Good CPI for Apps in India: Category Benchmarks

What CPI actually measures

Cost per install (CPI) is total paid media spend divided by the installs attributed to that paid spend, and nothing else. It is not spend divided by every install you got in the month.

That distinction sounds pedantic, but it is where most CPI figures go wrong. Keep three things in mind:

• Numerator: spend on the campaigns you are measuring, pulled from each ad network.
• Denominator: installs that an attribution tool matched to those campaigns, deduplicated across networks.
• Excluded: organic installs, installs from untracked links, and reinstalls you choose not to count.

What is a good CPI for an app in India?

A good CPI for an Indian app is one at or below the median for its category, which sits around ₹18 to ₹21 for e-commerce, productivity, social and fintech apps. Health and fitness runs higher at about ₹31, and EdTech, entertainment and astrology apps routinely pay several times that.

The figures below come from Linkrunner's published research on Indian app startups, which reports medians and the middle half of apps (the range between the 25th and 75th percentile).

A few things stand out:

• The cheap categories cluster tightly. E-commerce, productivity, social and fintech all sit within a few rupees of ₹20 at the median.
• Fintech has a long upper tail. The top quarter of fintech apps pay over ₹60 per install, usually because they target narrower, higher-value audiences.
• Astrology, entertainment and EdTech figures are directional. These categories have small samples, so use the range as a rough guide rather than a target.
• Gaming and food and beverage have no published CPI benchmark yet. Compare against the overall median until they do.

For per-category detail, see the fintech benchmarks or the full CPI benchmarks for India.

Why does my CPI look better than the benchmark?

Usually because organic installs or double-counted network installs have leaked into the calculation. A CPI that looks unusually low is more often a measurement problem than a media-buying win.

The common causes:

• Organic installs get mixed in. If you divide spend by all installs, organic users make paid look cheap. The median Indian app gets about 62% of its installs organically, so a blended CPI can understate true CPI by more than half.
• Every network claims the same install. Meta and Google each count installs they touched, so adding up their dashboards overstates paid installs and understates CPI.
• Reinstalls are counted as new users. Users who reinstall after a phone change inflate the install count without adding reach.

Why does my CPI look worse than the benchmark?

A high CPI is not automatically bad. It often reflects choices rather than inefficiency:

• Audience targeting. Narrow audiences (high-income, specific cities, specific professions) cost more per install but can convert far better.
• Platform mix. iOS installs usually cost more than Android installs, and Android makes up the large majority of Indian installs. An iOS-heavy campaign will sit above the median. See Android vs iOS in India.
• Optimisation event. Campaigns optimised for signups or purchases rather than installs pay more per install by design.
• Creative fatigue. CPI tends to creep up as the same creatives run for weeks without refresh.

CPI is only half the picture

CPI tells you what an install costs, not what a user is worth. A ₹15 install that never signs up costs more than a ₹30 install that does.

• Track cost per signup alongside CPI. The median Indian app pays about ₹48 per signup, which captures quality as well as price.
• Check install to signup rate by campaign. The median Indian app converts about 20% of installs to signups. Campaigns well below that are buying the wrong users, however cheap they look. See the install to signup benchmarks.
• Read the full breakdown in what a signup really costs.

How to measure CPI properly

To get a CPI you can trust, you need one source that deduplicates installs across networks and matches spend to installs by campaign. That is the job of a mobile measurement partner (MMP), such as Linkrunner, Tenjin or Kochava.

A working setup should:

1. Pull spend from each ad network automatically, rather than relying on manual exports.
2. Attribute each install to one campaign only, using a consistent lookback window.
3. Separate organic installs from paid ones, and report them side by side.
4. Report cost per signup and cost per purchase as well as CPI.

You can see true CPI and cost per signup for every campaign, with spend pulled from Meta and Google, in Linkrunner.

FAQ

What is the average CPI for apps in India? The median Indian app startup pays about ₹20 per paid install. The median is a better guide than the average, because a handful of high-cost categories such as entertainment and astrology pull the average up sharply.

Which app categories have the lowest CPI in India? E-commerce, productivity, social and fintech apps all have median CPIs between about ₹18 and ₹21. They are the cheapest categories to acquire users in, although fintech has a long tail of apps paying much more.

Why is CPI so high for entertainment and astrology apps? Both categories have median CPIs above ₹110, partly because of competitive audiences and partly because the figures come from small samples. Treat these numbers as directional and compare cost per signup as well, since astrology apps in particular tend to convert installs to signups at high rates.

Should I include organic installs when calculating CPI? No. CPI should only count installs attributed to paid campaigns. Including organic installs makes paid look much cheaper than it is and hides campaigns that are losing money.

Is a low CPI always good? Not on its own. A low CPI with a weak install to signup rate can cost more per real user than a higher CPI with strong conversion, so always read CPI next to cost per signup.