Mobile App LTV: How to Calculate, Measure and Grow Lifetime Value

Mobile app LTV (lifetime value) is the total net revenue an app earns from a user, or a cohort of users, over the whole time they stay active. For subscription apps that means renewals minus refunds and platform fees. For ad-monetized and hybrid apps it also includes in-app advertising revenue. LTV:CAC is the ratio between that value and the cost of acquiring the user, and cohort LTV is lifetime value measured for a group of users who installed in the same period, from the same source.

LTV is the number that decides how much an app can afford to pay for a user. Get it wrong on the high side and every campaign looks profitable until the cash runs out. Get it wrong on the low side and the app underbids, loses auctions to competitors that understand their own users better, and stalls. This guide explains how Admiral Media calculates mobile app LTV, how to read it early enough to make bidding decisions, how to connect it to ad platform algorithms, and which levers actually move it. It draws on Admiral Media’s work managing more than €500M in mobile ad spend for over 150 mobile brands, published case study results, and current third-party benchmarks.

What Mobile App LTV Actually Measures

Mobile app LTV measures the net cash a user generates for the business, not the gross price of what they bought. The useful version of the metric is always tied to a cohort, a time window, and a definition of revenue that everyone on the team agrees on.

In Admiral Media’s work across subscription, fintech, dating, health, and education apps, the first step in any account audit is to pin down four things:

• The revenue definition: gross or net of platform fees, taxes, refunds, and chargebacks. Admiral Media always recommends net revenue for bidding decisions, because that is the cash that pays for the next campaign.
• The revenue sources: subscriptions, one-off in-app purchases, ad revenue, and web purchases that happen outside the app stores. Web-to-app flows are growing, and a user who pays on the web is still an app user.
• The time horizon: Day 7, Day 30, Day 90, Day 180, Day 365, or a modeled lifetime. A “lifetime” with no horizon is a guess.
• The cohort key: install date, acquisition channel, campaign, platform, country, and, where the data allows, creative concept.

Realized LTV versus predicted LTV

Realized LTV is the revenue a cohort has actually generated to date. Predicted LTV (pLTV) is an estimate of what the cohort will generate by a later horizon, built from early signals. Both matter, and they answer different questions.

Realized LTV is accurate but slow: a Day 180 figure only exists six months after install, far too late to change a campaign spending today. Predicted LTV is available within days but carries forecast error that has to be monitored. Admiral Media covers the bidding side in its guide to predictive LTV bidding. This article focuses on how to calculate, validate, and grow the LTV those models are built on.

Why LTV Decides How Fast an App Can Grow

LTV sets the ceiling on what an app can pay per user, and the maximum affordable cost per user decides how much of each ad auction the app can win. An app with a higher, well-measured LTV can outbid competitors for the same users and still be profitable.

Paid user acquisition is an auction economy. On Meta, Google, TikTok, Apple Ads, and programmatic networks, the app that can bid more for a given user usually wins that user. Bidding more is only sustainable if the user is worth more. That is why the strongest growth teams spend as much energy on LTV as on creative or targeting: every improvement in value per user turns directly into more auction headroom.

There are three practical consequences.

• LTV caps your CPI and CPA targets. If a cohort’s net Day 180 LTV is lower than what you paid for it, the campaign lost money, regardless of how cheap the installs looked.
• LTV decides payback speed. The same LTV earned in three months funds far more growth than the same LTV earned over two years, because the cash can be reinvested sooner. Admiral Media’s guide to the CAC payback period for mobile apps covers how to set payback targets that match your runway.
• LTV differences between segments tell you where to spend. An average LTV hides the fact that some channels, platforms, and countries deliver users who are worth several times more than others. Budget should follow those differences.

The Inshallah case shows the third point clearly. When Admiral Media analyzed Inshallah’s user base, iOS users generated significantly higher revenue and nearly twice as high retention rates as Android users. Admiral Media shifted spend from Android to iOS, rebuilt the SKAN conversion value schema to track iOS revenue, and moved campaigns to purchase optimization in September 2023. In Admiral Media’s work with Inshallah, US iOS subscriptions grew by 824% and US iOS revenue grew by 1,253% since Admiral Media took over the account. The lever was not a cheaper install. It was buying more of the users who were worth more.

How to Calculate Mobile App LTV: Four Methods

There are four common ways to calculate mobile app LTV: the simple ARPU method, the churn-based formula, cohort-based realized LTV, and predictive modeling. Each trades speed for accuracy, and most apps should use at least two of them side by side.

Method 1: Simple ARPU multiplied by lifetime

The simplest formula is LTV = ARPU × average user lifetime, where ARPU is average revenue per user over a period (usually a month) and lifetime is how many periods the average user stays. It is quick and easy to explain to a board, which is why it appears in so many pitch decks.

Its weakness is that “average lifetime” is hard to know for a young app and easy to overstate for any app. Retention curves in mobile are steep at the start and flatten later. An average hides that shape. Use this method for back-of-the-envelope planning only, never as a bidding target.

Method 2: Churn-based LTV for subscription apps

For subscription apps with steady monthly churn, a common approximation is LTV = ARPPU ÷ churn rate, where ARPPU is average revenue per paying user per billing period and churn rate is the share of subscribers who cancel each period. If you want to include gross margin, multiply ARPPU by your margin after platform fees first.

The formula assumes churn is constant, which it rarely is. Subscription churn is front-loaded: most cancellations happen at the first renewal, and survivors are much stickier. RevenueCat’s 2026 renewal benchmarks put the median first renewal rate for monthly plans in Health and Fitness at 57%, which means 43% of those monthly subscribers are gone after one billing cycle. Plugging a single average churn rate into the formula would understate that first-month loss and overstate the value of a new subscriber. Use the churn formula for mature cohorts where the curve has flattened, and use cohort data for everything else. For more on the mechanics of churn itself, see Admiral Media’s guide to app churn rate.

Method 3: Cohort-based realized LTV

Cohort-based LTV is the cumulative net revenue per install for users grouped by install date and source, tracked at fixed checkpoints such as Day 7, Day 30, Day 90, and Day 180. It is the most honest method because it uses revenue that has actually been collected, not assumed.

The output is a curve for each cohort. Plotting cohorts on top of each other shows whether newer users are worth more or less than older ones at the same age, which is the fastest way to see whether product changes, pricing changes, or new channels are helping. Admiral Media builds these curves per channel and per platform as standard, because a blended curve can look stable while one channel is quietly deteriorating.

Method 4: Predictive LTV modeling

Predictive LTV uses early behavior (trial starts, first purchases, session depth, feature usage, onboarding completion) to forecast a cohort’s value at a later horizon. Models range from simple ratios, such as “Day 180 LTV has historically been a stable multiple of Day 7 LTV for this channel”, to machine learning models that score each user individually.

Predictive LTV is what makes value-based bidding possible, because ad platforms need a value signal within days, not months. The risk is that a model trained on last year’s users will mis-forecast when pricing, onboarding, or the channel mix changes. Every predictive model needs a standing back-test against realized LTV.

A Framework That Turns LTV Into Bids

The Admiral Media Cohort-to-Bid LTV Framework is a seven-step process that turns raw revenue data into bid targets an ad platform can act on. It exists because most apps have an LTV number, but very few have a chain that connects that number to what their campaigns bid every day.

The Admiral Media team developed the framework from the recurring gaps it finds in account audits: LTV defined differently by finance and UA, cohorts that are not split by channel, early signals that do not predict anything, and bid targets set once and never revisited. Each step closes one of those gaps.

The Admiral Media Cohort-to-Bid LTV Framework

1. Define net value. Agree on one revenue definition across finance, product, and UA: net of platform fees, taxes, refunds, and chargebacks, including web purchases and ad revenue where they exist. Write it down. Every later step depends on it.
2. Cohort by source. Group users by install week, channel, campaign type, platform, and country. Keep iOS and Android separate, because their measurement, monetization, and retention behave differently.
3. Read the early windows. Track value at Day 0 to 2, Day 3 to 7, and Day 8 to 35. These checkpoints match the three AdAttributionKit conversion windows on iOS, so the same milestones work on both platforms and your iOS measurement is not an afterthought.
4. Find the predictive signal. Test which early events best predict Day 90 or Day 180 value: trial start, first payment, onboarding completion, a key feature used in the first session. Keep the one or two events with the strongest and most stable relationship to later revenue.
5. Set LTV guardrails. Convert predicted LTV into a maximum CPA or a target ROAS for each channel and platform, using your payback target. A business that needs cash back in three months sets tighter targets than one funded for a two-year payback.
6. Feed value into bidding. Send the predictive event, or a value-bearing purchase event, to each ad platform and switch to value-based bidding once volume allows. On iOS, encode the same signal in your conversion value schema.
7. Back-test monthly. Compare predicted against realized LTV for every cohort that has reached its horizon. When the gap widens, retrain the model or reset the targets before the error compounds into overspend.

Proof: What LTV-Led Acquisition Looks Like in Admiral Media Accounts

LTV-led acquisition shows up in results as revenue and ROAS growing faster than install volume. When the value per user rises, the same budget produces more revenue even if install growth is moderate.

The NeuroNation account is a good illustration. Admiral Media ran intensive creative testing and channel exploration for the brain-training app, evaluating every test with its pRank methodology, which ranks creatives and campaigns against the client’s target KPIs rather than against install volume alone. In Admiral Media’s work with NeuroNation, from January to August 2019, the account delivered:

• +117% ROAS: return on ad spend more than doubled, the clearest sign that value per user rose along with volume.
• +66% installs: volume grew while efficiency improved, rather than at its expense.
• +42% net cohort revenue: the revenue each cohort generated, net, rose substantially.
• +32% purchases: more users converted to paying customers.
• -39% CPI: the cost of each install fell at the same time.

ChatPDF shows the same pattern in a subscription AI product. Admiral Media restructured the ChatPDF accounts, moved them to value-based bidding, and ran a weekly creative testing cadence with three variants per winning concept, during what was ChatPDF’s low acquisition season. In Admiral Media’s work with ChatPDF, comparing Year 1 against Year 2 year to date on an indexed baseline, results were:

• +320% ROAS overall: with Google Ads also at 320% ROAS growth and Meta at 280%.
• +156% subscriptions overall: +142% on Google Ads and +171% on Meta.
• -42% CAC overall: -38% on Google Ads and -45% on Meta.

The mechanism is the one the Cohort-to-Bid framework describes. Value-based bidding tells the algorithm which users are worth more, the algorithm shifts delivery toward them, and ROAS rises faster than subscriptions because the added subscribers skew toward higher value.

PURE, the dating app, shows the channel side of LTV. Admiral Media tested a programmatic DSP against an established self-attributing network for US Android campaigns, with separate budgets and ads tailored to each platform. The DSP delivered a $2.44 CPI against $9.43 on the self-attributing network and exceeded the app’s D7 ROAS goals, which made further market launches possible, according to Admiral Media’s PURE case study. A cheaper install only matters because the D7 ROAS held up: the users bought through the new channel were still worth enough.

Reading LTV Early: Signals That Predict Value

The most useful early LTV signals are the events that happen within the first days after install and correlate strongly with revenue months later. For most subscription apps, those are trial start, trial-to-paid conversion, and the plan chosen. For ad-monetized apps, they are session depth and early engagement.

Early reading matters because platforms learn fast and budgets move daily. A UA team that waits 90 days to judge a campaign has already spent three months of budget on it. Admiral Media’s standard approach is to find a signal that shows up within the first week and holds a stable relationship with Day 90 or Day 180 value.

Paywall model changes the early curve

The paywall model is one of the biggest structural drivers of how early an app can read value. RevenueCat’s State of Subscription Apps 2026 reports that hard paywall apps generate $3.09 revenue per install at Day 60, compared with $0.38 for freemium apps. Hard paywalls also show a median Day 35 trial-to-paid conversion of 10.7%, against 2.1% for freemium.

This does not mean every app should switch to a hard paywall. RevenueCat’s same report finds that year-one retention of yearly subscribers is statistically similar across the two models, at 28% for freemium and 27% for hard paywall apps. What it does mean for UA is that a hard paywall app can read value much earlier, because revenue arrives sooner. A freemium app has to rely more heavily on behavioral proxies, such as feature usage and session frequency, to predict value in the first week, and should expect its early LTV reads to be noisier.

Front-loaded cancellations

Cancellations in subscription apps cluster right at the start. RevenueCat reports that 55% of all trial cancellations happen on Day 0, and that the first month accounts for 35% of all annual subscription cancellations. For LTV modeling, that is useful information: a large share of the eventual churn is visible in the first days, so an early “trial started and not cancelled by Day 1” signal is often more predictive than a raw trial start.

Category and plan benchmarks for the retention side of LTV

Retention by plan type is a core input to any subscription LTV model. The table below shows RevenueCat’s median first renewal rates for a selection of categories that Admiral Media works in frequently.

Source: RevenueCat, Average subscription renewal rates by app category (2026 benchmarks), last updated April 2026. These are medians across RevenueCat’s dataset. Your own app’s cohorts are the only numbers that should drive bids.

Annual plans renew at lower rates than monthly plans in every category shown, but an annual subscriber has already paid for a full year before that decision. When Admiral Media models LTV by plan, it compares cumulative net revenue per subscriber at the same age, not renewal rates in isolation.

Measuring LTV on iOS Under AdAttributionKit

On iOS, the ad-side view of LTV is limited by Apple’s privacy-preserving attribution, so the ad network sees conversion values within fixed windows rather than a user’s full revenue history. The practical answer is to design the conversion value schema around the early signals that predict LTV.

According to Apple’s AdAttributionKit documentation, AdAttributionKit supports up to three postbacks for a winning ad attribution, covering Day 0 to 2, Day 3 to 7, and Day 8 to 35 after the first launch. A fine-grained conversion value is only available in the first postback, and only for higher postback data tiers. The second and third postbacks carry a coarse value. Postbacks arrive after a random delay: 24 to 48 hours for the first, and 24 to 144 hours for the second and third.

That has three consequences for LTV measurement:

• The first postback carries most of the detail. The Day 0 to 2 window is where your fine-grained value lives, so the events you encode there should be the strongest early predictors of LTV, not just “app opened”.
• Later windows are coarse. The Day 3 to 7 and Day 8 to 35 postbacks can tell you whether a user moved into a low, medium, or high value band, which is enough to confirm trial conversions and first renewals if the bands are designed around them.
• Delays affect optimization speed. Postback timing means iOS campaigns always react slower than Android. Budget decisions on iOS should use rolling windows, not daily snapshots.

The Inshallah account is an example of this in practice. Part of Admiral Media’s work was developing and refining the SKAN conversion value schema to track iOS revenue, which allowed purchase-optimized campaigns to work on a platform where users were worth more. Admiral Media’s Inshallah case study documents the resulting US iOS growth described earlier.

Connecting LTV to Bidding: From Number to Target

LTV only changes results when it becomes a bid target. The two standard routes are a maximum CPA derived from LTV and a payback target, or a target ROAS that tells the platform how much value it must return per unit of spend.

The logic is simple. If your predicted Day 180 net LTV for a channel is known, and your business needs to recover acquisition cost within a set window, the maximum you can pay for a user is the portion of LTV that arrives within that window, minus any margin you want to keep. That number becomes the ceiling for tCPA. For value-based bidding, the equivalent is a target ROAS that reflects the early-window value you expect relative to the full-horizon LTV.

What the platforms need

Value-based bidding needs enough conversion data to learn. Google’s guidance for App campaigns is explicit. Google Ads Help recommends moving to Target ROAS once a campaign reaches stable volume, typically 30 or more conversions in 30 days, and, when migrating from Target CPA, having at least 3 to 4 weeks of historical conversion value data. The recommended starting tROAS is the achieved ROAS (conversion value divided by cost) of a comparable campaign for the same app. Google also advises keeping target changes within a 15 to 20% range and not making major changes more often than every 1 to 2 weeks, because each change can push the campaign back into a learning phase of 7 to 14 days.

Those thresholds explain why LTV work and campaign structure are linked. In Admiral Media’s accounts, consolidating campaigns so each one clears the data threshold is often the first structural change, before any target is adjusted.

LTV-based guardrails by growth stage

Levers That Increase Mobile App LTV

LTV grows through four levers: acquiring higher-value users, converting more of them to paying customers, keeping them longer, and earning more per paying user. UA teams control the first lever directly and influence the other three through the feedback they give product and monetization teams.

1. Acquire higher-value users

Not every user is worth the same, and the acquisition source is one of the strongest predictors of value. Channel, platform, country, and creative angle all shape who installs. The Inshallah platform shift and the ChatPDF move to value-based bidding are both examples of increasing LTV by changing who you buy rather than what the product does. Creative matters here too: an ad that promises something the product does not deliver produces installs that churn on Day 0, which lowers LTV even when CPI looks good.

2. Convert more users to paying

Conversion from install to first payment is where most value is lost. Paywall placement, trial length, and onboarding all sit here. RevenueCat reports that trials of 17 to 32 days convert at 42.5%, compared with 25.5% for trials shorter than four days. Trial length is a product decision, but UA teams should know it, because a change in trial length changes the timing of revenue and therefore the early signals that bidding relies on.

3. Keep users longer

Retention is the multiplier in almost every LTV formula. Front-loaded churn is the main problem, which is why onboarding and the first week of product experience have outsized influence on lifetime value. For subscription apps, billing failures are a hidden churn source: RevenueCat reports that 31% of subscription cancellations on Google Play are due to billing errors, compared with 14% on the App Store. That is revenue lost from users who did not choose to leave, and it is fixable with billing retry and grace period handling.

4. Earn more per paying user

Pricing, plan mix, upgrades, and hybrid monetization (combining subscriptions with in-app purchases or ads) raise revenue per payer. RevenueCat finds that AI apps sustain a 41% Year 1 realized LTV premium over non-AI apps, with a median of $30.16 versus $21.37, but also that AI monthly plans churn 36% faster than non-AI plans over 12 months. Higher value per payer and weaker retention can coexist, which is exactly why LTV has to be measured as a cohort curve rather than a single number.

Common LTV Mistakes Admiral Media Sees in Account Audits

The most common LTV mistakes are definitional and structural, not mathematical. Most apps have enough data to calculate a reasonable LTV. The problems start when the number is defined inconsistently, averaged across segments that behave differently, or never checked against reality.

• Using gross revenue for bidding. Store commission, taxes, and refunds come off the top. Bidding against gross revenue systematically overpays for users.
• Blending iOS and Android. The two platforms differ in measurement, retention, and monetization. The Inshallah retention gap between platforms is a reminder that a blended LTV can hide a large difference in value.
• Optimizing to an event that does not predict revenue. Install, registration, and even trial start can all be poor predictors when Day 0 cancellation is high. Test the event against realized LTV before you optimize to it.
• Setting targets once. Pricing changes, onboarding updates, new markets, and new creative all shift LTV. A tROAS set six months ago is probably wrong today.
• Ignoring the time value of cash. Two cohorts with the same Day 365 LTV are not equally valuable if one pays back in month two and the other in month ten. Payback timing decides how much growth the business can fund.
• Treating a model as fact. Predictive LTV is a forecast. Without a monthly back-test against realized cohorts, forecast error goes unnoticed until it shows up as a cash problem.

Frequently Asked Questions

What is a good LTV for a mobile app?

There is no universal good LTV for a mobile app, because the number only means something relative to what the app pays to acquire users and how quickly that cost needs to come back. A useful LTV is one that is higher than your net cost per acquired user within your chosen payback window, measured per channel and platform. Category benchmarks, such as RevenueCat’s median renewal rates, help sense-check a model but should not replace your own cohort data. Admiral Media sets LTV targets per account based on realized cohort curves and the client’s payback requirements.

How do you calculate LTV for a subscription app?

The quick formula for a subscription app is average revenue per paying user, net of platform fees, divided by the periodic churn rate. That formula works reasonably for mature cohorts with steady churn but overstates LTV for new subscribers, because subscription churn is concentrated at the first renewal. The more reliable method is cohort-based: track cumulative net revenue per install for users grouped by install date and source at Day 7, 30, 90, and 180. Use the cohort curves to validate any predictive model before using it for bidding.

What is the difference between LTV and pLTV?

LTV usually refers to realized lifetime value, the revenue a cohort has actually generated so far. pLTV, or predicted lifetime value, is a forecast of what a cohort or individual user will generate by a future horizon, based on early behavior such as trial starts or first purchases. Realized LTV is accurate but slow to arrive, while pLTV is available within days but carries forecast error. Most app growth teams use pLTV to set bids and realized LTV to check whether those bids were right.

What is a good LTV to CAC ratio for mobile apps?

The right LTV to CAC ratio depends on the app’s margin, cash position, and how quickly it needs payback, so there is no single correct figure. What matters more than the ratio itself is that LTV and CAC are measured on the same basis, net revenue against fully loaded acquisition cost, and over the same time horizon. A high ratio that takes two years to realize can be riskier than a lower ratio that pays back in a few months. Admiral Media recommends pairing the ratio with a payback period target per channel.

How early can you predict mobile app LTV?

Many subscription apps can build a useful LTV prediction within the first week after install, provided they have historical cohorts to learn from. The best early signals are events that correlate strongly with later revenue, such as a trial that was not cancelled by Day 1, a first payment, or completion of a key onboarding step. On iOS, the first AdAttributionKit window covers Day 0 to 2 and is the only one with fine-grained conversion values, so the strongest early signal should be encoded there. Every early predictor should be back-tested against realized value before it drives bidding.

Does iOS or Android have higher LTV?

It depends on the app, the category, and the market, so each app should measure it for itself rather than assume. In Admiral Media’s work with the dating app Inshallah, iOS users generated significantly higher revenue and nearly twice the retention of Android users, which justified shifting budget toward iOS. Other apps see different results, particularly in markets where Android dominates. The safe practice is to calculate LTV separately for each platform and never bid against a blended figure.