season and competition. CPIs often rise when big advertisers push spend, for example around major shopping periods.

Why CPI is the wrong north star

CPI measures what it costs to get a download. It says nothing about whether that person signs up, pays or comes back. Optimizing toward it creates predictable problems:

• Cheap installs, weak users. Algorithms told to minimize CPI find the people most likely to install, not the people most likely to buy.
• Fraud looks like success. Fake installs are cheap by design. AppsFlyer's India report found install fraud averaged 12% across 2023–2025, and finance apps saw fraud rates above 30%. Campaigns judged on CPI can reward exactly the sources producing it.
• iOS measurement gaps. With privacy-preserving attribution on iOS, install-level data is limited, so CPI comparisons across channels are less reliable than they look.
• It ignores the funnel. An install at $2 with a 20% signup rate is cheaper per signup than one at $1 with a 5% rate.

What to track instead

CPI still has a place: as an input to these metrics and as an early warning when creative or auction conditions change. Our guide to app install campaign optimization covers how to move campaigns from installs to in-app events.

Turn your economics into a target CPI

Rather than copying a benchmark, work backward from what a user is worth to you. A simple version, using made-up numbers for illustration:

1. Value per paying user: say a paying user brings $60 in gross margin within your payback window.
2. Target CPA: if you want to recover acquisition cost within that window with room to spare, you might cap cost per paying user at $40.
3. Install-to-paying rate: if 5% of installs become paying users, your break-even CPI at that CPA is $40 × 0.05 = $2.00.
4. Compare to the market: if published benchmarks for your category and country are well above that figure, you need better conversion, higher value per user or a different market before scaling paid.

This calculation also shows where to focus. Doubling the install-to-paying rate doubles the CPI you can afford, which is usually easier than halving what the auction charges. For budget sizing, see how much to spend on app marketing.

Planning notes for the US and India

United States

• Expect higher CPIs than global averages, especially on iOS and in finance and subscription categories.
• Plan measurement around SKAdNetwork and AdAttributionKit for iOS, and lean on modeled and cohort data rather than install-level attribution. See our iOS attribution guide.
• Test budgets need to be large enough for campaigns to exit learning on a meaningful event, not just installs.
• Watch store conversion closely. At US prices, a small lift in the share of store visitors who install lowers effective CPI more cheaply than most bid changes.

India

• Installs are cheap, but quality varies widely. Judge channels on cost per KYC, first order or subscription.
• Android dominates volume, so plan creatives, testing and device coverage around it. The market is also shifting: AppsFlyer found India UA spend fell 15% year on year across Q1–Q3 2025, driven entirely by Android, while iOS UA spend rose 17%. A small iOS test can be worthwhile for apps targeting higher-income urban users.
• Run fraud protection from day one, particularly in finance. Our India app marketing guide covers vernacular creatives and tier-2 and tier-3 targeting.

How Appryze can help

Appryze plans and runs paid acquisition for apps in the USA, UK, UAE and India on Google App Campaigns, Meta and Apple Search Ads. We set targets from your unit economics rather than benchmarks, optimize toward the events that make money, and report cost per key event, retention and payback every week. See our Growth services.

Frequently asked questions

What is a good cost per install?

There is no universal good CPI. It depends on country, platform, category and what your users are worth. Adjust reported a global 2025 CPI of $0.98 for ecommerce and $1.13 for finance, but US costs run several times higher. A good CPI is one that keeps your cost per paying user below what those users earn you.

What is the average cost per install in the US?

It varies widely by category and platform. For shopping apps, Adjust data for the first half of 2026 put US CPI at $3.28, while Adjust's 2026 trends report showed North American finance app CPI at $4.13 in 2025. Subscription and finance apps often pay more because each user is worth more.

What is the average cost per install in India?

India has some of the lowest CPIs in the world. Adjust data for the first half of 2026 put shopping app CPI in India at $0.10. Low CPIs come with quality and fraud risks, though, so Indian campaigns should be judged on cost per signup, KYC, first order or subscription rather than installs.

Why is CPI higher on iOS than Android?

iOS users in many markets spend more in apps on average, so more advertisers compete for them and auction prices rise. iOS also has a smaller share of users in markets like India, where AppsFlyer data shows Android accounted for 94% of installs in 2025. The size of the gap varies by country and category.

What should I track instead of cost per install?

Track cost per key in-app event, such as signup, trial or first purchase; cost per paying user; day-7 and day-30 return on ad spend; payback period; and retention. Use CPI as an input and an early warning signal, not as the goal that campaigns optimize toward.

Keep reading

More in Cost per Install Benchmarks · All app marketing guides