Retention Rate Cohort Calculator

The Retention Rate Cohort Calculator turns a cohort size and a list of active-user counts into a complete retention picture: the retention rate at every checkpoint, an animated retention curve, a colour-coded period-over-period drop heatmap, the cohort half-life, a smile-curve / product-market-fit flattening detector, a power-law fit with steady-state extrapolation, and a quartile classification against published industry benchmarks for SaaS, mobile, consumer-subscription, e-commerce, and fintech cohorts. Unlike single-rate churn calculators, a cohort retention curve captures the shape of the drop-off — and the shape, not any single number, is what tells you whether your product is sticky.

How to Use

1. Enter the cohort size — the number of users acquired in the same window (sign-up week, paid-conversion month, install day, etc.).
2. Pick the measurement unit. Mobile apps usually measure in days, B2C subscriptions in weeks, and B2B SaaS in months.
3. Enter the checkpoint periods as a comma-separated list. The default 1, 7, 14, 30, 60, 90 is the classic mobile-app schedule; for B2B SaaS try 1, 2, 3, 6, 9, 12 months.
4. Enter the active users at each checkpoint, in the same order. Each value is the count of users from the original cohort still active at that checkpoint — never re-baseline the denominator.
5. Pick the closest industry benchmark — B2B SaaS, SMB SaaS, consumer subscription, mobile app, e-commerce, or fintech.
6. Click Build Retention Curve to see the retention rate at every checkpoint, the animated curve, the drop heatmap, the half-life, the smile-curve verdict, the power-law fit, and the quartile classification.

Formulas Used

What Makes This Cohort Calculator Different

• Curve, not a single number. Most "retention rate calculators" just compute one rate. This one builds the full retention curve at up to ten user-defined checkpoints — the shape is what reveals whether the cohort hits a floor or keeps leaking.
• Smile-curve / PMF flattening detector. A built-in test for the last three checkpoints classifies the cohort as strong flatten (the classic Sean Ellis PMF signal), mild flatten, or still decaying.
• Cohort half-life via interpolation. Linear interpolation between checkpoints gives a precise estimate of the period at which 50% of the cohort has churned — not just whether you crossed it.
• Power-law fit and steady-state projection. OLS on the log-log transform fits R(t) = a · t^(−k), reports R², and projects retention out to roughly 3× the last observed period when the fit is good.
• Drop-off heatmap. Period-over-period drops are colour-coded from green (≤2 pp) to deep red (>50 pp) so the highest-leverage leak is visible at a glance.
• Industry quartile benchmarks. Six segment-specific benchmarks (B2B SaaS, SMB SaaS, consumer subscription, mobile app, e-commerce, fintech) classify your Day-1, Day-7 and Day-30 equivalents against published P25 / P50 / P75 bands.
• Animated SVG curve. The line draws, the area fills, and the value labels pop in sequence so the drop-off is visually obvious — not just a static plot.
• Day, week, or month units. Pick the time scale that matches your product's natural cycle. Mobile apps live in days; B2B in months.
• Live preview before submission. Type any number and the mini-bar preview, percentage, and PMF badge update instantly — no full-page reload needed.
• Step-by-step math. Every calculation is broken down line by line so you can verify, document, or learn from the result.

What Is a Cohort, Exactly?

A cohort is a group of users who share an acquisition event in the same time window — most commonly the week, day, or month they signed up, installed, or made their first purchase. Cohort analysis follows that same fixed group across subsequent periods, so the denominator never changes. This is what makes cohort retention strictly more honest than "monthly active divided by total users", which can hide accelerating churn behind acquisition growth.

Retention Rate vs Churn Rate

Retention rate and churn rate are inverses for any given period: retention = 1 − churn. The reason teams choose one or the other is psychological more than mathematical — framing a number as "95% retained" rather than "5% churn" tends to shift attention from preventing losses to celebrating saves. The cohort version of retention is more informative than period churn because it follows the same users over time, exposing where the leaks are: a 50% Day-1 churn followed by perfectly flat retention is a fundamentally different (and better) cohort than a slow, persistent decay.

Reading the Retention Curve Shape

• Steep early drop, flat tail (the "smile"). Most acquired users were the wrong fit, but a stable core finds the product indispensable. This is the textbook product-market-fit shape — and the floor is the size of your true addressable market within the acquired segment.
• Gentle, steady decay. Familiar shape for low-frequency products (annual purchase items, tax tools, wedding-planning apps). Not necessarily bad — but unit economics depend on the slope, not the level.
• Initial flat then sudden drop. Almost always a forced-cancellation event — a trial ending, a payment failing, an annual contract not renewing. Look for the structural cause, not the user behaviour cause.
• Still falling at the last checkpoint. No floor in sight. Either you need more observation time, or the product has not yet found a sticky use case for this cohort.
• Curve going up. Almost always a measurement artefact: re-activations being counted as cohort members, or activity definitions inflated by a new feature. Investigate before celebrating.

Retention Benchmarks by Category

Quartiles inferred from AppsFlyer, Mixpanel, Amplitude, OpenView, RJMetrics public reports. Treat as rough orientation, not industry standards.

Common Cohort Analysis Pitfalls

• Re-baselining the denominator. Always divide by the original cohort size at every checkpoint. Dividing by "users still active last period" inflates retention every step and hides the leak.
• Mixing definitions of "active". Was Day-30 retention measured by login, by a meaningful action, or by paid status? Each gives a different curve. Pick one and report consistently.
• Confusing cohort retention with revenue retention. Cohort logo retention treats every user equally; net revenue retention weights by dollars. They diverge sharply for products with wide pricing tiers.
• Insufficient cohort size. Cohorts smaller than ~500 users have wide confidence intervals at every checkpoint. Roll up several adjacent acquisition weeks rather than over-interpreting a small cohort.
• Reading a still-falling tail as a floor. Three near-flat points in a row is the minimum signal for PMF flattening. One flat-looking point can be noise.
• Ignoring the curve shape. A 30% Day-30 retention from a sharp drop + flat tail is fundamentally different from a 30% from a slow steady decay. Same number, opposite implications.
• Forced-cancellation artefacts. Trial ends, payment failures, and annual-contract renewals create artificial drops at predictable periods. Annotate them; don't read them as product-driven churn.

How Retention Drives Customer Lifetime Value

Higher retention means longer customer lifetimes, which directly increases customer lifetime value (CLV). The simplest CLV formula is CLV = (monthly ARPU × gross margin) ÷ monthly churn; equivalently CLV = (monthly ARPU × gross margin) × average customer lifetime in months. Because lifetime is the inverse of churn, a one-percentage-point retention improvement on a 5% monthly churn base extends lifetime from 20 months to 25 months — a 25% CLV uplift at the same ARPU. This is why retention investments typically out-perform acquisition investments at the same dollar level once a product has reached even modest scale.

Where to Look First to Improve Retention

1. First-period activation. The steepest drop on most curves is between Day 0 and Day 1 (or week 0 → week 1). Onboarding clarity, time-to-first-value, and proactive activation outreach typically have the highest leverage.
2. Forced-cancellation moments. Trial endings and renewal moments concentrate churn at predictable periods. Map your cohort drops against these structural moments before blaming product engagement.
3. Failed-payment recovery. A surprising fraction of "churn" is just expired credit cards. Card-update tooling and dunning sequences recover 5–15% of monthly churn for most B2C SaaS.
4. Power-user identification. Cohort members still active at the last checkpoint are your stickiest segment. Look at what they share — feature usage, integration adoption, team size — and target acquisition to look-alikes.
5. Cohort segmentation. If the overall cohort never flattens, segment it. Often one acquisition source or one persona is hitting PMF inside an otherwise mediocre overall curve.

FAQ

What is a cohort retention rate?

A cohort retention rate is the percentage of users from a single acquisition cohort still active at a later period. It equals active users at the checkpoint divided by the original cohort size, expressed as a percentage. Because the cohort denominator is fixed, the retention rate falls monotonically — every churn lowers it, and re-activations cannot inflate it.

How is cohort retention different from monthly retention?

Monthly retention typically divides monthly active users by total registered users — a number that can stay flat or grow even while acquired cohorts are leaking, because new acquisitions plug the gaps. Cohort retention follows a fixed group over time, so it cannot be hidden by acquisition growth. Cohort is strictly more honest for assessing product stickiness; monthly retention is more useful for top-line growth reporting.

What checkpoints should I use?

The classic mobile-app schedule is Day 1, Day 7, Day 14, Day 30, Day 60, Day 90 — short enough to catch early drop-off and long enough to see a floor form. For B2C subscriptions try weekly checkpoints; for B2B SaaS try monthly. The calculator supports any 2–10 strictly-increasing integer values.

What is a smile curve in cohort analysis?

A smile curve is a retention curve that drops sharply early and then flattens to a stable floor. The flattening is the classic Sean Ellis signal of product-market fit: a stable share of acquired users find the product genuinely indispensable. A curve that keeps decaying never reaches the smile and indicates the product has not yet found a sticky use case for that segment.

What is good Day-30 retention?

Good Day-30 retention depends heavily on the product category. B2B SaaS top-quartile cohorts retain 65% or more at Day 30. SMB SaaS top quartile is around 45%. Consumer subscriptions are typically 20% at the median. Mobile apps are typically 9% at the median and 18% at the top quartile. The calculator positions your Day-30 against the right benchmark for the segment you select.

How does the half-life calculation work?

Half-life is the period at which retention crosses 50%. The calculator scans your checkpoints for the first pair that straddles 50% retention and linearly interpolates between them. If retention never falls below 50% in the observed window, the half-life is "beyond the data" — a strong stickiness indicator.

What does the power-law fit tell me?

Most cohort retention curves approximate a power law R(t) = a · t^(−k). The calculator fits this model to your checkpoints with ordinary least squares on the log-log transform and reports the goodness-of-fit (R²). When R² exceeds 90% the power law summarises your curve well, and the model's extrapolation to roughly 3× the last observed period is a meaningful projection of where retention is heading. Lower R² usually indicates a structural inflection (trial end, renewal event) that the smooth power-law model cannot capture.

Can I use this for revenue retention instead of user retention?

Yes — enter the cohort's total revenue at acquisition as the cohort size, and the revenue still being collected from that cohort at each checkpoint. The calculator does not know whether you fed it users or dollars, so the math works identically. This gives gross revenue retention (GRR). For net revenue retention (NRR) that includes expansion, you would need to add expansion revenue to the numerator — the calculator does not do this directly.

How many cohort members do I need for the numbers to be meaningful?

As a rough rule, cohorts below 500 users have noisy retention curves at long horizons, and individual user re-activations can shift the curve by visible amounts. For statistical robustness either roll up adjacent acquisition periods (multiple weeks of sign-ups into one cohort) or treat each retention point as a confidence interval rather than a single number.