Skip to content
App Imp

Benchmarks and cutoffs.

Every threshold this site publishes, and every industry figure it cites — kept apart, because they are not the same kind of number.

Counted by us

01

When a complaint counts as a gap

Both bars have to clear at once.

Reviews behind itbelow this, the share is noise
at least 20, across the shelf
Breadthone app's bug is not a market's gap
at least half the apps counted
02

Graded words

The only three words allowed to lead a claim, each a fixed multiple of the store's typical rate. Below the first cutoff no word is used at all.

elevated
from 1.15 times typical
high
from 2 times typical
extreme
from 3.5 times typical
03

The Opportunity score

The site's one composite, and only on markets. Each factor maps a counted number onto 0–100 through a fixed anchor, then the weighted mean is taken over whatever could be counted.

Nobody's fixed itweight 25 · share of the shelf carrying the top gap
straight share — 12 of 13 apps scores 92
How far above typicalweight 20 · the gap's share against the store's typical rate
tops out at 4 times typical
How angryweight 15 · mean star on the gap theme; needs a deep report
star3.0 scores 0 · star1.5 scores 100
People in the marketweight 15 · Apple's rating counts across the shelf, log scale
10,000 scores 0 · 10 million scores 100
Room to enterweight 15 · one minus the leader's share of shelf ratings
a leader holding 59% scores 41
Willingness to payweight 10 · confirmed charging out of actually checked
straight share — the unchecked are named, never assumed free

A factor that cannot be counted is dropped and the page says which; no qualifying gap means no score at all. The weights are a v1 judgment, published so they can be argued with — they are not themselves measured.

Cited, not counted

Everything below is published by RevenueCat, State of Subscription Apps 2026 — 115,000+ apps · $16B+ tracked revenue · >1B transactions. We did not measure any of it, none of it feeds a score of ours, and we attach no verdict to it. Rows from an older edition say so.

04

Paywalls and trials

Cited
hard paywalls out-convert freemium (install → paid, ~day 35)
10.7% vs 2.1%
hard paywalls out-earn freemium, revenue per install
D14 $2.32 vs $0.27 · D60 $3.09 vs $0.38
the catch is not retention: 1-year annual-plan retention is near-identical across models
~27% vs ~28%
longer trials convert better: 17–32-day vs under-4-day, median trial→paid
42.5% vs 25.5%
half of all paid conversions happen the first day the app is opened
~50% on day 0
the market moves the other way anyway: under-4-day trials were ≈ half of all trials in 2025
46.5% share
05

Plan length and renewals

Cited
yearly plans renew far better than short cycles
83.4% yearly renewal — more than four times weekly, about twice monthly
the annual-churn cliff is month one
35% of all annual cancellations land in month 1
cancelled annuals rarely return
annual reactivation ~5%; monthly subscribers return at about four times that rate
weekly plans are now the revenue engine for most categories
55.5% of all subscription revenue; best 12-mo LTV when paired with a trial (Health & Fitness is the exception)
duration mix sold skews annualHealth & Fitness
68% of subscriptions sold are annual
the monthly-first categoryProductivity
77% of subscriptions sold are monthly · 91% of revenue from monthly (different denominators — both true)
renewal rates by planHealth & Fitness
weekly 54% (top-3 of 11) · monthly 57% · annual 25% (mid-table)
steady mid-table renewals across all plan lengthsUtilities
habitual use keeps churn flat regardless of billing cycle
06

Price, revenue and refunds

Cited
cheaper apps keep annual subscribers better
1-yr annual retention: low-priced 36% · mid 26% · high 23%
highest install LTV of any categoryHealth & Fitness
$1.21
highest refund rate of any categoryEducation
4.86% (2025 edition)
second-highest refund rateHealth & Fitness
4.71% (2025 edition)
07

The market itself

Cited
the median app barely grows; the tails are extreme
median MRR +5.3% YoY · top decile +306% · bottom decile contracting
the launch flood vs the incumbency moat
~2,000 → 14,700+ new subscription apps per month (Jan 2022 → Jan 2026), yet pre-2020 apps still hold 69% of subscription revenue
even top categories take over 100 days to reach $10k revenue
>100 days to the first $10k
AI-powered apps churn annuals faster
annual retention 21.1% vs 30.7% non-AI; cancellations ~30% faster

A further 11 cited rows are held back rather than shown, because each reaches us through a relay or carries an edition ambiguity we have not resolved at the source. They ship when that is settled, not before. Our own numbers are re-counted weekly from App Store reviews; nothing here is modelled, and per-keyword search-volume estimates are refused outright — see the counted app ideas for what the cutoffs above produce.