What is Churn? Definition, Formula, and Examples for Business

Churn is the percentage of customers, users, or revenue lost from a business over a defined time period. The term comes from the idea of customers "churning" in and out of a product, and it's become the standard vocabulary in SaaS, subscription, and consumer app businesses.

Churn is the rate at which customers stop using a product, service, or subscription over a given period.

The word "churn" gets used loosely, and that's where confusion starts. Depending on the context, someone saying "our churn is 5%" might mean customer churn, revenue churn, or user churn — three different things with three different formulas and business implications. We'll cover each type in the next section.

Churn matters because it's the denominator of sustainable growth. You can spend heavily on acquisition and still go backwards if customers leave faster than you replace them. The math is brutal: if a product adds 100 new customers a month and loses 50 to churn, it's running flat no matter how aggressive the top-of-funnel spend gets.

The gap between top performers and everyone else is enormous. Amplitude's 2025 Product Benchmark Report found that the top 10% of products retain roughly four times more users at month 1 than the median — 26% versus 6.5%. That gap compounds. A product retaining a quarter of its cohort after the first month has fundamentally different unit economics than one retaining less than a tenth.

Churn is also the inverse of retention rate, which is why the two metrics get discussed together. If your month-1 churn is 30%, your month-1 retention is 70%. They measure the same thing from opposite angles. Most product teams end up tracking retention curves rather than churn rates, because curves make cohort patterns easier to see over time.

Churn isn't one metric. It's a family of metrics that measure different things. Product and finance teams typically track several in parallel.

• Customer churn is the percentage of customers who stop doing business with a company in a given period. A SaaS product with 1,000 customers that loses 50 the next month has 5% monthly customer churn.
• Revenue churn (MRR churn) is the percentage of recurring revenue lost from existing customers, whether through cancellations or downgrades. A company with $100K MRR that loses $5K from canceled contracts has 5% MRR churn. Revenue churn often diverges from customer churn, especially when larger accounts behave differently than smaller ones.
• User churn is the percentage of active users who stop engaging with a product, independent of account status. This matters most in freemium or consumer apps where one account can have many users, or where "active" is defined by behavior rather than billing.
• Voluntary churn happens when a customer actively chooses to leave — cancels, switches to a competitor, or decides the product no longer fits their needs.
• Involuntary churn happens when a customer leaves because of external factors like a failed payment, an expired credit card, or a deactivated account. A meaningful share of involuntary churn is recoverable with dunning, payment retry, and card update flows.
• Gross vs. net churn measures the same loss from two angles. Gross churn counts only revenue or customers lost. Net churn subtracts expansion revenue — upgrades, seat growth, add-ons — from the losses. Net churn can go negative when expansion outpaces losses, which is the hallmark of a healthy subscription business.

The basic formula for customer churn is straightforward.

Churn rate = (Customers lost during period / Customers at start of period) × 100

Worked example. A SaaS product starts the quarter with 500 customers and loses 25 during the quarter. Churn rate = (25 / 500) × 100 = 5% quarterly churn.

The formula for revenue churn follows the same shape, substituting dollars for customer counts.

MRR churn rate = (MRR lost during period / MRR at start of period) × 100

To run the numbers on your own product, plug them into Amplitude's churn rate calculator.

A couple of things to watch for when you calculate churn:

• Pick a period that matches the cadence of your business. Monthly churn is standard for SaaS; weekly churn is common for consumer apps; annual churn is standard for enterprise contracts with yearly renewals. Don't mix periods without converting.
• Short-period churn compounds. A 5% monthly churn rate doesn't mean you lose 60% of customers over a year — it compounds to roughly 46% annual churn. The monthly-to-annual math matters for forecasting.
• Handle mid-period additions carefully. If you add new customers during the measurement period, you need to decide whether they're in the denominator. Most teams use the starting customer count as the denominator and track cohorts separately for new customers.
• Anniversary vs. rolling-window measurement. Anniversary churn counts whether a customer was active on day 30 (or day 90, or day 365). Rolling-window churn counts any cancellation in the last 30 days. They tell different stories.

There's no universal "good" churn rate. The honest answer is that it depends on your industry, business model, customer segment, and cohort maturity — and the most useful comparison is your own trend line over time, not an external benchmark.

That said, rough ranges help teams orient. Annual churn of 5–7% is often cited as acceptable for SMB-focused B2B SaaS; enterprise SaaS targets well under 5%. Monthly churn in consumer mobile apps routinely runs 20–40% in the first month, especially for free-tier users. These are starting points, not targets.

Amplitude's 2025 Product Benchmark Report gives concrete context on the size of the gap between top performers and everyone else. At month 3, the top 10% of products retain 18.5% of users, compared to 3.8% for the median. That works out to roughly 81.5% month-3 churn for top performers versus 96.2% for the median — a nearly 5x gap in the share of a cohort still around three months in.

Two things to take from that data. First, churn is much higher across the industry than most teams assume. Second, the difference between a top-decile product and an average one isn't incremental. It's categorical.

Customers churn for specific, identifiable reasons, and most of them trace back to the earliest moments of the relationship. The most common drivers include:

• Poor onboarding and no clear "aha moment" in the first session
• Product-market misfit for the specific segment or use case
• A better competitor offering or a more compelling alternative
• Pricing or perceived-value mismatch after the initial honeymoon
• Support failures, reliability problems, or bugs that block core workflows
• Natural life-cycle endpoints where the customer no longer needs the product

Of these, onboarding is the most important — and the most controllable. Amplitude's 2025 Product Benchmark Report found that 69% of top performers in 7-day activation were also top performers in 3-month retention. The connection is strong enough that early activation is the single best predictor of long-term retention in the data. Teams that want to reduce churn should start where customers form their first impression of whether the product is worth their time.

For the full playbook on diagnosing and reducing churn, see Amplitude's guide to churn analysis.

Reducing churn is a loop, not a one-time fix. The pattern that works across product teams is: detect who's leaving, understand why, intervene where it matters, and measure whether the intervention worked. Amplitude Analytics supports each step of that loop with a shared data foundation, and the practical setup for tracking churn rate in a product analytics tool starts with defining what "churned" means for your business.

Start by segmenting churned users into behavioral cohorts — by plan, acquisition channel, feature usage, or activation pattern. Cohort analysis shows which customers leave and which stay, and which behaviors separate the two groups. A common finding: customers who complete a specific action in their first session retain at dramatically higher rates than those who don't.

Next, watch the behavior before the churn. Session Replay gives you the qualitative signal that quantitative analysis can't — the exact moment a user got stuck, the feature they tried to find, the error they hit. Combining cohort data with replay turns a chart into a cause.

Once you know who's at risk and why, build a prediction model. Amplitude's AI Agents use behavioral data to flag at-risk users before they leave, which gives the team a window to act. The window is usually short — most churn happens in silence.

Then test interventions. Use Feature Experimentation to run controlled tests on onboarding changes, pricing variants, or feature exposure to at-risk cohorts. And close the loop in-product with Guides and Surveys, which lets you reach at-risk users with contextual messages or targeted offers before they churn.

The workflow looks like: run the cohort analysis, watch the sessions of users who left, build a predictive cohort of at-risk accounts, launch an experiment on a retention intervention, and measure the lift. That's how teams move churn as a KPI rather than just reporting it.

Several metrics get confused with churn. Here's how they relate:

Churn vs. retention

Retention is the inverse of churn. Month-1 churn of 30% equals month-1 retention of 70%. Most product teams track retention curves because curves show cohort behavior over time, while churn rates collapse that information into a single number.

Churn vs. attrition

Often used interchangeably. "Attrition" is more common in HR (employee attrition) and traditional B2B contexts. "Churn" is standard vocabulary in SaaS, subscription, and consumer product businesses. Both measure the rate at which a group is lost over time.

Churn vs. bounce rate

Bounce rate is a single-session metric — did the visitor leave the site after viewing one page? Churn is a longer-horizon measure of whether customers stop using the product over days, weeks, or months.

Churn vs. dormancy

A dormant user hasn't churned yet. They've stopped engaging but haven't canceled, unsubscribed, or formally left. Dormancy often precedes churn, which is why many teams build dormant-user cohorts as an early-warning signal.

Churn is one metric. Retention, activation, and engagement tell the rest of the story — and they're all connected. Amplitude helps product teams see how the pieces fit together and find the specific behaviors that keep customers around.