How much can a subscription app pay per install on Apple Ads?
A breakeven worksheet for Apple Ads (formerly Apple Search Ads): maximum CPI, maximum cost per tap and the four campaign types I start with. The examples come from my own iOS apps, which I use as a lab.
Most Apple Ads accounts start from the wrong number. Someone picks a bid because the suggested range looked reasonable, the campaign spends, and only later does anyone ask what an install is worth.
Run it the other way round. Work out the most you can pay for an install, turn that into the most you can pay for a tap, then build the campaigns under that ceiling. Everything below is arithmetic you can do in a spreadsheet in ten minutes.
The formula
For a subscription app, the ceiling for one install is:
max CPI = net revenue per payer × install-to-paid rate
Net revenue per payer is what reaches you after tax and the store fee, for the period you are willing to wait to be paid back. Install-to-paid rate is the share of installs that start paying within that same period.
From there, the ceiling for one tap:
max cost per tap = max CPI × tap-to-install rate
Apple Ads shows the tap-to-install rate as the conversion rate. On search results you pay per tap, so this is the number your bids have to respect.
Step 1. Net revenue per payer
- Start from what the user pays, not the list price on your pricing page. Introductory offers and discounts count.
- Take tax out where the storefront price includes it. Most European storefronts do.
- Take the store fee out. It is 30% by default. It is 15% if you are enrolled in the App Store Small Business Program, which you have to apply for. It is not automatic. For auto-renewable subscriptions it also drops to 15% once a subscriber passes one year of paid service, which for an annual plan only helps from the first renewal.
- Pick the payback period. If you can only wait for the first payment, use the first payment. If you can fund three months, use three months of revenue. This choice moves the ceiling more than any bid strategy.
Example: an annual plan at $79.99 with a 30% fee leaves $55.99 per payer before tax effects. At $35.99 it leaves $25.19.
Step 2. Your install-to-paid rate
Use your own cohorts, not a benchmark. Published conversion benchmarks mix apps with and without trials and across categories, so the median tells you little about your app. If you don't have enough installs for a stable number yet, write the sample size next to it and treat the ceiling as a range.
This table shows the install-to-paid rate you need to break even on the first annual payment, at a 30% store fee and before tax:
| CPI | At $35.99 a year | At $79.99 a year |
|---|---|---|
| $1 | 4.0% | 1.8% |
| $2 | 7.9% | 3.6% |
| $3 | 11.9% | 5.4% |
| $4 | 15.9% | 7.1% |
| $6 | 23.8% | 10.7% |
Read the $3 row. At a 30% store fee, a $3 CPI needs about 11.9% install-to-paid at $35.99 a year, against 5.4% at $79.99. The price you charge decides which channels you can afford before you open any ad account.
With the Small Business Program rate of 15%, the same $3 install needs 9.8% at $35.99 and 4.4% at $79.99.
Step 3. From CPI to cost per tap
Multiply by your tap-to-install rate. If your maximum CPI is $3 and half of your taps install, your maximum average cost per tap is $1.50. If 30% install, it is $0.90.
Brand searches usually convert better than generic ones, so their ceiling is higher. Set a ceiling per campaign, not one number for the whole account.
What this looked like on my own apps
I run Apple Ads for my own apps with my own money before I recommend anything to a client. The scale is small and I treat it as a lab, not as a case study.
I paused two Apple Ads campaigns for my own apps once measured CPI (EUR 2.35 and 3.24) sat far above each app's breakeven ceiling (EUR 0.44 and 0.49).
The measured CPI was between five and seven times the ceiling. No bid change closes a gap like that. The fix has to come from the offer: price, paywall and conversion, which is exactly what the table above shows.
The second lesson came from organic search, which tells you where paid search is needed:
I tracked 324 App Store search terms for my own apps: they showed up for only 13% of the terms taken from the keyword field, against 53% of brand terms and 39% of title and subtitle terms.
So I don't assume the keyword field earns visibility on its own. The terms you don't show up for organically are the ones you may need to pay for, and the ones you already own may need little paid support.
The four campaign types I start with
- Brand. Your app name and close variants, exact match. High intent and usually the cheapest installs. Keep it in its own campaign so its strong numbers don't hide weak generic ones.
- Category. Generic terms that describe what the app does, exact match, with bids under the ceiling you calculated.
- Competitor. Other apps' names, exact match. Usually the lowest tap-to-install rate, so the lowest ceiling of the four.
- Discovery. Search Match and broad match at a lower bid. Its job is to find terms. Move the ones that convert into the category or competitor campaigns as exact match, and add them as negative keywords in discovery so the two campaigns don't compete.
Negative keywords go in from day one: searches for free alternatives, unrelated apps that share a word with yours, and job or tutorial searches.
Check that attribution reaches your analytics
Apple Ads attribution comes through Apple's AdServices framework. If the app doesn't collect the attribution token and pass it to your MMP or analytics, the installs exist in Apple Ads and disappear downstream, and every calculation above runs on the wrong denominator.
I recovered Apple Ads attribution that my analytics pipeline was dropping (reconciled 100% with GA4) and issued a standard SDK checklist to 11 app repositories.
Before scaling, compare the installs Apple Ads reports with the installs your analytics attributes to Apple Ads for the same days. They won't match to the unit. They should be in the same range, and the gap should be stable.
Put a fuse on every launch
Give every new campaign a daily budget and an end date. If nobody looks at it for a week, it stops on its own.
Write the stop rule before launch too. For example: pause a keyword group that spends twice the maximum CPI without an install, or one whose CPI stays above the ceiling after enough taps to judge. When the rule is written first, nobody argues with it later.
If the ceiling sits far below what installs cost in your category, more optimization won't save it. The levers are price, paywall, onboarding and the plan mix. That is why I put the offer and the measurement before the channel.
Checklist
- Net revenue per payer after tax and the store fee, for a payback period you can fund
- Install-to-paid rate from your own cohorts, with the sample size written next to it
- Maximum CPI, then a maximum cost per tap for each campaign
- Brand, category, competitor and discovery campaigns, with negatives from day one
- AdServices attribution checked against your analytics
- A daily budget, an end date and a written stop rule on every launch
Want these numbers worked out for your account? The 30-minute diagnostic is free.
