Putting iOS and Android into a single app install campaign feels efficient. One budget, one set of creative, one number to report. The problem shows up the moment that number moves and you cannot say which platform moved it.
The two platforms do not behave the same
Install costs sit at different levels on iOS and Android in most markets, and they move for different reasons. Store review timing, device mix, and how each store handles attribution all pull the two apart. When they share a campaign, automated bidding quietly decides the split for you, and it optimises towards whichever platform is cheaper rather than whichever one is worth more after the install.
That last part matters. A cheap install that never opens the app again is not a win. You only see the difference if the two platforms are reported separately.
What splitting actually gives you
Two numbers instead of one, and both of them are decisions. If Android installs cost a third of iOS but iOS users convert to a listing enquiry more often, that is a budget conversation you can have with real figures. Blended, it is invisible.
On a recent seven month TikTok program I ran the app install campaigns separately for iOS and Android for exactly this reason, alongside parallel reach campaigns during Ramadan and Eid. The full numbers are in the TikTok case study, including what the split let us change week to week.
Do this before you scale
Splitting is cheap while budgets are small and expensive to retrofit once a campaign has months of learning behind it. If you are planning to increase spend, separate the platforms first, let each gather its own data, then scale the one that earns it.
Paid install campaigns only pay off if the listing they point at converts, which is a separate job. That side is covered on the app store optimisation page, and the campaign side on paid social.