Installs are the easy part

What app marketing strategies actually work?

Effective app marketing strategies run in a fixed order: raise store conversion first, buy installs against the better rate second, and build retention loops before paid spend scales. The order matters because the listing and the first session are multipliers on everything bought downstream, while media budget multiplies nothing. Global app uninstall rates sat at 46.1 percent within 30 days in 2024, against 46.9 percent in 2023.⁵ Close to half of what most teams buy disappears inside a month.

That reframes the usual budget argument. Acquisition, store optimization and retention are not three teams competing for one pot of money. They are three terms in a single equation, and two of them are cheaper to move than the third.

How is app store optimization different from app marketing?

App store optimization is the narrower discipline: improving visibility and conversion inside store search and browse through keyword targeting, metadata, icon, screenshots, preview video, ratings and Custom Product Page testing. App marketing is the whole demand system around it, covering paid acquisition, owned channels, social discovery and lifecycle messaging. ASO is one channel. It is also the one channel that changes the price of every other channel.

The benchmarks explain why. Average store conversion runs at 8.56 percent on iOS and 16.15 percent on Google Play across categories.¹ Custom Product Pages lift conversion rate by an average of 5.9 percent, and by as much as 8.6 percent on generic campaigns.¹ A listing that converts better does not only win more organic installs. It lowers the effective cost of every paid install pointed at it.

The shift worth making is to stop treating the listing as a setup task. Icon, screenshot order, the first screenshot caption and Custom Product Pages are all testable against traffic segments the way a landing page is, and the gains compound rather than reset. Teams that already run web experimentation have the muscle for this, because the discipline transfers directly from mobile conversion rate work.

What should user acquisition cost, and when should paid start?

Cost per install sits in a wide band. Business of Apps puts the broad range at $1.50 to $5.00, with Android averaging about $1.20 against roughly $3.60 on iOS.⁶ Region widens it: roughly $0.50 to $2.00 in Latin America, $1.50 to $3.00 across APAC, $2.00 to $4.00 in EMEA and $2.50 to $5.00 in North America.⁶ Vertical widens it most, with fintech and insurance at $10 to $25 and sports apps highest at an average of $26.81.⁶

None of those numbers mean anything without a retention figure beside them. The arithmetic is blunt: a $3.00 install in a category that keeps 4 percent of users to Day 30 costs $75 per retained user, and the same $3.00 install in a category that keeps 15 percent costs $20.³⁴ Cost per install is not a performance metric on its own. It is one half of a fraction, and most teams only report the top half.

On timing, organic should lead, but not to the exclusion of paid. Apple Search Ads top of search placements convert at over 60 percent on average,² which means paid search inside the store can buy the rank and download velocity that organic position then inherits. The pattern that works is a concentrated paid push on a defined keyword set, followed by a taper once organic holds the position, with listing tests running underneath the whole time. Treating paid user acquisition and organic ASO as separate budgets owned by separate people is what breaks that handover.

Which channels carry an app launch?

App launch marketing has three jobs in the first eight weeks: reach people who already have the problem, land them on a listing that converts, and keep enough of them to hold rank. Channel choice follows from those jobs rather than from a list of what is available.

Store search is the base layer. It is intent traffic, it is where the conversion benchmarks apply, and it is the only channel where a listing improvement pays back permanently. Paid store search comes second, for the rank reasons above. Short form social video comes third, and it does a different job: a demo clip on TikTok, Reels or Shorts qualifies the viewer before they ever reach the store page, which raises the quality of arriving traffic rather than the volume.

That changes how social creative should be scored. View counts and install counts both flatter clips that overpromise, and an overpromising clip shows up later as an uninstall. Score short form creative on the Day 7 and Day 30 retention of the cohort it produced, and what survives will be the creative that showed the product honestly. Owned channels, email lists, existing web traffic and partner audiences complete the mix, and they belong in the same plan as the rest of the digital marketing programme rather than in an app-only silo.

Why do users uninstall, and what holds them?

Uninstalls cluster where the store promise and the first session diverge. Dating apps see roughly 59 percent Android uninstall rates and gaming roughly 52 percent, while news and magazine apps sit at roughly 27.35 percent.⁵ That spread is not a ranking of category quality. It tracks how quickly each app delivers the thing its listing promised.

Retention benchmarks need reading with care, because the published averages disagree. Pushwoosh puts average Day 30 retention at 3.10 percent on iOS and 2.82 percent on Android,³ while Adjust puts the all category average at approximately 6 percent.⁴ Both are credible. They count different app panels and different definitions of an active user, which is why a single retention target imposed across a portfolio produces arguments instead of decisions.

Category benchmarks are more usable. Social and productivity apps sit at the top of the range, transactional and hyper-casual categories at the bottom.

Read across the range: social apps hold 15 to 20 percent at Day 30, productivity 10 to 18 percent, fintech 10 to 15 percent, mid-core gaming around 10 percent, ecommerce 3 to 6 percent and hyper-casual gaming around 4 percent.³⁴

An ecommerce app holding 5 percent of users at Day 30 is performing. A social app holding 5 percent is failing. The number is identical and the verdict is opposite.

What holds users is a loop rather than a feature. A trigger brings the user back, an action inside the app produces a result they actually wanted, and that result creates the reason for the next trigger. Onboarding is where the loop is either established or lost, which makes first session design a user acquisition cost lever and not only a product concern.

How many push notifications are too many?

Two to five messages in a week is where the damage starts. Pushwoosh reports that 46 percent of users opt out of push after two to five messages in a single week, and a further 32 percent opt out in the six to ten range.⁷ The permission is also getting scarcer: Android push opt-in has fallen from 85 percent to 67 percent, and iOS from 58 percent to 56 percent.⁷

The engagement data points the same way as the opt-out data, which is unusual and worth acting on. Contextual campaigns triggered by user behaviour open at 14.4 percent, against 4.19 percent for generic blasts.⁸ Segmented, event triggered messaging therefore wins twice. It earns more engagement per send, and it spends less of a permission that is harder to obtain each year.

In practice: cap weekly sends per user rather than per campaign, treat the push permission as a finite budget with a running balance, and route any message that cannot name the behaviour which triggered it to in-app messaging instead.

Which metrics predict revenue rather than installs?

Install volume is the number most likely to be reported and least likely to predict anything. The measures that carry signal combine a cost with a survival rate.

None of these need new tooling. They need the attribution data already being collected to be divided by the retention data already being collected, then reported in one view rather than in two decks owned by two teams.

How should you sequence these app marketing strategies?

A workable first 90 days, ordered by payback.

1. Baseline the listing. Measure current conversion against the 8.56 percent iOS and 16.15 percent Google Play averages¹ before changing anything, so later tests have something to beat.
2. Ship two Custom Product Pages. Match them to the two clearest audience segments. Average lift is 5.9 percent and reaches 8.6 percent on generic campaigns.¹
3. Fix the first session. Remove everything standing between install and first delivered value. This is the cheapest available reduction in cost per retained user.
4. Buy rank deliberately. Run store search ads on a defined keyword set, then taper as organic position holds.
5. Build one retention loop, not five. One behaviour triggered notification tied to one repeated action, judged on Day 7 retention.
6. Report cost per retained user weekly. Retire install count as a headline number.

Discovery keeps moving upstream of the store, into social feeds and into assistants that answer a question before they link out to anything. Structuring product content so those systems can read and quote it is the same job covered in our guide to optimizing for voice and AI answer engines, and it feeds the store listing rather than competing with it. If you want a read on where your own app funnel leaks between impression and Day 30, talk to our team.

Written by the group's editorial team with the practice leads who run these builds. Reviewed before publish. Spotted an error? Tell us and we will fix it.