Every product team tracks sign-ups. Far fewer can tell you, cohort by cohort, how many of those users are still around ninety days later — and that second number is the one that decides everything. Acquisition is rented; retention is owned. A product that keeps its users compounds quietly in the background, while a product that loses them has to re-buy its growth every single month.
User retention is the percentage of users who keep using your product over time. It is the mirror image of churn, the source of every durable growth curve, and — for most SaaS products — the cheapest growth lever available. Acquiring a new customer costs 5 to 25 times more than retaining an existing one (Harvard Business Review), and research by Bain & Company shows that increasing retention by just 5% can boost profits by 25% to 95%. Retention is not a metric you report; it is the business model.
This guide covers the full picture: what retention means (with a precise definition), the retention rate formula with a worked example, the measurement variants that change the number (N-day vs rolling vs unbounded, user vs revenue retention), what a good retention rate looks like in SaaS, real retention examples, why users actually stay — and where to start if you want to improve it.
Key Takeaways
- User retention is the percentage of users still active after a given period. It is churn's complement: 97% monthly retention = 3% monthly churn.
- The formula is simple: (users at period end − new users) ÷ users at period start × 100. Subtracting new users is the step that keeps acquisition from hiding real losses.
- The definition changes the number. N-day, rolling, and unbounded retention can describe the same product with wildly different figures — always check which one a benchmark uses.
- User retention and revenue retention answer different questions. One grades the product, the other grades the business — and NRR above 100% means growth with zero new sign-ups.
- Small differences compound enormously. 95% vs 98% monthly retention sounds close; after a year it is the difference between keeping half your users and keeping four out of five.
- Retention is decided early. Users who reach the aha moment in their first sessions form the plateau of your retention curve; users who don't were never really retained at all.
What Is User Retention? (Definition & Meaning)
User retention — also called customer retention when measured at the account level — is the percentage of users who continue using your product over a given period. Take a group of users who signed up together, come back after a day, a week, or a month, and count how many are still active: that share is your retention. It answers the only question that ultimately matters about a product: once people have tried it, do they keep choosing it?
Retention and churn are two views of the same number — a product with 97% monthly retention has 3% monthly churn — but the framing does different work. Churn asks "who did we lose, and why?", which makes it the alarm system. Retention asks "who stays, and what makes them stay?", which makes it the strategy: find the behavior your retained users share, and build your onboarding to produce more of it.
Retention, defined. User retention is the percentage of users from a starting group who are still active after a given period, excluding anyone who joined along the way. If 2,000 users start the month and 1,940 of them are still active at its end, monthly retention is 97%. Measured across many time points for one sign-up cohort, these percentages form the retention curve — the single most honest chart in SaaS.
The most useful way to picture retention is as a cohort thinning over time. A hundred users sign up on day zero. Some never return after the first session. More drift away over the first weeks. And then — if the product delivers real recurring value — the thinning stops, and a stable core keeps coming back indefinitely. That surviving core is what "retained" actually means:
Retention is a cohort thinning until it stabilizes. The filled dots that survive to day 30 — the plateau — are the users the product genuinely retained.
Why retention is the metric that decides everything else
Because everything downstream inherits it. Retention sets customer lifetime value, which sets what you can afford to spend on acquisition, which sets your growth rate. Retained users are also where expansion revenue, referrals, and honest product feedback come from — none of which a churned user provides. Two products with identical sign-up numbers and different retention are two different companies: one is compounding, the other is refilling a leaky bucket. It is also the truest test of product-market fit: marketing can manufacture sign-ups, but only genuine recurring value can make people come back week after week — which is why a flattening retention curve is the classic signature of product-led growth working.
Retention Rate: The Formula and How to Calculate It
The retention rate formula compares the users you kept against the users you started with, after removing anyone new who joined during the period:
Subtract the new users — retention measures the starting group only.
A worked example: you start June with 2,000 users. During the month, 300 new users sign up, and you end the month with 2,240 active users. Retention = (2,240 − 300) ÷ 2,000 = 1,940 ÷ 2,000 = 97%. The subtraction is the whole trick: skip it and the same month reads as 112% "retention" — growth masquerading as loyalty, hiding the 60 users you actually lost.
The measurement decisions that make or break the number
- Define "active" honestly. A login is a weak signal; a core action — the thing your product exists to do — is a strong one. Retention measured on logins flatters you; retention measured on value delivered tells the truth.
- Match the period to the product's natural frequency. A daily habit product should watch day-1 and day-7 retention; an invoicing tool used monthly should not panic over quiet weeks. Judging a monthly-use product on daily retention manufactures a crisis out of normal behavior.
- Cohort it. One blended number mixes yesterday's sign-ups with loyal three-year-old accounts and hides both improvement and decay. Track each sign-up cohort separately over time — the format of the retention curve — so you can see whether this month's onboarding beats last month's.
- Segment it. Retention by plan, persona, and acquisition channel. A 90% blended rate that splits into 98% for enterprise and 70% for one ad channel is not one number — it is a to-do list.
Why small retention differences are enormous
Retention compounds, which makes close-looking numbers wildly different a year out. Retained users after 12 months = (monthly retention)12 — and that exponent is merciless:
Retention compounds: after 12 months = (monthly retention)¹². The gap between 95% and 98% monthly retention is the gap between keeping half your users and keeping four out of five.
Beware the vanity version. Total active users can climb for months while retention quietly falls — strong acquisition papering over cohorts that each keep less than the last. This is exactly the failure mode the formula's "subtract new users" step exists to expose. If you only track one growth chart, make it retention per cohort, not total actives.
The Types of Retention: N-Day, Unbounded, User vs Revenue
"Retention" is one word covering several different measurements, and the differences are big enough to turn the same product into a success story or a crisis. Three distinctions do most of the work:
Retained = returned on exactly day N. Strict; the standard for daily-habit products.
Retained = returned on day N or within a window around it. Smooths frequency noise.
Retained = returned on day N or any day after. Fits low-frequency products.
People and accounts that stay. The product-health signal: every retained user is value delivered.
MRR that stays. The business-health signal — and NRR can exceed 100% through expansion.
Does the user still use the product at all? The number this guide is mostly about.
Does the user keep using a specific feature? Reveals which features carry the product — and which shipped into silence.
N-day vs unbounded: same product, different number
Take a project-management tool someone uses every Monday. Day-30 classic retention might record them as churned if day 30 lands on a Thursday; unbounded retention counts them as retained because they came back on day 33. Neither is wrong — but a benchmark measured one way is meaningless against your number measured the other. The rule: pick the definition that matches your product's natural usage frequency, state it next to every chart, and never change it silently.
User retention vs revenue retention
Gross revenue retention (GRR) is the share of recurring revenue you keep from existing customers, ignoring expansion — its ceiling is 100%, and it grades how leak-proof the base is. Net revenue retention (NRR) adds upgrades and expansion, so it can exceed 100%: at that point your existing customers grow in value even if you sign nobody new, the revenue-side twin of negative churn. The two can diverge from user retention in both directions — you can retain most of your users while revenue quietly downgrades away, or lose small accounts while the big ones expand. Track user retention to grade the product and revenue retention to grade the business, and read them side by side.
What Is a Good Retention Rate? (SaaS Benchmarks)
Benchmarks depend on audience, price point, usage frequency, and — as the previous chapter showed — which definition of retention is being used. With those caveats, the broad SaaS bands are consistent enough to be useful:
| Segment | Healthy | Excellent | Notes |
|---|---|---|---|
| B2B SaaS — monthly user retention | > 98% | > 99% | Annual retention above 90% is broadly considered sustainable |
| B2C SaaS — monthly user retention | 93–97% | > 97% | Lower tolerance is normal: cheaper acquisition, less sticky relationships |
| SMB-focused — net revenue retention | > 90% | > 100% | Small accounts churn more; volume smooths the number |
| Mid-market — net revenue retention | > 100% | > 110% | Longer contracts and deeper integrations hold accounts |
| Enterprise — net revenue retention | > 110% | > 120% | Best-in-class enterprise SaaS grows mostly through expansion |
Three cautions before comparing yourself to any table, this one included. First, check the measurement matches — N-day vs unbounded, user vs revenue, monthly vs annual all produce different numbers from the same business. Second, early-stage products naturally retain less; benchmarks describe products past product-market fit. Third, the benchmark that actually drives decisions is your own history: retention rising cohort over cohort is success at any absolute level, and retention falling is a warning at any absolute level.
Retention Examples: What It Looks Like in Practice
Definitions stick better with numbers attached. Here are three worked examples of retention measured the way real teams measure it:
1. Monthly user retention — the account view
A B2B SaaS starts March with 1,500 customers. During the month it signs 200 new ones and ends with 1,650 total, of which 1,455 are from the original group. Retention = 1,455 ÷ 1,500 = 97% — which also means 3% monthly churn, and (0.97)12 ≈ 69% of any cohort still present after a year. Good for self-serve SMB; worrying for enterprise.
2. Day-N cohort retention — the product view
1,000 users sign up in the first week of April. 400 come back at least once the next day (day-1 retention: 40%), 220 are active in day 7 (22%), and from day 30 onward the cohort stabilizes around 150 (15%). Plotted over time this is the classic retention curve: a steep early cliff, a slope, and — crucially — a flattening plateau. The plateau is the pass mark: a curve that flattens above zero means a real retained core exists; a curve that slides to zero means the product has no floor, and no acquisition spend can fix that.
3. Net revenue retention — the business view
A company begins the quarter with $100,000 MRR from existing customers. It loses $5,000 to cancellations and $2,000 to downgrades, but existing customers add $12,000 in upgrades and seats. NRR = (100,000 − 5,000 − 2,000 + 12,000) ÷ 100,000 = 105% — the base grew 5% with zero new sign-ups. Note the same quarter's gross revenue retention is 93%: same company, both numbers true, different questions answered.
Reading the examples together. Notice that each example could look "fine" while another flashes red — solid NRR can coexist with a decaying day-N curve, because today's revenue reflects decisions users made months ago. User retention is the leading indicator; revenue retention is the lagging one. When they disagree, believe the leading one.
What Actually Drives User Retention
Retention is not produced by retention tactics. It is produced upstream, by a handful of forces that decide — usually within the first days — whether a user's visit becomes a habit. These are the five that matter most:
-
Fast time-to-value
The strongest predictor of long-term retention is how quickly a new user experiences the product working for them. Every extra step between sign-up and first value bleeds a slice of the cohort — which is why user onboarding is retention work in disguise, and why activation is the metric that predicts retention before retention can be measured.
-
Reaching the aha moment
Every product has an action that separates users who stay from users who vanish — the aha moment. Users who reach it populate your retention plateau; users who don't were only ever visiting. Finding yours and shortening the path to it is the single highest-leverage retention project.
-
A habit loop, not a one-off win
One good session retains nobody. Retention comes from the product attaching itself to something recurring — a weekly report, a daily standup, an inbound customer message — so returning is triggered by the user's own workflow. Sustained user engagement is the visible evidence of that loop forming.
-
Continuous value discovery
Users judge the price against the value they know about, and most know a fraction of it. Ongoing feature discovery — contextual tooltips, in-app feature announcements — keeps deepening product adoption, and every newly discovered capability is a fresh reason to stay.
-
A feedback loop that catches problems early
Most users who leave never complain first. An in-app pulse like NPS surfaces the quietly dissatisfied while their problem is still fixable — every detractor you win back is a retained user acquisition never had to replace.
How to Improve User Retention: Where to Start
Improving retention is its own discipline — we wrote a full playbook in how to increase user retention — but the short version has a clear order of operations, because retention work has a steep leverage gradient: the earlier in the user journey you intervene, the more of the cohort is still there to save.
- Instrument first. Set up cohort retention tracking so every change you ship can be verified against the curve it was supposed to bend. Without cohorts, retention work is guesswork with extra steps.
- Fix the first session. Replace the empty first screen with a focused product tour that walks each new user to one real outcome, backed by an onboarding checklist that carries momentum into sessions two and three — the steepest part of the curve is where checklists earn their keep.
- Drive everyone to the aha moment. Identify the action your retained users share, and rebuild onboarding as the shortest path to it.
- Keep value visible. Announce what you ship in-app, and use contextual tooltips to surface the features each user hasn't touched yet — discovery is retention fuel for users past their first week.
- Listen before they leave. Run in-app NPS, read the detractor verbatims, and close the loop on the fixable ones.
Every one of those levers ships without engineering time using Kompassify: no-code product tours, checklists, feature announcements, NPS surveys, and product analytics that report engaged, activated, and churning users out of the box — so you can watch each cohort's retention respond to what you shipped.
Retention vs Churn vs Engagement: Untangling the Metrics
Retention lives in a family of overlapping metrics, and conversations go better when everyone means the same thing:
| Metric | The question it answers | Relationship to retention |
|---|---|---|
| Retention | Who is still here after N days? | — the outcome the others explain |
| Churn | Who did we lose, and why? | The exact complement: 97% retention = 3% churn |
| Engagement | How deeply and often do users interact? | The leading indicator: fading engagement is tomorrow's lost retention |
| Activation | Did the new user reach first value? | The gatekeeper: unactivated users almost never retain |
| Adoption | How much of the product's value is in use? | The deepener: broader adoption raises the retention plateau |
The sequence to remember: activation → engagement → retention. Activation opens the door, engagement builds the habit, retention is the habit persisting — and adoption and feedback loops keep raising the plateau after that. Teams that try to fix retention directly usually discover the real problem two steps upstream.
Measuring & Improving Retention: Do vs. Don't
A quick reference for keeping your retention work honest and effective.
✅ Do
- Subtract new users — measure the starting cohort only
- Define "active" as a core action, not a login
- Match the retention window to natural usage frequency
- Track retention per cohort and per segment
- State which definition (N-day, rolling, unbounded) every chart uses
- Read user retention and revenue retention side by side
- Fix activation and the first session before anything else
- Verify every onboarding change against the cohort curve
- Benchmark against your own earlier cohorts first
❌ Don't
- Let total active users stand in for retention
- Compare your N-day number to someone's unbounded benchmark
- Blend all cohorts into one number and call it insight
- Judge a monthly-use product on daily retention
- Celebrate NRR while the day-N curve decays underneath it
- Treat retention as a late-funnel problem — it's decided early
- Ship discounts as a retention fix — they defer churn, expensively
- Chase resurrection campaigns while onboarding still leaks
- Change your definition of "active" silently between quarters
Ready to Bend Your Retention Curve?
Kompassify combines the retention levers in one no-code platform: product tours, onboarding checklists, in-app announcements, NPS surveys, and product analytics that report engaged, activated, and churning users. GDPR compliant, EU-hosted, and free for under 100 monthly active users.
Start for Free →Frequently Asked Questions
What is user retention?
User retention is the percentage of users who keep using your product over a given period — the share of a starting group that is still active after a day, a week, a month, or a year. It is the opposite of churn: a product with 97% monthly retention has 3% monthly churn. Retention matters because it measures the only thing that compounds — users who stay, form habits, expand, and refer — whereas acquisition has to be re-bought every month.
How do you calculate retention rate?
Retention rate = ((users at the end of a period − new users acquired during it) ÷ users at the start of the period) × 100. Example: you start the month with 2,000 users, gain 300 new ones, and end with 2,240 total — of which 1,940 are from the original group. Retention is 1,940 ÷ 2,000 = 97%. Subtracting new users is the crucial step: counting them would let acquisition inflate the number and hide real losses.
What is a good retention rate for SaaS?
For B2B SaaS, monthly user retention above 98% is healthy and above 99% is excellent; annual retention above 90% is broadly considered sustainable. B2C products typically run lower — 93–97% monthly — because relationships are less sticky. For revenue, net revenue retention above 100% means your existing customers grow in value even with zero new sign-ups; best-in-class SaaS runs 110%+. Because definitions vary, your most reliable benchmark is your own earlier cohorts.
What is the difference between retention and churn?
Retention and churn are mathematical complements — two views of the same number. A product with 97% monthly retention has 3% monthly churn. The framing differs: retention asks who is still here and what makes them stay, which drives strategy and habit-building; churn asks who was lost and why, which drives leak-fixing. Mature teams use both — retention as the strategy, churn as the alarm system.
What is the difference between N-day, rolling, and unbounded retention?
N-day (classic) retention counts a user as retained only if they return on exactly day N — strict, and standard for high-frequency products. Rolling (bracket) retention counts users who return on day N or within a window around it, smoothing out usage-frequency noise. Unbounded retention counts users who return on day N or any day after it, which suits low-frequency products where a monthly visit is healthy. None is wrong — but they produce very different numbers, so never compare retention figures without checking which definition was used.
What is the difference between user retention and revenue retention?
User retention counts people or accounts that stay; revenue retention counts the money that stays. Gross revenue retention (GRR) measures recurring revenue kept from existing customers excluding expansion — its ceiling is 100%. Net revenue retention (NRR) includes upgrades and expansion, so it can exceed 100%: existing customers then grow in value even with zero new sign-ups. User retention grades the product; revenue retention grades the business. Track both, because they can move in opposite directions.
How do you improve user retention?
Retention is decided early, so start where the curve is steepest: shorten the path from sign-up to first value with a focused product tour and an onboarding checklist, drive every user to the aha moment before they drift, then keep value visible with in-app feature announcements and contextual tooltips. Watch your retention curve per cohort to verify each change actually bends it, and run NPS surveys to catch dissatisfaction while it is still fixable. Kompassify combines these levers in one no-code platform, free for under 100 monthly active users.