Guide · Apple Ads

What is a good ROAS for Apple Search Ads?

Every answer to this question that gives you a single number is wrong, and the ones quoting "aim for 3x" are usually selling ad management. ROAS for a subscription app is not a score, it is a reading taken at a moment you choose, and choosing a different moment changes the answer completely.

The number moves with the window

Return on ad spend is attributed revenue divided by spend. For a one-time-purchase app that settles quickly. For a subscription app, the revenue from a single install arrives over months, so the window you pick decides what you see:

• 7-day ROAS mostly measures your trial-to-paid conversion, not the keyword.
• 30-day ROAS catches the first renewal for monthly plans, and almost nothing for annual ones.
• 12-month ROAS tells the truth and arrives far too late to run a campaign on.

An annual-plan app can show 30-day ROAS of 0.3 and still be one of the healthiest businesses in its category. The number is not lying, it is simply answering a question you did not mean to ask.

Use payback period, not a benchmark

The question worth asking is: how long until this keyword pays back what it cost? That reframes the decision from an arbitrary threshold to a cash-flow one you can actually act on.

1. Take your average revenue per subscriber over their lifetime, or a conservative estimate of it.
2. Divide the spend on a keyword by the number of subscribers it produced. That is your acquisition cost for that keyword.
3. Work out how many months of subscription revenue cover that cost.
4. Compare that against how long you can afford to wait. That is the only benchmark that is yours.

An indie developer funding ads from revenue might need payback inside two months. A funded studio might happily wait nine. Same keyword, same ROAS, opposite decisions, both right.

Per keyword, not per campaign

Campaign-level ROAS averages away the thing you need to see. A campaign returning 1.8 can easily contain one keyword returning 6.0 and four returning 0.2. The average says the campaign works. The breakdown says to move the entire budget onto one term.

This is why the useful unit is the keyword row: what it cost, what came back, and what its organic position was while that happened.

The catch with attribution

Calculating ROAS per keyword needs two datasets joined: Apple Ads spend by keyword, and subscription revenue by user, tied back to the keyword that delivered them. Apple gives you the first. Your subscription platform gives you the second. Nothing gives you both, which is why most teams stop at cost per install and call it a strategy.

Attributing revenue with RevenueCat covers how that join works in practice.

Common questions

What is a good ROAS for Apple Search Ads?

There is no single number, because ROAS depends entirely on the window you measure. A subscription app measuring 30-day ROAS will look far worse than the same app measuring 12-month ROAS, and both can be correct. The useful question is whether a keyword clears your payback period, not whether it beats someone else's benchmark.

How do I calculate ROAS per keyword?

Attributed revenue from that keyword divided by spend on that keyword, over the same window for both. The hard part is not the arithmetic, it is attributing subscription revenue back to the keyword that produced the install, which needs your ads data and your subscription data joined on the same window.

Is ROAS below 1.0 always bad?

No, for a subscription app it is often expected early. If a keyword returns 0.6 at 30 days but your average subscriber pays for eight months, that keyword may be strongly profitable by month four. It is only bad if you never measure the later window and never find out.

Should I optimise for ROAS or cost per install?

Cost per install tells you what a person cost, not what they were worth. Two keywords with identical CPI can differ several times over in revenue. Use CPI to spot a keyword getting expensive; use ROAS to decide whether it deserves the money at all.

A working rule

If you want one sentence to carry away: judge a keyword on whether it pays back inside a window you can afford, measure it per keyword rather than per campaign, and treat any published benchmark as a description of somebody else's business.