This article is based on my interview with Jan Gerwin for App Vision Updates. The original German conversation appears there alongside App Vision’s reporting on Apple’s foldable phone.
When a team brings me a new consumer app and asks where paid acquisition should start, my default answer is Meta, especially on iOS.
That is not because Meta is simple. Tracking iOS campaigns on Meta is harder than on other channels, and some teams respond with web redirects or complete web funnels. Android brings the opposite problem for many apps, because its users are harder to monetize. Neither platform gives you an automatic win.
Meta is usually the best first learning environment because it connects campaign structure, creative testing and funnel diagnosis. With reliable tracking and a simple account, you can see which messages attract the right people and where users drop off after the click.
Start with a structure you can read
A new account does not need a maze of campaigns splitting its signal. It needs an optimization event that represents real value, enough budget for the campaign to produce evidence, and a structure simple enough that someone can explain where the money went.
Then put real energy into the ads.
Whether the first round contains three ads or one hundred, creative volume and iteration decide whether Meta becomes a scalable channel. More useful variations create more chances to find a message that works. Winners are not finished assets either, but starting points for the next set of hooks, formats and executions.
Facebook and Instagram are the obvious placements, while Audience Network can extend delivery into other publishers’ apps. The placement list matters less than whether the account has enough signal and enough distinct creative ideas to learn.
When Google should come before Meta
Google deserves an early test when Android revenue is already strong.
App campaigns are easy to launch. A few headlines and descriptions can travel a long way because Google combines them with the Play Store title, icon and screenshots, then distributes the result across Search, Google Play, YouTube and other inventory.
The tradeoff is patience and test size. Many teams wait weeks before Google reaches stable performance. If I would reserve about $10,000 a month for a meaningful Meta test, I would plan closer to $20,000 a month for Google. Once it is working, Google usually needs less day to day attention than Meta.
But cheap Android installs do not repair weak Android revenue. Existing revenue data is evidence that the platform deserves a test, while a low CPI is not.
When TikTok should come first
TikTok moves forward in the order when the product itself gives it a reason to.
Some apps are built around an idea people naturally want to show, talk about or share with friends. That can create an organic acquisition path before paid media begins. I would consider TikTok early when there is real viral pull, a heavily Gen Z audience, or a compelling offer that saves money.
Being easy to demonstrate is not enough on its own. In the accounts I have managed, TikTok often required frequent creative refreshes. That is an operational observation, not a universal finding across channels. Videos need refreshing before performance visibly declines. Meta has the same requirement, but TikTok often compresses the cycle.
For an established brand without one of those product advantages, TikTok normally comes later in my channel order.
When Apple Ads should come first
Apple Ads make sense when people already search for your brand, a competitor or the problem your app solves. Your own name does not have to be famous, but the category needs language people already use.
A completely unfamiliar idea is harder because users may not know what to type. Search Results are the logical starting placement when demand exists. Search Tab and Today Tab have different economics, so treat them as different inventory.
Keep brand terms separate. Some users who click a brand ad already intended to find the app. Test whether that spend creates incremental customers before counting all of its return as new growth.
What real channel data showed
In one AI utility subscription app, which I wrote up in full here, Meta reached 1.61x gross media ROAS by day 28 on $181,800 of spend. Google reached 1.09x, despite delivering much cheaper installs.
The full comparison for that app was:
One AI utility subscription app, gross media revenue by day 28, with the spend behind each ROAS. The dashed marker is 1.0x, gross media parity. Green bars clear it and red falls below it. The budgets are very different, and TikTok's ROAS is an estimate because its cost was not tracked in the MMP.
This is evidence from one app, not a promise of what every app will see or a controlled test of four channels at equal spend. It does show why CPI and ROAS have to be read together. Google bought installs at roughly a quarter of Meta’s cost and still produced less revenue per dollar by day 28. Optimizing for cheap installs would have selected the wrong winner.
The revenue figures are gross. Your own ROAS definition must account for store fees, taxes, refunds and the payback window your cash flow can support. A 3:1 LTV to CAC ratio is a rough guide for when an app is set up to scale, not a threshold that fits every app.
A channel needs a product reason
The choice is about what evidence the product already gives you, not which platform is fashionable.
| Channel | When it comes first | What to watch |
|---|---|---|
| Meta | My default, especially on iOS · about $10K a month | iOS tracking, creative volume |
| Google Ads | Android already monetizes · about $20K a month | Slow to stabilize |
| TikTok | Viral pull, a Gen Z audience or a savings offer | Frequent creative refreshes |
| Apple Ads | Search demand already exists | Whether brand spend is incremental |
The signal each platform needs before it can learn is in my article on which event a subscription app should optimize for.
Then judge the result on revenue and payback, not the cheapest install. Paid acquisition becomes useful when it tells you which customers you can profitably buy. Everything before that is an input metric.
Source and scope
The channel order and $10K/$20K test reserves are my planning preferences, not platform minimums or cutoffs the industry has proven. Google’s July 2025 guidance ties App campaign budgets to target CPI or CPA and warns about conversion delay. That supports budgeting from your economics, not a universal dollar amount. The figures for the four channels above are my private account comparison, not independently verified research.