Freemium

Freemium is the standard mobile-app business model in 2026. A user can download and use the app for free at some baseline level. Revenue comes from the fraction of users — typically 2-8% in healthy apps — who convert to a paid tier (subscriptionGlossarySubscriptionA recurring in-app purchase that charges the user on a fixed cadence (weekly, monthly, annual) until they cancel., one-time upgrade, in-app content packs). The free tier needs to be useful enough to build habit but incomplete enough that power users feel pulled to upgrade. Getting that balance wrong is the most common reason freemium apps fail to monetize.

The model arithmetic is simple but unforgiving: paying-user share × paying ARPPUGlossaryARPPU (Average Revenue Per Paying User)Total revenue divided by the number of users who paid at least once in the period — the monetization-per-payer metric, excluding free-tier dilution. = blended ARPUGlossaryARPU (Average Revenue Per User)Total revenue in a period divided by the user base for that period — the blended revenue-per-user metric, including non-paying users.. To hit a $5 monthly ARPU you need either 5% of users paying $100/month each, or 25% paying $20/month each — different funnels, very different products. Most healthy subscription apps land at 3-8% conversion with a paying ARPPU of $30-80, giving blended ARPU somewhere in the $1-5 range.

The MWM catalog data exposes how harsh freemium actually is. Of all free apps with measurable downloads, only ~25% earn any IAPGlossaryIn-App Purchase (IAP)Any purchase made inside an app after installation — consumables, non-consumables, or subscriptions. Processed by Apple or Google, who take a 15-30% commission. revenue at all — three out of four free apps never convert a single user. Most "freemium" apps are free in name only; they ship with monetization hooks that no one ever triggers. The ones that DO monetize collectively capture 99% of total mobile app revenue — almost all the money is in the freemium tail.

Median ARPU among monetizing freemium apps is $0.14 per installer over a 30-day window. The top decile starts at $4.50; the top 1% above $40. The spread tells you everything: freemium is a power-law game where a few apps make almost all the money, and the median freemium app is barely paying for its server costs. Aim for the top decile or accept that you're building a hobby business.

Freemium is not "free with ads" — those are different business models with different unit economics. Ad-supported monetizes every user via impressions at very low ARPU per impressionGlossaryAd ImpressionA single instance of an ad being served to a user — the atomic unit of ad inventory. What counts as "served" vs "viewable" vs "verified" matters enormously for pricing.. Freemium monetizes a small fraction at high ARPU per converted user. Many modern apps hybridize: show ads to free users, remove ads at the paid tier, layer subscription content on top. The hybrid extracts more from the long tail without killing the conversion funnel.

When to NOT pick freemium: when your value proposition is binary (a tax-prep app, a one-off utility) and a paid up-front purchase makes more sense than a continuous monetization layer. When your user acquisition cost exceeds your realistic blended ARPU × tenure — freemium amplifies that gap, paid up-front contains it. And when your free tier would require giving away the entire product — a game with no progression friction, a content app with no usage cap.

Operational tips: instrument the point of friction that triggers paywallGlossaryPaywallThe in-app screen that asks the user to subscribe or buy — the single most revenue-sensitive surface in a subscription app. conversion. Most successful freemium apps have one specific moment — a usage quota hit, a premium feature attempt, a contextual win-backGlossaryRe-EngagementMarketing activity aimed at bringing lapsed users back to the app — through push, email, retargeting ads, deep-linked notifications. — where 70%+ of their conversions happen. Find yours, A/B testGlossaryA/B Testing (Mobile App)Running controlled experiments where users are randomly assigned to variants of a feature or design, then comparing results to identify the better-performing version. it relentlessly, and don't waste paywall surface area on users who haven't reached it.