How Much Does It Cost to Acquire an App User?
Ask what an app user costs and you'll get a CPI range. Ask what a valuable app user costs and the honest answer is: it depends on what a user is worth to you. This guide gives you both — the market price ranges and the math that tells you what you can actually afford to pay.
What the market charges
Install prices move with vertical, GEO tier, platform and format, but the structure is stable. Casual games and utilities buy installs cheapest — cents in tier-3 GEOs, low single-digit dollars in tier-1. Dating and social sit mid-range. Finance, trading and casino verticals pay the most per install because the user behind it is worth the most — and there, CPI alone is a misleading lens: the metric that matters is cost per funded account, which runs orders of magnitude above cost per install.
Format shifts the price too: rewarded and playable inventory costs more per install than banners because it converts deliberate, higher-intent users. Push traffic is the cheapest test volume available. iOS typically prices above Android per install — and frequently below it per payer. Any rate card quoting one number per GEO without asking about your event economics is selling you a guess.
Compute your allowable CAC before asking prices
The only acquisition cost that matters is the one your unit economics can absorb. Work backwards from revenue: average revenue per paying user, payer conversion rate, gross margin, and your payback window — how many days or months a cohort has to return its cost. Multiply through and you get the allowable cost per acquired user; divide by the event-to-install conversion rate and you get the allowable CPI.
Two apps in the same vertical can have allowable CPIs that differ by five times, because their monetization differs. This is why 'what does an install cost' is the wrong first question and 'what is a retained, paying user worth to us' is the right one. A buying partner who starts with the first question is selling traffic; one who starts with the second is building a campaign.
The costs hidden inside the number
The quoted CPI is never the full cost of a user. Fraud tax: on unfiltered sources, a meaningful share of paid installs are emulators, farms or misattributed organics — your effective CPI on real users is proportionally higher. Churn tax: an install cohort that retains at half the expected rate doubles your cost per retained user. Measurement tax: if attribution is broken, you're optimizing blind and the waste doesn't appear on any invoice.
Add creative production, tooling, and the team's time. Then compare against channels honestly: a 'cheap' source delivering bots costs more than an 'expensive' one delivering payers. We wrote a full field guide to reading these signals in our traffic quality guide.
What actually lowers your cost per valuable user
Four levers move the real number. Better creative: a stronger hook raises install rate at the same impression price, directly cutting CPI — creative iteration is the highest-ROI cost reduction in UA. Better targeting of value, not volume: optimizing toward the billable event makes the algorithm hunt future payers instead of cheap installs. Better sources: filtered, placement-scored inventory removes the fraud tax at the root. Better funnels: store page conversion and onboarding completion multiply every upstream dollar.
The one lever that doesn't work long-term is simply bidding less. Underpriced bids buy leftover inventory and worse placements; the savings evaporate in quality. Cost control in app acquisition is earned through conversion quality, not through thrift at the auction.
Planning a test budget that produces answers
A test budget's job is to buy a decision, not to buy users. Size it from event frequency: if your billable event converts at a few percent of installs, you need enough installs for the event count to be statistically readable per source — otherwise you're reading noise. Vertically this ranges from modest budgets for fast-funnel utilities to serious commitments for deposit-driven finance apps.
The planning conversation should end with three numbers: the test spend, the measurement window, and the target economics that decide scale-or-kill. If you want those numbers computed for your app before spending anything, register as an advertiser — the offer audit is where every Mobadly campaign starts.
Benchmarks: use ranges, not points
Any benchmark table you find online — ours included — is a starting range, not a promise. Vertical averages hide the variance that decides your outcome: the same vertical shows three-fold CPI differences between a polished funnel with strong creative and a raw store link with a generic video. Benchmarks are useful for exactly one thing: sanity-checking that your quote isn't disconnected from the market. Beyond that, your own cohort data replaces benchmarks within the first measurement window, and should.
The benchmark that actually matters is internal and weekly: your cost per valuable user this week versus last week, per source. Trends you own beat industry points you don't. Build that series from your first campaign and you'll never need to ask a forum what installs cost again — you'll know what they cost you, which is the only version of the question with a usable answer.