On this page
- Where to read IPM
- Why installs per mille matters for app campaigns
- IPM vs CTR, click-to-install rate, CPI, and hook rate
- How creatives move IPM
- How to test creatives on IPM
- Gaming IPM by region in Adjust's 2026 report
- Is a high IPM always good?
- IPM for CTV app campaigns
- Creative testing with Starti AI Studio
- Frequently asked questions
IPM, or installs per mille, is the number of app installs an ad generates for every 1,000 impressions. Also written as installs per thousand impressions, the IPM metric measures the whole path from an ad being seen to the app being installed, which makes it the first number mobile game and app teams read when they test creatives.
The formula:
IPM = installs ÷ impressions × 1,000
An ad with 40,000 impressions and 120 installs has an IPM of 3.0. Unity's glossary describes the same calculation: "multiplying the number of campaign installs by 1000, and then dividing that number by the number of impressions."
Where to read IPM
Some ad platforms report the rate directly, and on others it comes from two columns:
- Unity Ads: the dashboard metrics include "Installs per Mille," and Unity counts impressions as "Starts."
- Google Ads: App campaign asset reports include "Installs per (1000) impressions."
- TikTok: Ads Manager reports impressions and conversions, so IPM on an install campaign is conversions ÷ impressions × 1,000.
- Other networks: divide attributed installs by the network's own impression count.
- iOS: SKAdNetwork and AdAttributionKit report installs in postbacks that Apple sends after a random delay of 24 to 48 hours for the first one, so a same-day iOS IPM is incomplete.
Compare creatives inside one network and one set of ad placements. Unity's creative testing guide gives the reason: "The same creative will produce different IPM and CPI on Unity Ads and on a social network."
Why installs per mille matters for app campaigns
The metric folds two steps of the funnel into a single number: it "reflects both click-through and conversion performance, offering a unified view of campaign effectiveness," in AppsFlyer's metric definitions.
On networks that rank campaigns by eCPM, such as Unity Ads, installs per mille also shapes how much inventory a CPI campaign wins. The user acquisition formula in Unity's eCPM guide is "eCPM = CPI * IPM," and the guide explains that networks rank campaigns by eCPM and serve "the campaigns with the highest eCPMs first and most often." The company's bidding documentation puts the rule plainly: "The higher your campaign's eCPM, the more exposure it receives."
Two consequences follow for a team buying installs on such a network:
- Delivery: at the same CPI bid, a creative with a higher IPM produces a higher eCPM and wins more impressions.
- Cost: rearranging Unity's formula gives CPI = eCPM ÷ IPM. At an estimated $10 eCPM, an IPM of 5 means a $2 CPI, and an IPM of 8 means $1.25.
IPM vs CTR, click-to-install rate, CPI, and hook rate
Each metric covers a different stretch of the path to install:
- IPM: installs ÷ impressions × 1,000. Covers the whole path, from impression to install.
- CTR: clicks ÷ impressions, which AppsFlyer writes as "Number of clicks / number of ad views." Shows whether the ad earns the tap.
- Click-to-install rate (CTI): installs ÷ clicks. Shows whether the store page delivers what the ad promised.
- CPI: spend ÷ installs. Shows what each install cost, and moves with IPM and the CPM paid.
- Hook rate: the share of impressions that keep watching past the opening. Unity's creative testing guide lists hook rate, completion rate, and CTR as diagnostic metrics that "explain why a creative wins or loses."
Multiplying the definitions out gives IPM = CTR × CTI × 1,000, so a change can come from either step. On Unity's dashboard, the conversion rate, or CVR, "is calculated by dividing the number of Installs by the number of Starts," the same ratio as IPM on a scale of one rather than 1,000.
Installs per mille can rise while CTR stays still. In Singular's creative volume analysis, published October 26, 2026, IPM "roughly doubles and install rate climbs nearly 2.6x, while click-through rate (CTR) stays essentially flat" from the lowest tier of creative density to the tier with 13 to 54 creatives per $1,000 of spend.
How creatives move IPM
At the test stage, teams are told to "read installs per mille (IPM) first" in Unity's creative testing guide, published October 15, 2026. When two creatives run in the same campaign to the same audience, it says, "the gap in IPM comes largely from the creative." The parts the guide names include:
- Hook: the first 1 to 3 seconds, which the guide calls "the largest single driver of whether users see the rest."
- Ad format: video, playable, or video followed by a playable; the guide notes that "Often, changing format moves performance more than any single element."
- Gameplay depiction: how much real gameplay, or real app use for a non-game app, the ad shows.
- Copy and call to action: on-screen text and the final prompt to install.
- Layout: portrait or landscape, interface density, and where the end card sits.
In terms of the formula, the hook and the format work mostly on the tap, while gameplay depiction, the call to action, and the end card work on whether a tap becomes an install.
The link between creative variety and IPM holds only up to a point. In Singular's data, drawn from gaming, fintech, food delivery, and other advertisers, IPM fell back to 1.2x the lowest tier above 54 creatives per $1,000 of spend, and Singular adds that "Adding creative only helps when the new creatives are different ideas."
A falling rate also signals fatigue, and Unity's guide treats it as the cue to refresh: "When a top creative's IPM slides while spend holds steady, fatigue has set in, and that's your refresh trigger." Platform signals for the same problem are covered in ad fatigue.
How to test creatives on IPM
Run the variants in one campaign, to one audience, over the same days, so the creative is the only difference:
- Variants: two to four, which Unity's guide says "suit most budgets," each differing in one large element such as the hook or the format.
- Volume: Unity's testing documentation asks for "at least 20 thousand impressions for each creative pack," and its scaling guide asks for at least one week and "at least 10,000 starts or 100 installs per creative pack." Meet the higher floor, and check it with a sample size calculator, as Unity's documentation recommends.
- Read: IPM for the shortlist, with hook rate and CTR to explain the result.
- Decide: day-7 revenue and retention for the final call. Unity warns that "A creative can win on IPM and lose on ROAS, usually because it attracted players the game didn't retain."
D7 ROAS explains how to read the day-7 numbers, and creative testing sets out the order of a full test.
Example: An IPM test for a word puzzle game
The plan:
- Variants: three creative packs with the same gameplay and end card, and three hooks: a near-miss puzzle, a timer running out, and a hand solving a word in one move.
- Volume: 25,000 impressions per pack over eight days on one network.
- Result: 80, 160, and 140 installs, an IPM of 3.2, 6.4, and 5.6.
- Shortlist: the timer hook and the one-move hook go forward, and the near-miss hook, with half the installs of the leader, is dropped.
- Final call: the two remaining packs run against each other until their day-7 cohorts mature, and the pack with the higher D7 ROAS scales.
Gaming IPM by region in Adjust's 2026 report
Adjust's Gaming App Insights Report 2026, covering January 2024 to January 2026, puts global gaming IPM at 8.62 in 2025, up from 8.59. Regional figures for 2025 from the same report:
- APAC: from 7.59 to 8, with Indonesia at 8.58, South Korea at 6.13, Japan at 5.27, and Singapore at 3.23.
- Europe: DACH at 4.92, and the UK and Ireland at 5.46.
Genre targets from the same report are set out in mobile game marketing strategy; within a single test, the target is the creative it has to beat.
Is a high IPM always good?
A high rate wins impressions and lowers CPI, yet it reveals nothing about the users who installed. A hook that promises something the app does not deliver can raise IPM and lower retention, the pattern behind Unity's warning that a creative can "win on IPM and lose on ROAS." Use IPM to pick which creatives deserve spend, and day-7 revenue to decide which ones keep it.
IPM for CTV app campaigns
On connected TV, the impression runs on the television while the install happens on a phone, and measurement partners join the two through view-through attribution. Each partner sets its own rules: AppsFlyer's cross-platform links default to a 24-hour window, and Adjust's CTV measurement documentation ranks CTV impressions last by default. A CTV campaign's IPM counts only the installs matched under those rules, so compare it across CTV creatives rather than against a mobile network's IPM. The measurement paths are set out in CTV ad measurement.
Creative testing with Starti AI Studio
AI Studio produces the variants a creative test needs, and Video Agent can revise an individual scene, such as the hook, without rebuilding the entire ad. Smart Insight brings Meta, TikTok, Google Ads, and configured custom networks into the same analysis and compares nine creative attributes against CTR, completion, and other outcomes. Its Winner, Scaling, Watch, and Fatigue signals help teams prioritize which campaigns and creatives to review.
At Starti, AI handles planning, generation, revisions, and analysis, and the creative team and the client make the final calls, with the findings feeding the next brief.
Book a demo to plan a first creative test.
Frequently asked questions
What is IPM in advertising?
IPM, or installs per mille, is the number of app installs an ad generates per 1,000 impressions. It combines click-through and click-to-install performance in one number.
How do you calculate IPM?
Divide installs by impressions and multiply by 1,000. An ad with 120 installs from 40,000 impressions has an IPM of 3.0.
What is a good IPM?
IPM varies by genre, market, and network. Adjust's 2026 gaming report puts global gaming IPM at 8.62 in 2025 and APAC at 8. Within a test, a good IPM is one that beats the current best creative on the same network and audience.
What is the difference between IPM and CTR?
CTR counts clicks per impression; IPM counts installs per impression. It includes the click and the install, so an ad can raise its CTR and still lose on IPM if the store page does not convert.
How does IPM affect CPI?
On networks that rank campaigns by eCPM, such as Unity Ads, Unity's formula eCPM = CPI × IPM means a higher rate wins more impressions at the same bid. Rearranged, CPI = eCPM ÷ IPM, so at an estimated $10 eCPM an IPM of 5 gives a $2 CPI.






