Consumer Subscription App Retention Benchmarks: How to Calculate LTV, ARPU vs. ARPPU, and Cohort Churn
Learn how to calculate subscription app LTV, ARPU vs. ARPPU, and cohort churn rates using verified benchmarks across weekly, monthly, and annual billing cycles.
Key takeaways
- Billing duration dictates retention baselines: According to benchmark data from RevenueCat, annual subscriptions show a median 1-year retention rate of 28% (with 17% in the lower quartile and 43% in the upper quartile), compared to 11% for monthly subscriptions and 3% for weekly subscriptions. Monthly and annual cohorts cannot be evaluated against the same retention curve.
- Separate active users from paying subscribers: Average Revenue Per User (ARPU) divides total revenue across all active users in a period, whereas Average Revenue Per Paying User (ARPPU) isolates users who made at least one purchase. Blending non-payers into revenue metrics hides shifts in paywall conversion efficiency.
- Track churn by acquisition cohort and renewal cycle: App churn equals the share of users lost during a window relative to the starting count. In subscription businesses, churn must be measured across specific renewal points (Cycle 1, Cycle 2, and Cycle 3) rather than as a single aggregate monthly percentage.
- Early LTV proxies guide ad scaling: Waiting 12 months to measure annual cohort payback stalls marketing budgets. Practical subscription growth relies on early predictive signals (such as Day 3 to Day 30 cohort revenue curves and early retention checkpoints) to establish allowable Cost Per Acquisition (CPA) thresholds across ad networks.
Deconstructing subscription unit economics: CPI, CPA, ARPU, and ARPPU
When managing paid user acquisition across networks like Meta, Google, TikTok, Apple Search Ads, Unity Ads, Moloco, and Appier, looking only at top-of-funnel install metrics creates a revenue blindspot. A low Cost Per Install (CPI) often masks campaigns that fail to generate recurring revenue.
| Metric | Formula |
|---|---|
| CPI | Total Ad Spend ÷ Attributed Installs |
| CPA | Total Ad Spend ÷ Paying Subscribers (or Trial Starts) |
| ARPU | Total Revenue ÷ Total Active Users (in defined period) |
| ARPPU | Total Revenue ÷ Total Paying Users (in defined period) |
Cost Per Install (CPI) vs. Cost Per Acquisition (CPA)
- CPI (Cost Per Install) measures ad spend divided by attributed installs: CPI = Total Ad Spend ÷ Attributed Installs
- CPA (Cost Per Acquisition) measures the cost to secure a monetization event, such as a paid trial activation or a first subscription payment: CPA = Total Ad Spend ÷ New Paying Subscribers (or Validated Trials)
If an ad campaign spends $2,000 to generate 1,000 attributed installs, the CPI is $2.00. However, if only 40 of those installers start a trial and convert into a paying subscriber, the actual CPA is $50.00 ($2,000 / 40). Evaluating creative performance solely on the $2.00 CPI will lead a growth team to scale ad sets that bring in non-monetizing installs.
ARPU vs. ARPPU: Why the denominator matters
According to definitions from Mixpanel and AppsFlyer, Average Revenue Per User (ARPU) calculates the average revenue generated across your entire active user base:
ARPU = Total Revenue in Timeframe ÷ Total Active Users in Same Timeframe
As outlined by Adjust and Sendbird, Average Revenue Per Paying User (ARPPU) narrows the denominator strictly to users who completed a financial transaction:
ARPPU = Total Revenue in Timeframe ÷ Unique Paying Users in Same Timeframe
According to Mixpanel, separating paying users from the broader active user base is critical for freemium and trial-based models because a large influx of free users lowers blended ARPU, even when monetization among subscribers remains strong. Tracking ARPU and ARPPU side by side reveals whether revenue changes stem from shifts in conversion rates, pricing changes, or subscriber engagement.
Retention benchmarks across billing cycles and app lifecycles
Evaluating retention requires distinguishing between general in-app engagement (whether users open the app) and subscription renewal retention (whether users renew their paid plans).
In-app product engagement retention (Day 1, Day 7, Day 30)
According to benchmark research from Sendbird, in-app retention across mobile app categories drops sharply over the first month:
- Day 1 Retention: 25.3% average
- Day 7 Retention: 12.0% average
- Day 30 Retention: ~5.0% broad average across all categories (with category-specific engagement ranges reaching 27% to 43% in specialized verticals, as reported by Sendbird)
According to Business of Apps, comparing retention at Day 1, Day 7, and Day 30 pinpoints where onboarding friction occurs. A steep drop between Day 1 and Day 7 indicates that new users are missing the core value proposition before reaching a paywall.
12-month subscription renewal retention benchmarks
Subscription retention decay curves vary significantly depending on billing duration. According to data published by RevenueCat analyzing consumer subscription apps, 1-year subscriber retention breaks down as follows:
| Billing Duration | Median 1-Year Retention |
|---|---|
| Annual Plans | 28% |
| Monthly Plans | 11% |
| Weekly Plans | 3% |
As highlighted by Blustream, monthly and annual subscriptions require separate target baselines. An annual plan experiences a single renewal decision point at Month 12, whereas a monthly plan faces twelve separate renewal decisions over that same year.
Calculating cohort churn, customer lifespan, and realized LTV
Cohort churn formulas
According to Sendbird and Ethora, app churn and retention rates are calculated using cohort boundaries:
Retention Rate (%) = (Active Users at End of Period ÷ Active Users at Start of Period) × 100
Churn Rate (%) = (Users Lost During Period ÷ Active Users at Start of Period) × 100
For subscription models, point-in-time churn (e.g., total cancellations divided by total active subscribers this month) can be misleading when new user acquisition fluctuates. Subscription churn should instead be tracked across distinct acquisition cohorts tied to signup dates, as supported by Airbridge and RevenueCat.
Modeling customer lifespan and LTV
According to Sendbird and Mixpanel, basic Customer Lifetime Value (LTV) models customer revenue over time:
Customer Lifespan = 1 ÷ Cohort Periodic Churn Rate
LTV = ARPPU × Average Customer Lifespan
Blended LTV = ARPU × Average User Lifespan
When factoring in free trials, the calculation incorporates the trial-to-paid conversion rate:
Realized LTV per Install = Trial Start Rate × Trial-to-Paid Conversion Rate × Subscriber LTV
Worked calculation: Monthly vs. annual LTV
Assume an app charges $10/month for a monthly subscription and $60/year for an annual subscription:
-
Monthly Plan Cohort:
- Average monthly churn across 12 cycles = 18% (0.18).
- Average lifespan = 1 ÷ 0.18 ≈ 5.55 months.
- Gross Subscriber LTV = 5.55 × $10 = $55.50.
- Net Subscriber LTV (after 15% platform fee) = $55.50 × 0.85 = $47.18.
-
Annual Plan Cohort:
- Year 1 renewal rate = 28% (median benchmark).
- Expected renewals over 3 years: Year 1 = 1.0, Year 2 = 0.28, Year 3 = 0.28 × 0.50 = 0.14.
- Average billable billing cycles = 1.0 + 0.28 + 0.14 = 1.42 years.
- Gross Subscriber LTV = 1.42 × $60 = $85.20.
- Net Subscriber LTV (after 15% platform fee) = $85.20 × 0.85 = $72.42.
According to research from RevenueCat, comparing monthly and annual subscriber LTV helps teams determine plan pricing, set promotional discounts, and decide which package to emphasize on onboarding paywalls.
Actionable early-window LTV modeling: Turning early signals into acquisition decisions
Waiting 12 months to measure actual cohort renewal revenue limits how quickly growth teams can deploy ad spend. If a growth marketer spends $4,000 on paid ads this month across Apple Search Ads and Meta, they need to know whether the acquisition is profitable within days, not after a full annual billing cycle.
To bridge this observation gap, growth teams use early predictive modeling approaches:
- Ratio-driven early LTV modeling: According to AppAgent, teams apply a historical D90-to-D3 or D180-to-D30 multiplier to early revenue cohorts: Forecasted D180 LTV = D30 Realized Cohort Revenue × (Historical Mature Cohort D180 Revenue ÷ Historical Mature Cohort D30 Revenue)
- Machine learning predictive LTV: Platforms analyze early in-app engagement patterns, initial trial choices, session depth, and early churn signals during the first 72 hours to forecast long-term cohort value. For example, Airbridge Predictive LTV forecasts up to 180 days of lifetime value using 3 days of cohort data, providing predictive LTV and ROAS checkpoints at Days 30, 60, 90, and 180.
As detailed in the Airbridge Core Plan documentation, subscription measurement tools can also hold postbacks for up to 24 hours so canceled trials and immediate refunds are filtered out before conversion signals reach ad network algorithms. This prevents automated bid strategies from optimizing toward low-quality trial activations.
Translating LTV to paid user acquisition: ROAS, ROI, and allowable CPA
Understanding subscription LTV allows growth teams to set clear guardrails for customer acquisition costs.
| Metric | Formula |
|---|---|
| ROAS (Return on Ad Spend) | (Attributed Ad Revenue ÷ Ad Spend) × 100 |
| ROI (Return on Investment) | (Net Profit ÷ Total Costs) × 100 |
| Max Allowable CPA | Net Realized LTV × (1 - Target Margin) |
ROAS vs. ROI in subscription apps
- ROAS (Return on Ad Spend) evaluates gross revenue generated directly by an ad channel relative to its ad spend. A campaign that spends $1,000 and generates $2,000 in gross subscription revenue achieves a 200% ROAS ($2,000 / $1,000).
- ROI (Return on Investment) accounts for all operational costs, including app store platform fees (15% to 30%), server hosting, attribution platform costs, and payment infrastructure expenses: ROI = ((Gross Revenue - Ad Spend - Store Fees - COGS) ÷ (Ad Spend + COGS)) × 100
Establishing allowable CPA by ad channel
To maintain cash flow positive acquisition across channels like Meta, Apple Search Ads, TikTok, Unity Ads, Moloco, or Appier, teams set their maximum allowable CPA based on expected cohort payback windows.
- Calculate Net Realized LTV per Paying Subscriber
- Gross 1-Year Expected LTV: $60.00
- App Store Fee (15%): -$9.00
- Net Realized Subscriber LTV: $51.00
- Apply Target Profit Margin (30%)
- Target Margin per Subscriber (30%): -$15.30
- Maximum Allowable CPA (Payer): $35.70
- Derive Maximum Target CPI from Funnel Conversion Rates
- Trial Activation Rate (Install to Trial): 8%
- Trial-to-Paid Conversion Rate: 50%
- Overall Install-to-Paid Conversion: 4% (0.08 × 0.50)
- Maximum Allowable CPI = $35.70 × 0.04 = $1.43
By connecting attribution data directly with subscription lifecycle events, growth marketers avoid optimizing for low-cost installs that churn before paying, focusing spend instead on campaigns that deliver profitable long-term subscribers.
FAQS
Frequently asked questions
What is a good 1-year retention rate for a consumer subscription app?
According to benchmark data from RevenueCat, the median 1-year subscriber retention rate for annual plans is 28%, with top-quartile apps achieving 43% and bottom-quartile apps retaining 17%. For monthly plans, median 1-year retention drops to 11%, while weekly plans average 3%.
What is the difference between ARPU and ARPPU?
According to Mixpanel and Adjust:
- ARPU (Average Revenue Per User) divides total revenue by the entire active user base (both free and paying users) over a specified timeframe.
- ARPPU (Average Revenue Per Paying User) divides total revenue strictly by unique users who completed a payment in that timeframe. ARPPU isolates paying customer value, while ARPU reflects overall app monetization efficiency across the full user base.
How is subscription app churn rate calculated?
According to Sendbird, general app churn is calculated as:
Churn Rate (%) = (Users Lost During Period ÷ Active Users at Start of Period) × 100
In subscription apps, this formula should be applied to specific cohort renewal windows (e.g., Cycle 1 churn, Cycle 2 churn, and annual renewal churn) to track where cancellations occur along the billing lifecycle.
What is the difference between ROAS and ROI for subscription marketing?
ROAS measures gross attributed ad revenue divided by advertising cost (e.g., $3,000 revenue from $1,000 ad spend = 300% ROAS). ROI measures net profit after subtracting ad spend, app store commission fees (15% to 30%), server infrastructure, and tooling expenses, divided by total investment.
How do you estimate subscription LTV before an annual cohort matures?
Teams use two primary methods:
- Historical Ratio Multipliers: Applying mature cohort growth factors (e.g., multiplying Day 30 revenue by the historical D365-to-D30 ratio), as described by AppAgent.
- Machine Learning Predictive Models: Analyzing early in-app engagement and trial behavior from the first 3 days to forecast revenue up to 180 days out, available in tools like Airbridge ($40+/mo Core Plan with a 30-day free trial).
Scale subscription acquisition with predictive LTV
Unify campaign attribution with renewal cohort analytics and Day 3 predictive ROAS modeling.
