No plan length has the best acquisition economics for every app. Annual plans win on cash: the whole first year arrives on the day of purchase, and each paying subscriber is worth the most. Weekly plans usually win on conversion, and if they turn enough more installs into payers they can earn more per install over a year, despite RevenueCat's 2026 data showing only 1% to 2% of weekly subscriptions survive twelve months. Monthly plans sit between the two and are the most forgiving. The deciding number is not retention. It is how many more of your installs buy the shorter plan, and much of the annual plan's famous retention advantage comes from who chooses it, not from the plan itself.
This page compares the three plan lengths from the ad account. It covers what the 2026 benchmarks say, why the two biggest datasets disagree, a worked example that runs the same 1,000 paid installs through each plan, and what each plan length does to your Meta signal, attribution and cash. If you want the ROAS windows themselves in detail, our subscription app ROAS benchmarks cover how D7, D30 and D60 grow for each model.
Key takeaways
- Annual plans earn the most per paying subscriber and collect it up front, which makes early ROAS and payback simple to read.
- Shorter plans usually convert more installs. Adapty's 2026 report finds weekly plans convert 2 to 7 times better than annual plans across its price buckets.
- Per install, the winner depends on the conversion ratio between plans. In our hypothetical example a 2.3x conversion advantage is enough for weekly to beat annual over a year; at 1.8x they tie.
- Year-one retention flatters annual plans twice: they face one renewal decision instead of about fifty, and the people who choose them are self-selected for stronger need.
- Weekly and monthly cohorts earn most of their revenue after Meta's 7 day click window, so they need cohort reporting, not platform ROAS.
What the 2026 data says about each plan length
The figures below come from RevenueCat's State of Subscription Apps 2026, which covers more than 115,000 apps and more than $16 billion in revenue. Ranges are the spread of category medians, not the spread of individual apps, and every figure includes organic as well as paid users.
| Metric | Weekly | Monthly | Annual | Notes |
|---|---|---|---|---|
| Typical price architecture | $4.99 to $6.99 | $7.99 to $9.99 | $29.99 to $39.99 | RevenueCat's description of the dominant pricing across geographies. |
| First renewal rate (median range by category) | 35% to 58% | 53% to 61% | 23% to 40% | Share of subscriptions that renew at the first billing cycle. |
| Third renewal rate (median range by category) | 77% to 81% | 73% to 82% | 56% to 70% | Survivors of the first renewal are far more stable on every plan. |
| Retained after one year (median range by category) | 1% to 2% | 6% to 14% | 20% to 40% | Share of paid subscriptions still active after a year. 2024 cohort medians: 1.2%, 8%, 28%. |
| Churned subscribers who come back within a year | 9% | 20% | 5% | Overall reactivation rates. |
| Revenue per install, day 14 and day 60 (median) | $0.19 / $0.32 | $0.18 / $0.29 | $0.36 / $0.46 | Apps grouped by their most-sold plan; all of each app's revenue counted. Chart figures, see the note below. |
Two things in the report need flagging before you use it. The revenue per install row uses the chart figures, which match the growth percentages in the report's own commentary (+68% for weekly-dominant apps and +61% for monthly between day 14 and day 60). A summary elsewhere in the same report quotes much lower weekly figures, $0.07 at day 14 and $0.09 at day 60, and a lower monthly day 60 figure of $0.24. Only the yearly figures agree in both places. The report also publishes two charts of plan mix by category whose category labels contradict each other, so we have not quoted category-level plan shares except Gaming, where both charts agree that 82% of sales are weekly.
The trend is clearer. Daphne Tideman, writing in the report, notes that annual plans fell from 41.4% of subscription durations in the previous report to 33.6% in this one, with monthly and, in some categories, weekly plans taking the share. Prices barely moved: RevenueCat puts the median yearly price at $34.80, up from $31.60, with weekly and monthly flat.
Why RevenueCat and Adapty disagree about which plan earns the most
Read the two big 2026 reports side by side and they appear to describe different markets. Adapty's State of In-App Subscriptions 2026 says weekly plans generated 55.6% of all app revenue in its dataset in 2025, up from 43.3% two years earlier, and that weekly plans with trials bring in 1.5 times the average one-year LTV of every other paywall setup. RevenueCat says monthly subscriptions dominate overall revenue share and that weekly plans rarely exceed 30% of a category's revenue outside Gaming.
Both can be accurate. Each report describes the apps that use that vendor, and the customer bases differ. Adapty covers about 16,000 apps and $3 billion in revenue, reports gross revenue, and counts one-time purchases in its revenue-share denominator. Its category list does not include Gaming. RevenueCat's dataset is roughly seven times larger by app count and includes Gaming, whose 82% weekly mix still produces a small revenue share because of low prices. Neither is the market, and neither tells you what your app should sell. Revenue share also ignores what it cost to acquire the revenue, which is the question this page is about.
Why year-one retention flatters annual plans
The retention row in the table is the number most often used to argue for annual plans, and it compares unlike things. A yearly subscriber retained after one year has survived one renewal decision. A monthly subscriber has survived twelve, and a weekly subscriber about fifty-two. The same underlying satisfaction will always produce lower twelve-month retention on a shorter plan, because there are more doors out.
The second problem is selection. Nobody is randomly assigned a plan length. People who pay for a year up front have already told you they expect to need the product for a long time. Alice Muir Kocourková, writing in RevenueCat's report, puts it directly: yearly plans retain better "not because users are inherently more loyal, but because higher upfront commitment filters for stronger need." The same applies in reverse to Adapty's weekly-with-trial LTV figure: the people who start a trial and then keep paying weekly are a filtered group too.
The practical consequence matters more than the statistics. If you push more people onto the annual plan, by defaulting it, hiding the monthly option or deepening the discount, the extra annual buyers are the least committed ones. They will renew below your current annual rate. RevenueCat's own pricing data hints at the effect: low-priced yearly plans retain 36% at the median after a year against 23% for high-priced ones. A cheaper annual plan attracts a different mix of buyers, not just a cheaper version of the same ones.
Worked example: the same 1,000 installs, sold three ways
The example below is hypothetical. It is three versions of one app, each selling only one plan, so the only thing that changes is plan length. Every assumption is set inside RevenueCat's 2026 median ranges, and none of it is client data.
- Media: 1,000 paid installs at £1.50 each, so £1,500 of spend. Every purchase is a direct purchase on install day, with no trial, to isolate plan length.
- Weekly, £4.99: 8% of installs pay. Renewals of 45%, 70% and 78%, then 94.5% a week, which leaves about 1.5% after a year.
- Monthly, £9.99: 5% of installs pay. Renewals of 57%, 72% and 78%, then 86.8% a month, which leaves about 9% after twelve renewals.
- Annual, £39.99: 3.5% of installs pay, and 30% renew at month twelve.
- Netting: UK App Store prices include 20% VAT. Apple keeps 30% in a subscriber's first year of paid service and 15% after it, per its subscription terms, and 3% of first payments are refunded.
The conversion ratio, weekly converting 2.3 times as many installs as annual, sits at the bottom of the 2x to 7x range Adapty reports. The table shows net revenue as a share of the £1,500 spend, which is net ROAS, at each point in the cohort's life.
| Cohort age | Weekly, £4.99 | Monthly, £9.99 | Annual, £39.99 |
|---|---|---|---|
| Day 0 | 15% | 19% | 53% |
| Day 7 | 22% | 19% | 53% |
| Day 30 | 34% | 30% | 53% |
| Day 60 | 47% | 30% | 53% |
| Day 90 | 57% | 38% | 53% |
| Day 180 | 77% | 54% | 53% |
| Day 364 | 92% | 72% | 53% |
| Day 365 (annual renewal) | 92% | 74% | 73% |
Three things stand out. First, annual leads for almost three months. The weekly cohort passes it on day 77 and the monthly cohort on day 152, so any decision made on a 30 or 60 day read favours annual. Second, at the end of year one the weekly cohort has earned the most per install, 92% net ROAS against 72% for monthly and 53% for annual, even though only 1.5% of its subscribers are left. Third, per paying subscriber the order reverses: year-one net value is about £17.29 weekly, £21.55 monthly and £22.63 annual, rising to £31.13 once the annual renewal lands on day 365.
None of the three has paid back its media within a year, net of VAT and Apple's cut. That is normal at these prices and this CPI, and it is why the netting matters. Our guide to break-even CPI explains why gross dashboard revenue overstates what a cohort can fund.
How sensitive the answer is
The ranking is fragile, which is the main lesson. If the annual version converted 4.4% of installs instead of 3.5%, it would match the weekly cohort's year-one revenue on the day its renewals land. That is a conversion ratio of 1.8x rather than 2.3x. If the weekly first renewal were 35% instead of 45%, the bottom of RevenueCat's range, the weekly cohort's year-one ROAS would fall from 92% to about 75%. Small changes in two numbers decide the winner, and both numbers are specific to your app and your traffic. A benchmark cannot settle it. A test can.
What each plan length does to your Meta account
Using the same hypothetical app, this is how the three versions look from inside the ad account.
| Per 1,000 installs at £1.50 | Weekly | Monthly | Annual |
|---|---|---|---|
| Paying subscribers per 1,000 installs | 80 | 50 | 35 |
| Media cost per paying subscriber | £18.75 | £30.00 | £42.86 |
| Monthly spend for about 50 purchases a week | About £4,100 | About £6,500 | About £9,300 |
| Gross value of each first purchase | £4.99 | £9.99 | £39.99 |
| Share of year-one revenue collected by day 7 | 24% | 26% | 100% |
Signal volume
Meta's learning phase guidance says an ad set usually exits learning after about 50 results in the week after its last significant edit. Optimising on first purchases, the weekly version gets there on about £4,100 a month and the annual version needs about £9,300. If your budget sits between those, plan length decides whether Meta can optimise on the event you care about or has to settle for trial starts or installs. Our guide to hard paywall vs freemium runs the same calculation for the access model, which usually matters even more.
Value signals
If you send purchase values to Meta, an annual purchase reports eight times the value of a weekly one on the day it happens, and the weekly subscriber's later renewals arrive long after the ad was judged. Any value-based bidding or ROAS ranking inside the platform will therefore lean towards whatever creative attracts annual buyers. That is right only if annual buyers are genuinely worth more per install for your app, which the example above shows is not guaranteed.
Attribution and reporting windows
The longest click window in Meta's standard attribution settings is 7 days. In the example, the annual cohort has collected all of its first-year revenue by then, while the weekly and monthly cohorts have collected about a quarter. Platform ROAS will make annual look about four times better, relative to the other two, than it really is. Judge creative against creative inside the platform, where every ad on the same plan suffers the same distortion, and judge budgets on cohort revenue from your subscription data. Our Day-7 ROAS guide explains how to set an early checkpoint that predicts the mature result for your plan mix.
Cash
Annual plans fund the next month of spend fastest, which is the strongest argument for them. In the example the annual cohort has returned 53% of its media cost on day zero; the monthly cohort takes until its sixth payment, around day 150, to match that. The store then adds its own delay on every plan, because Apple pays proceeds up to 45 days after the end of its fiscal month. Our CAC payback guide works through what that does to your bank balance at real spend levels.
How plan length changes the creative
This is the part we work on most with clients, and it is where plan length reaches the ad most directly. A weekly plan is a small, low-commitment purchase, so the ad can sell an immediate result and the paywall can ask for money quickly. The risk is that the ad sells a moment and the product cannot repeat it, which shows up as a weak first renewal two weeks later. An annual plan is a considered purchase, so the ad has to sell a transformation that takes months and make the price feel small against it. Adapty notes that a median annual plan costs only about three times the monthly price globally, so the saving is real and worth showing.
The promise in the ad and the plan on the paywall need to match. An ad that sells a quick fix in front of an annual-only paywall produces installs that do not convert. An ad that sells a long journey in front of a weekly plan produces buyers who churn when the journey turns out to be slow. Our hook patterns for mobile app ads are a useful starting point for matching the promise to the commitment you are asking for.
Which plan mix suits which app
Most apps sell more than one plan, and the real decision is which one to lead with and how hard to push it. The table below is how we would frame that for an app that grows mainly through paid social. It is a judgement framework, not a rule.
| If this describes your app | Lead with | Why |
|---|---|---|
| The value is felt in the first session and the need is short-lived or impulsive | Weekly, usually with a short trial | Buyers pay before the novelty fades, and the plan only has to earn the first few renewals to work. |
| The product solves an ongoing problem people expect to work on for months | Annual as the anchor, monthly as the alternative | The annual plan matches how long the value takes to arrive, and the monthly plan catches buyers who will not commit yet. |
| Usage comes and goes with the problem, such as projects, trips or exam seasons | Monthly | RevenueCat finds monthly churners are the most likely to come back, so the relationship survives gaps that an annual renewal would not. |
| Your Meta budget is modest and you need purchase signal quickly | A plan that converts more installs, often weekly or monthly | More purchases per pound gets an ad set through learning on a deep event at a lower budget. |
| You need campaigns to return cash within weeks to keep spending | Annual | The first year's revenue arrives at once, so payback is decided on day one rather than over months. |
| Your audience is concentrated in lower-income or pay-as-you-go markets | Weekly or monthly | RevenueCat finds Latin America leans weekly and the Middle East and Africa lean monthly; a large upfront price is a harder sell there. |
The pattern underneath: match the billing cycle to how often the user feels the value. Where value is immediate and episodic, shorter plans convert more and lose little by churning. Where value builds over months, the annual plan is simply describing the product honestly.
How to test plan mix without being fooled by selection
Adapty's experiment data suggests plan structure is worth testing: in its 2026 report, 58.7% of plan duration tests improved LTV and 37.8% improved conversion rate, against 34.6% and 31.4% for visual or text-only changes. These are the rules we would apply.
- Randomise the paywall, not the buyers. Compare cohorts that saw different plan defaults or line-ups. Comparing annual buyers with monthly buyers inside one paywall only measures who chose what.
- Measure per install. Paying users and revenue per install decide the media budget. Share of annual sales and revenue per subscriber can both rise while revenue per install falls.
- Wait for the first renewal on short plans. The first renewal is where most weekly and monthly subscribers are lost, so a weekly test needs about three weeks and a monthly test about two months before the read means much.
- Use early warnings on annual. You cannot wait a year for the renewal. RevenueCat finds 35% of annual cancellations happen in the first month, so early auto-renew turn-offs and refund requests are the best available read on whether a push towards annual brought in weaker buyers.
- Hold the ads constant. Run the same creative and optimisation event through the test. Changing plan mix changes purchase values and event volume, so expect Meta to need time to settle afterwards.
How we approach plan length with subscription app clients
We do not set clients' pricing, and plan mix is a product and monetisation decision before it is a marketing one. What we do on our Meta Ads service is build the account around the plans the app actually sells: an optimisation event the budget can support, reporting windows that match when each plan's money arrives, and creative whose promise fits the commitment the paywall asks for. When the numbers say the plan line-up is the constraint rather than the ads, we say so.
Plan length is one of the inputs that set what you can afford to pay for a subscriber, and our subscription app marketing strategy guide covers how it fits with the others. If your app is already spending on Meta and you want the account built around how your plans really convert and renew, see how we work with consumer and subscription apps. We work with accounts spending at least £3,000 a month on Meta, and current fees are on our pricing page.
Frequently asked questions
Is a weekly, monthly or annual subscription best for a mobile app?
No plan length is best in general. Annual plans collect the most cash on day one and earn the most per paying subscriber. Weekly plans usually convert more installs into payers, and if they convert enough more, they can earn more per install over a year despite near-total churn. Monthly plans sit between the two. Which one wins for your app depends mainly on how many more people buy the shorter plan, so measure it per install rather than per subscriber.
Do annual subscribers retain better than monthly or weekly subscribers?
On the headline numbers, by a wide margin. RevenueCat's State of Subscription Apps 2026 puts median year-one retention at 20% to 40% for yearly plans, 6% to 14% for monthly and 1% to 2% for weekly. Two cautions apply. An annual subscriber faces one renewal decision in the first year while a weekly subscriber faces about fifty, and people who choose annual plans are self-selected for stronger need. Pushing more users onto annual will not give them the retention of people who chose it.
Do weekly subscriptions really earn the most revenue?
It depends whose data you read. Adapty's State of In-App Subscriptions 2026 reports that weekly plans generated 55.6% of all app revenue in its dataset in 2025. RevenueCat's 2026 report says monthly subscriptions dominate overall revenue share. The two vendors have different customer bases and Adapty's revenue figures are gross, so neither is the whole market. Revenue share also says nothing about profit after acquisition cost.
Which plan length is easiest to run on Meta ads?
Weekly plans usually generate the most purchase events per install, so Meta has more to learn from at a given budget. Annual plans report their whole first-year value on the day of purchase, which suits short attribution windows and early ROAS reads, but they produce fewer purchases. Monthly and weekly cohorts keep earning long after Meta's 7 day click window closes, so judge them on cohort revenue from your subscription data, not on platform ROAS.
Should I push users towards the annual plan on my paywall?
Test it rather than assume it. Defaulting or discounting the annual plan usually raises cash collected up front, but the extra annual buyers are the least committed ones and will renew below your current annual rate. Judge the test on paying users and revenue per install, not on the share of annual sales, and do not declare a winner until you have at least an early read on cancellations and refunds.
The short version
Annual plans give you the most money per subscriber, all of it up front, and the easiest early read on a campaign. Weekly plans give you more paying users per install and more purchase signal, and can earn more per install over a year if they convert enough more people. Monthly plans split the difference and are the easiest to come back to. The retention gap that dominates most comparisons is partly arithmetic and partly selection, so moving buyers onto a longer plan will not buy you the retention of people who chose it. Work out how many more of your installs buy the shorter plan, test the line-up per install, and build the ad account around the answer.