Every account we audit that runs both platforms has the same conversation waiting inside it. Someone opens the platform breakdown, sees that Android installs cost a fraction of iOS installs, and reaches the obvious conclusion: move the budget.
The iOS vs Android CPI gap is real, it is large, and it is almost entirely structural. It is not an inefficiency sitting in the account waiting to be arbitraged, and treating it as one is how growth teams spend a quarter buying more installs and the same amount of revenue.
Here is how big the gap actually is by category, what causes it, why the cheaper platform is usually not the better buy, and how to structure campaigns so the comparison stops misleading you.
Key takeaways
- The iOS vs Android CPI gap runs around two times on global averages and stretches to five times or more in gaming, where the spread between platforms is widest.
- The gap is caused by who holds the device, where the install base sits geographically, and how much competition there is for a smaller pool of high value users. Meta does not price iOS higher.
- A cheaper install is not a better install. Platform allocation should be decided on the ratio of lifetime value to cost per install, not on cost per install on its own.
- The two numbers are not measured the same way. iOS runs through SKAdNetwork with delayed and modelled postbacks, so an early iOS cost per install is systematically incomplete.
- Run the platforms as separate campaigns with separate budgets and separate benchmarks, and judge each one against itself over time.
How big the iOS vs Android CPI gap actually is
Start with the aggregate. Global average cost per install in 2026 runs at 2.24 dollars on iOS and 1.12 dollars on Android, and in Western Europe at 3.40 dollars against 1.85 dollars (Searchlab, citing Adjust and AppsFlyer data). That is a clean two times at the top level, and it is the number most people have in their heads.
The aggregate hides the interesting part, because the gap widens sharply once you look inside gaming. Liftoff's 2025 Casual Gaming Report, built on 1.4 trillion impressions, puts casual titles at 1.41 dollars on iOS against 0.14 dollars on Android. Midcore comes in at 3.65 dollars on iOS against 0.73 dollars on Android. Casino sits at 21.03 dollars on iOS, the highest single category Liftoff records.
So the working band is two to five times for most consumer apps, with gaming capable of stretching well past that. If your own split sits inside that range, nothing is broken. If iOS is running ten times Android outside gaming, that is worth investigating, and it is usually a geography or a creative coverage problem rather than a platform one. Our mobile app CPI benchmarks for 2026 break the numbers down by vertical and geography if you need a closer comparison for your category.
Why iOS costs more, in three parts
It is the people, not the platform. Meta does not charge a premium for iOS inventory. Advertisers do, collectively, through the auction. iOS device holders spend more inside apps on average, which raises the expected value of an iOS install, which raises what everyone is willing to bid for one. The platform gap is a demand gap that shows up as a price.
Geography is doing more work than you think. Android's global install base is weighted towards markets with lower purchasing power, so a meaningful share of what looks like a platform discount is really a geography mix. Liftoff's midcore data makes the point better than any platform comparison: 0.27 dollars in LATAM against 5.45 dollars in North America, a twenty times spread inside a single genre. Even between developed markets the effect is visible, with UK cost per install running 20 to 40 percent below the US across most genres (Amps33, 2026). A lot of "Android is cheap" is actually "Android is where the cheap geographies are".
Competition concentrates on the smaller pool. In most Western markets iOS holds a minority of devices but a majority of consumer app revenue. More advertisers chasing fewer impressions produces exactly what you would expect, and it is why the gap is widest in categories where monetisation is strongest.
Why you usually should not shift budget on CPI alone
Cost per install is a cost input. It is not a result, and no account has ever been profitable because its installs were cheap. The number that should decide platform allocation is the ratio between what a user is worth and what they cost, and FoxData puts the minimum viable threshold at 1.5 times lifetime value to cost per install (FoxData, 2026).
Run that arithmetic and the intuition inverts. An iOS install costing three times an Android install is the better buy whenever the iOS user is worth more than three times as much, which in subscription categories is routinely the case. The platform with the worse cost per install is frequently the platform with the better economics, and the only way to see that is to stop looking at cost per install on its own.
In our experience, accounts that reallocate heavily to Android on cost per install alone see the same thing happen: install volume climbs, the dashboard looks better, and revenue stays flat. The installs were cheaper because they were worth less, which was the information the price was carrying all along.
There are real cases where Android is the right answer, and they are worth naming so this does not read as a blanket rule. Ad monetised apps, where revenue comes from impressions rather than purchases, care about install volume in a way subscription apps do not. So do apps with strong viral or network effects, where scale is the product. And tier 3 expansion is a genuine Android play. The distinction is that each of those is a decision about the revenue model, not a decision about a price tag.
The two numbers are not measured the same way
Even if you accept the economics, there is a measurement problem underneath the comparison that makes the raw gap look wider than it is.
Android reports conversions deterministically and quickly. iOS runs through SKAdNetwork, where postbacks arrive aggregated, delayed and partially modelled, and where installs Meta never receives credit for simply do not appear. Under attribution inflates the apparent iOS cost per install, because the spend is fully counted and the conversions are not. We covered the timing in detail in our piece on SKAdNetwork postback delays, and the practical consequence here is simple: comparing yesterday's iOS cost per install to yesterday's Android number is not a comparison at all.
Use a seven day rolling view for any platform comparison, and expect the iOS number to improve as postbacks land. The gap you see at twenty four hours is not the gap you will see at the end of the week.
How to structure campaigns across platforms
The structural fix is unglamorous and it matters more than any bid adjustment.
- Separate campaigns, separate budgets. Never let one campaign allocate freely across both platforms, because it will optimise towards the cheaper install and you will have bought a cost reduction rather than a revenue increase.
- Separate benchmarks. Judge iOS against iOS and Android against Android, month over month. A cross platform target is a target nobody can hit on one of the two.
- Separate testing minimums. Your per concept testing spend should follow the cost per install of the platform it is running on, so an iOS test needs materially more budget than the equivalent Android test to produce a readable result.
- Separate reading windows. Give iOS longer before you judge it, for the postback reasons above.
Creative is the one place where separation is optional. Concepts that work usually travel across platforms, because the psychology of the audience does not change with the operating system. What changes is how much you pay to find out, which is an argument for validating new angles on the cheaper platform and amplifying the winners on the expensive one. It is also why creative coverage matters more on iOS: at a 3.40 dollar cost per install, a fatiguing portfolio is an expensive problem, which is the case we make in our guide to Meta ads creative fatigue for mobile apps.
Frequently asked questions
Why is iOS CPI higher than Android CPI?
Because advertisers bid more for iOS users, not because Meta charges more for iOS inventory. iOS device holders spend more inside apps on average, so the expected value of an iOS install is higher and the auction prices it accordingly. Android's global install base also skews towards markets with lower purchasing power, which drags its average cost per install down further.
What is a normal iOS vs Android CPI gap in 2026?
On global averages the gap is roughly two times, with iOS at 2.24 dollars and Android at 1.12 dollars according to Searchlab citing Adjust and AppsFlyer data. In gaming it stretches much further, with Liftoff reporting midcore titles at 3.65 dollars on iOS against 0.73 dollars on Android. Anything from two to five times is normal, and gaming can exceed that.
Should I move budget from iOS to Android if CPI is cheaper?
Usually not on cost per install alone, because cost per install is an input rather than a result. The number that decides platform allocation is the ratio of lifetime value to cost per install, and FoxData puts the minimum viable threshold at 1.5 times. An iOS install costing three times more is still the better buy if the user is worth four times more.
Can I compare iOS and Android CPI directly in Ads Manager?
Not on a like for like basis, because the two platforms are measured differently. Android reports conversions deterministically and quickly, while iOS runs through SKAdNetwork with delayed, aggregated and partially modelled postbacks. Early iOS numbers are systematically incomplete, so compare the platforms on a seven day rolling view rather than on yesterday.
Should iOS and Android run in the same campaign?
No. Run them as separate campaigns with separate budgets, because a single campaign allowed to allocate across both will chase the cheaper installs regardless of what those installs are worth. Separating them also lets you set different benchmarks, different testing minimums and different creative expectations for each platform.
Want this run for you?
If your platform split is being decided by whichever number is smaller, the fix is not a bid change. It is separating the campaigns, fixing the measurement window, and rebuilding the comparison around what a user is worth rather than what they cost.
That is the work we do as a performance creative agency for mobile apps. If you would rather have the whole system run for you, apply to work with us. We take a small number of mobile app clients per quarter.