A US impression costs 14× an Argentine one
And the US ranks 26th of 27 countries on D7 ROAS, at 33%. The Netherlands, Norway and Portugal return 122–133% in week one.
FreeBenchmark report · Q3 2026
Meta Ads benchmarks for subscription apps
What reach, installs, trials and paying users really cost on Meta, and where spend pays back in week one. Built from campaign data of 5K subscription apps across 45 countries.
- median CPM
- $8.87
- per paying user
- $64.93
- D7 ROAS
- 39%
Get the full report
29 pages of benchmarks. A free PDF, ready to download as soon as the form is sent.
Ready to read
Your report is ready
Thanks! The PDF is yours to keep: 29 pages, about 10 MB.
5Start with page 5, the benchmark at a glance. It takes two minutes and shows where your account sits against the median app.
Signal · What you tell Meta to find
The cheapest trial isn’t the cheapest customer
Optimizing for Start Trial buys trials for $20.28. Optimizing for Subscribe costs more per trial, yet buys paying users for $42.77 instead of $57.32, and returns 75% of spend in week one instead of 45%.
Cost per payer$42.77Subscribe
Cost per payer$57.32Start Trial
Cost per payer$82.23Purchase
A preview of the findings
Numbers that change where your budget goes
5% of ads take 81% of the spend
Meta picks a few winners fast. Testing volume decides how often you find one, not how evenly budget spreads.
Reaching 55+ costs 4× more than 18–24
But they click 3–4× more, so cost per click stays between $0.44 and $0.67 across every age band.
Inside the report
29 pages of benchmarks, from impression to paying user
Every number is the median app, with P25–P75 ranges where the spread matters. Compare your account against the market, line by line.
Get the report- 01
Auction and audience
CPM by month, placement, age and gender.
- 02
Creative
Spend concentration, video vs image, hook rate and frequency.
- 03
Installs and subscribers
CPI, cost per trial, cost per paying user and D7 ROAS, iOS vs Android.
- 04
Optimization events
Subscribe vs Start Trial vs Purchase, and Advantage+ vs manual.
- 05
App categories
Seven categories by CPM, D7 ROAS and cost per payer.
- 06
Countries
CPM and D7 ROAS in 27 markets, mapped into four quadrants.
Frequently asked questions
Is the report free?
Yes. Fill in the form and download the PDF straight away. There's no sales call, and you don't need an Adapty account.
What's in the report?
29 pages in four parts: auction and audience, creative, installs and subscribers, and countries. It opens with the eight headline benchmarks on one page and closes with five takeaways you can act on.
It covers CPM, CTR, CPC, hook and hold rate, CPI, cost per trial, cost per paying user and D7 ROAS, cut by month, placement, age and gender, creative format, platform, optimization event, app category and country.
Where does the data come from?
Anonymised, aggregated Meta Ads campaign data from 5K subscription apps connected to Adapty, across 45 countries, July–September 2026. April–June is used only for the monthly CPM trend.
Installs are Meta-reported (7-day click + 1-day view). Trials, paying users and revenue come from Adapty's D7 install cohorts for Meta campaigns, so they reflect what actually happened in the app, not what Meta modelled.
How are the benchmarks calculated?
Every benchmark is the median app, with P25–P75 ranges where the spread matters. That way one huge advertiser can't skew the number, and you can see whether you sit in the top or bottom quartile.
Ratios such as CTR and CPM use summed spend, impressions and clicks. Thin samples are left out: a cut is shown only when enough apps sit behind it, which is why 27 of the 45 countries make the D7 ROAS ranking.
What does D7 ROAS mean here?
Gross revenue in the first 7 days after install, divided by spend. It's an early read on payback: trials that convert later and renewals aren't included yet.
Use it to compare campaigns, events and markets against each other, not as your final payback target. A campaign well below 100% in week one can still pay back over a longer window.
Can any app, campaign or ad be identified?
No. Data is anonymised and aggregated before analysis, and no app, campaign or ad can be identified from the report.
Who is it for?
UA managers, growth leads and founders running Meta Ads for a subscription app who want to know if their costs and returns are normal.
It's most useful when you're setting targets, planning next quarter's budget, choosing an optimization event or deciding which countries to test.
Does it cover Facebook and Instagram separately?
Yes. The report breaks down CPM, CTR and CPI for six placements: Instagram Reels, Stories and Feed, and Facebook Reels, Feed and Stories. It also shows how much of each placement's installs come from views rather than clicks.
Does it cover both iOS and Android?
Yes. The report compares CPI and installs per 1,000 impressions on each platform, and shows how iOS apps perform against the median app down to cost per paying user and D7 ROAS.
iOS installs are partly modelled by Meta, so treat the iOS install figures as estimates.
Do I need to use Adapty to get value from it?
No. The benchmarks apply to any subscription app running Meta Ads, whatever your stack. Two pages describe Adapty Attribution and are clearly marked.
How do I act on these benchmarks?
Start by placing your own account against the medians and P25–P75 ranges, then fix the biggest gap first. For the how-to side, read our free Meta Ads playbook for subscription apps. It covers creative testing, the signals you send Meta, and scaling spend without hurting payback.
How often is it updated?
Every quarter. This edition covers Q3 2026.
Know what pays before it’s paid back
Benchmark your Meta campaigns against 5K subscription apps.
Get the free report