The short answer
Budget an app launch from what a payer is actually worth to you, not from a cost per install benchmark, because the revenue side is far better documented than the cost side.
Read the full answer
Per RevenueCat's State of Subscription Apps 2026, built on more than 115,000 apps and over $16 billion in revenue, the median subscription app makes about $72 a month one year after launch, the top quartile sits above $429 and the top 10% above $2,574, while median realized lifetime value per payer after year one is $23 globally.
Against that, install costs in tier-one markets are high and category-dependent: per Adjust data reported in October 2026, a shopping app install cost $3.28 in the US and $3.85 in the UK and Ireland in the first half of 2026, against $0.56 for a game in 2025.
At a $23 median realized LTV per payer and a median download-to-paid rate of 2.0% by day 35, a few dollars per install does not pay back on the first subscription, which is why the paywall, the onboarding funnel and the measurement have to work before the spend scales rather than after.
Treat the first 90 days as the cost of finding out what a user is worth, and size it so that you can afford to be wrong twice.
The distribution, not the average, is the planning input. A launch budget sized for the top decile outcome is a bet; one sized against the median is a test you can afford to repeat.
Source: RevenueCat, State of Subscription Apps 2026, more than 115,000 apps · log scaleMost launch budgets are built the wrong way round. Somebody finds a cost per install figure, multiplies it by a target install count, and that becomes the marketing number in the deck.
The install number is the least reliable input available and the least important one. What a user is worth is better documented, and it decides everything the install cost does not.
How much should you budget to market an app launch?
Budget in two parts: a fixed amount to make the product monetize, and a variable amount to find out what a user costs. The first part is paywall optimization, onboarding funnel work, store listing and measurement, and it is mostly labour rather than media. The second part is paid spend, and its size should follow from the arithmetic of your own conversion rate rather than from a benchmark.
Work it out backwards from a payer. If a paying customer is worth around $23 in realized lifetime value, per RevenueCat’s State of Subscription Apps 2026, and 2% of downloads become payers by day 35, then every install is worth roughly 46 cents on the first subscription cycle.
In a tier-one market that is well below what an install costs, so the first year is funded by renewals, upsells and improvements to that conversion rate, not by the first purchase.
That arithmetic is why the fixed half comes first. Doubling download-to-paid from 2% to 4% halves your cost per subscriber without touching the media buying, and no bidding strategy delivers the same effect.
Set the paid number as the cost of a decision rather than a growth target. Enough spend on one audience and one creative angle to get a readable cohort, repeated two or three times, is a launch budget. An install count multiplied by somebody else’s CPI is a forecast you have no way to hold anyone to.
One limit on all of this. The revenue figures below come from subscription apps, so an app monetized by advertising or a one-time purchase needs its own payer arithmetic before any of it transfers.
Why can’t you budget from a published cost per install benchmark?
Published cost per install benchmarks cannot carry a launch budget because install cost moves with category, platform, country, season and creative quality, so a blended figure describes nobody in particular. The most rigorous published sets also skew towards games, which are the cheapest installs in the market and the least like a subscription app.
Look at how wide the real spread is within one source. Per Adjust data reported by PPC Land on 28 October 2026, a shopping app install cost $3.28 in the US and $3.85 in the UK and Ireland in the first half of 2026, while a game install cost $0.56 in 2025. Same measurement house, same broad method, nearly six times the difference by category.
The direction of travel is upward, which matters for a plan written on last year’s figures.
Per AppsFlyer’s 2026 Performance Index, reported in that same PPC Land piece, 9 of the top 10 and 17 of the top 20 media sources grew their year-over-year ad spend, which AppsFlyer puts down to two things: more attributed conversions and higher media costs.
That index covers 27.2 billion installs from nearly 25,000 apps across 57 media sources, according to the same report.
So treat any CPI figure you find as a sanity check on your own test, not as an input to a budget. The only install cost that belongs in your model is the one your own campaign produced, in your category, in your market, against your creative.
What does a new subscription app actually earn in its first year?
A new subscription app earns far less in year one than most launch plans assume, and the distribution matters more than the average.
Per RevenueCat’s State of Subscription Apps 2026, built on more than 115,000 apps, over $16 billion in revenue and more than a billion transactions, the median app makes about $72 a month one year after launch, the top quartile starts above $429 and the top 10% above $2,574.
The milestone data says the same thing in a harder form. Per that report, 17.3% of new apps reach $1,000 a month within two years and 4.6% reach $10,000 a month, so the outcome most plans are written for is roughly a one-in-twenty event.
Growth is concentrated too. Per that report, median year-over-year monthly recurring revenue growth is 5.3% while the top 10% of apps grew 306% or more, which means a median app does not grow its way out of a monetization problem.
Read all of this as a budgeting instruction rather than a discouragement. A launch budget that only works if the app lands in the top decile is a bet on the least likely outcome; one sized against the median is a test you can afford to run again with what you learned. Our piece on how to price a subscription app covers the lever that moves these numbers most.
What should the first 90 days of paid spend actually buy?
The first 90 days of paid spend should buy answers to four questions, in order, and install volume is not one of them. Which audience responds, which creative angle holds attention, what an install costs in your category and market, and what share of those installs become trials and then payers.
Sequence the spend so each answer informs the next. A small, concentrated test on one platform produces a readable signal faster than the same money split across Meta, TikTok and Apple Ads at once, because thin data in three places answers nothing in any of them.
Judge on the right day. Per RevenueCat’s State of Subscription Apps 2026, download-to-paid conversion is measured at day 35, so a campaign assessed on week-one installs is being graded on the cheapest available signal rather than the one that pays the bills. Plan for a read on D7 and D30 ROAS and accept that the first cohort arrives late.
Expect to spend some of the budget proving a negative. A clear answer that this audience does not convert is worth paying for, provided it is cheap and arrives early, and that is what separates a test budget from a launch campaign. Our piece on why app installs do not convert into trials covers where that break usually sits.
How much should you spend on creative testing versus media?
Spend enough on creative testing that the media buying has something to work with, because on paid social the creative does most of the audience-finding. In practice that means a steady supply of new angles rather than one hero asset with a large budget behind it, and the production cost of that supply is a budget line rather than an afterthought.
The ratio shifts with stage. At launch, with no proven angle, the share going into making and testing creative is at its highest, and it falls as winners emerge and the money moves behind them. A brand that spends everything on media in month one usually spends month two wondering why performance decayed.
Creator-led acquisition is the common source of that supply for consumer apps, because creators produce volume and variety that a studio shoot does not.
We have run creator campaigns at that scale on mobile titles: on the Genshin Impact campaign, five TikTok waves delivered 4.7 million views on live posts with nine of ten videos passing 100,000.
Note what that case study is and is not: a creator campaign for a game, not an app launch engagement with monetization attached.
| Budget layer | What it buys | What it costs you to skip | When to spend it |
|---|---|---|---|
| Paywall and onboarding funnel | A higher share of installs becoming payers | Every install costs more, permanently | Before launch |
| Store listing and ASO | Conversion on traffic you already pay for | A tax on all paid media | Before launch |
| Measurement, MMP and SKAN setup | The ability to read any of the above | You cannot tell which test worked | Before launch |
| Creative production and testing | Angles for the media to find | Media spend decays with nothing to replace it | Continuous |
| Paid media | Install volume and a read on cost | No data on what a user costs | After the first three are done |
What has to be fixed before you spend anything on user acquisition?
Before the first dollar of user acquisition, fix the three things that change what every install is worth: the paywall, the onboarding funnel and the measurement. All three are cheaper than media and all three multiply it.
The paywall decision is the largest single lever in the data. Per RevenueCat’s State of Subscription Apps 2026, median download-to-paid conversion by day 35 is 2.0% overall, but 10.7% for apps using a hard paywall against 2.1% for freemium, with the top quartile of hard paywall apps above 20%.
Those models suit different products and the choice is not free, which our comparison of freemium, free trial and hard paywall works through.
Measurement comes before spend rather than after it, because retrofitting an event spec onto a live campaign means throwing away the first cohort. Server-side conversions, an MMP, a sane SKAN configuration and agreed definitions of install-to-trial and trial-to-paid are the minimum for reading a result at all.
This is the seam most launches fall into, and it is the reason we take both halves. Acquisition and monetization are one experiment: the promise the ad makes and the promise the paywall makes have to be the same promise, or the install arrives primed for a product it does not meet.
Our app launch work covers creators, UGC and paid social alongside pricing, paywall, onboarding, store optimization and measurement, with the order set by where the launch is actually leaking.
So what is the short version?
The short version is to budget from the payer, not from the install. Per RevenueCat’s State of Subscription Apps 2026, median realized lifetime value per paying customer after year one is $23 and the median app makes about $72 a month twelve months in, so the arithmetic has to work at those numbers before it works at anybody’s growth-case numbers.
Do not plan from a published CPI figure. Per Adjust data reported in October 2026, a US shopping app install cost $3.28 in the first half of 2026 against $0.56 for a game in 2025, a spread of nearly six times inside one measurement house, so the only install cost worth modelling is the one your own test produces.
Spend the fixed half first. The paywall, the onboarding funnel, the store listing and the measurement are cheaper than media and they change what every install is worth: per that same RevenueCat report, hard paywall apps convert at a 10.7% median by day 35 against 2.1% for freemium.
Then buy answers rather than installs. One audience, one angle, enough spend for a readable cohort, judged on D30 rather than week one, repeated two or three times with what you learned. A launch budget that only pays back on a top-decile outcome, which per that report is roughly a one-in-twenty event, is a bet rather than a plan.
Frequently asked
How much does it cost to market an app in the first month?
Enough to get a clean read on one audience and one creative angle, which is usually a few thousand dollars rather than a token test.
The number follows your install cost and your conversion rate: per Adjust data reported in October 2026, a US shopping app install cost $3.28 in the first half of 2026, and at a median day-35 download-to-paid rate of 2.0% per RevenueCat's State of Subscription Apps 2026 you need hundreds of installs before a paid conversion rate means anything.
Is CPI or CAC the right number to budget against?
Cost per acquisition of a paying customer, because cost per install tells you nothing about monetization. An app with a $2 CPI and a 1% download-to-paid rate is paying $200 per subscriber; an app with a $5 CPI and a 10% rate is paying $50.
Per RevenueCat's State of Subscription Apps 2026, hard paywall apps convert at a 10.7% median by day 35 against 2.1% for freemium, so the paywall model moves CAC more than the media buying does.
Should you spend on ASO or paid ads first for a launch?
Do the store listing work first, because every paid install lands on it and a weak listing taxes all of your media. App Store and Google Play optimization is cheap relative to media and it compounds, while paid spend stops working the day you stop paying.
Our guide to what ASO costs covers the market rates, and our piece on ASO for an app with no reviews covers the launch version of the problem.
How long should a launch test run before you judge it?
Long enough to see a paid conversion cohort rather than installs, which in practice means weeks rather than days.
Per RevenueCat's State of Subscription Apps 2026, download-to-paid is measured at day 35, and trial-to-paid varies sharply by trial length, so judging a campaign on install volume in week one reads the cheapest signal rather than the one that matters.
Can creators replace paid media for an app launch?
Creators can supply the demand and the creative, but the measurement problem does not go away. Organic creator posts drive installs you cannot attribute cleanly, so brands usually run creator content as ads too, which is where the attribution lives.
Our piece on measuring creator and influencer campaigns under SKAN covers how to read those numbers.
What is a realistic first-year revenue target for a new subscription app?
Lower than most founders plan for. Per RevenueCat's State of Subscription Apps 2026, the median subscription app makes about $72 a month one year after launch, 17.3% of new apps reach $1,000 a month within two years and 4.6% reach $10,000 a month.
A plan whose first year assumes the top decile is not a plan, and the budget should be sized so a median outcome is survivable.



Affiliate Marketing vs UGC: Which Should a Consumer Brand Spend On?
Should You Hire an App Growth Consultant, an Agency, or Do It In-House?
Why Aren't My App Installs Converting Into Trials?