mobile game UA benchmarks 2026: CPI, IPM and ROAS by genre.
what 'good' looks like by genre and platform — and a three-lever framework for telling whether a soft number is your creative, your audience, or your product.
you can't tell whether a campaign is healthy without knowing what "normal" looks like. this is a data-first map of CPI, IPM, ROAS, and retention for mobile game user acquisition in 2026 — and a framework for turning those numbers into a decision. read every figure as a directional range, not a target: benchmarks move by market, season, and source.
CPI and IPM by genre
directional 2026 ranges, drawn from public benchmarks (Business of Apps, Admiral Media, and network-reported data). your real numbers will vary — the point is to know roughly where your genre sits before you panic or celebrate.
| genre | android CPI | iOS CPI | typical IPM |
|---|---|---|---|
| hypercasual | $0.20–1.50 | $0.50–2.50 | high (25–40) |
| casual / puzzle | $1–5 | $2–7 | 8–15 |
| mid-core (rpg / strategy) | $5–20 | $8–25 | 2–5 |
| hardcore / mmo | $15–40 | $20–50 | low (1–3) |
two patterns hold almost everywhere: iOS costs meaningfully more than android for the same game, and tier-1 markets like the US run well above emerging markets. install costs have also drifted up year over year as competition for inventory rises on meta and tiktok — last year's "good" CPI is this year's average.
the retention floor under every ROAS target
D7 ROAS is what most teams scale on, but the realistic target depends entirely on your monetization model — iap-heavy and ad-funded games do different math, so set the bar against your own LTV curve, not a blog's. retention is the leading indicator underneath it. as a rough guide: D1 in the 40s and D7 in the mid-20s is strong; D7 in the mid-teens is average; D7 under ten percent is a core-loop problem no UA budget can outspend. experienced teams read retention before they pour money into scaling, because cheap installs that churn are just an expensive way to prove the product isn't ready.
the three-lever framework: which one is broken
benchmarks only matter if they change what you do monday. when a number underperforms, it's almost always one of three levers:
- creative — weak IPM or hook rate against your genre median. the ad isn't earning the click.
- audience or bid — healthy IPM but high CPI. the ad works; you're paying too much to show it, or aiming it wrong.
- product — cheap installs, low D1/D7. the funnel's fine; the game isn't holding.
isolate the weakest lever, build a 30-day test against that one thing, and re-audit against fresh benchmarks each quarter. chasing all three at once is how teams stay busy and flat.
catching the drift before it costs you
the hard part isn't the table — it's noticing a live campaign sliding off it a week before the monthly report does. numu watches CPI, IPM, ROAS, and retention continuously, tells you which lever is moving a number, and acts inside your guardrails. see it on your own account.
