Most teams find out about churn the same way: a cancellation notice arrives, someone updates a spreadsheet, and the account moves from "customer" to "lost" without much more investigation than that. By the time churn shows up as a monthly percentage on a board slide, the moment to actually understand what happened has usually already passed.

Churn analysis is the discipline of figuring out why customers leave before you try to fix it. It sits between the moment a customer cancels, which you capture in cancellation flow optimization, and the moment you build a systemic response, which is what churn reduction framework covers. Skip the analysis step and you end up building retention programs based on guesses instead of evidence.

Separating the Types of Churn

The first mistake in most churn analysis is treating churn as one number. It isn't. Different types of churn have different causes and require different fixes, and blending them together hides the pattern you're actually trying to find.

Voluntary versus involuntary churn. Voluntary churn is a customer actively deciding to cancel. Involuntary churn is a customer losing access because a payment failed, a credit card expired, or a billing error went unresolved. These have completely different root causes: voluntary churn is a product or value problem, involuntary churn is often a payments and dunning problem that has nothing to do with whether the customer wanted to stay.

Logo churn versus revenue churn. Logo churn counts the number of accounts that cancel. Revenue churn counts the dollar value lost. A single enterprise account canceling can move your revenue churn number dramatically while barely affecting logo churn, and the reverse is true for a wave of small-account cancellations. Reporting only one of these numbers can make your retention picture look better or worse than it actually is.

Gross churn versus net revenue churn. Gross churn measures revenue lost from cancellations and downgrades, full stop. Net revenue churn nets that loss against expansion revenue from existing customers, and it's possible to have healthy net revenue retention above 100% while still having a meaningful gross churn problem that expansion happens to be masking.

Full cancellation versus partial downgrade. A customer who downgrades from a $500/month plan to a $100/month plan hasn't churned in the traditional sense, but they represent a similar signal: something about the value they were getting no longer justified the price they were paying. Many churn analyses only count full cancellations and miss this earlier warning sign entirely.

The Formulas That Matter

Customer (logo) churn rate = (Customers lost during period / Customers at start of period) × 100

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

Net revenue churn rate = ((Revenue lost - Expansion revenue) / Revenue at start of period) × 100. A negative net revenue churn rate means expansion revenue more than offset losses, which is the goal most mature SaaS companies aim for.

Calculate these monthly and annually, and always specify which type you're reporting. According to CRV's SaaS churn rate benchmark data, median annual revenue churn for Series A companies ($1M-$10M ARR) sits around 12.5%, with top-quartile performers holding it below 5.48%. The same data shows a sharp divide by customer segment: SMB accounts (roughly $500-$5,000 ACV) see monthly logo churn around 4.1%, while enterprise accounts ($100K+ ACV) typically stay below 1% monthly. That gap exists because enterprise deals tend to have longer contracts, higher switching costs, and deeper integrations, all of which make cancellation a bigger decision to reverse.

If your churn rate looks unusually high or low compared to these benchmarks, segment by customer size before concluding anything. A blended churn number across SMB and enterprise accounts can look "average" while hiding a serious problem in one segment that a healthy number in the other segment is masking.

Cohort Analysis: Finding the Pattern Over Time

A single churn percentage tells you what happened last month. Cohort analysis tells you whether things are getting better or worse, and when in the customer lifecycle problems actually emerge.

Group customers by the month or quarter they signed up, then track what percentage of each cohort remains active at 3, 6, 12, and 24 months. Two patterns to look for:

Cohort curves that flatten out. If churn is heaviest in the first 90 days and then the curve flattens, meaning customers who survive the early period tend to stay long-term, your problem is concentrated in onboarding and early value delivery, not the product itself for established users.

Cohort curves that keep declining steadily. If customers keep leaving at a consistent rate no matter how long they've been a customer, that points to a more fundamental product or value problem that doesn't resolve itself once someone gets past onboarding.

Improving or worsening cohorts over time. Compare the 6-month retention of customers who signed up a year ago against customers who signed up this quarter. If the newer cohort curve is meaningfully better, whatever product or onboarding changes you made in between are working. If it's worse, something regressed, and cohort analysis is often how you catch that before it shows up in an aggregate churn number.

Root Cause Analysis: Why, Not Just How Many

Once you know the size and shape of the problem, the harder work is understanding why. Four sources of evidence matter most.

Exit surveys and cancellation reasons. Direct feedback from the cancellation flow itself, but treat self-reported reasons with some skepticism. Customers often cite price as the reason they're leaving because it's the easiest, least confrontational answer, when the real issue was that they never found enough value to justify any price.

Usage decline before cancellation. Look at product usage in the weeks and months leading up to cancellation. A steady decline in login frequency, feature usage, or active seats is often a more honest signal than the stated cancellation reason, and it's usually detectable well before the customer formally cancels, which is the entire premise behind churn risk detection.

Support ticket patterns. Customers who churn often show a distinct support pattern beforehand: either a spike in tickets around a specific problem that never got resolved, or the opposite, complete disengagement with no tickets at all because they've already mentally moved on.

Segment and cohort comparison. Compare churned customers against retained customers on onboarding completion, feature adoption breadth, initial use case, and account tier. Patterns that show up consistently across churned accounts, and don't show up in retained accounts, are strong candidates for root causes rather than coincidences.

The goal of root cause analysis isn't to find a single universal reason customers leave. It's to find the two or three most common, most fixable patterns, since most SaaS companies discover that a small number of root causes explain a disproportionate share of total churn.

Building a Repeatable Churn Analysis Process

Treat churn analysis as an ongoing operating rhythm, not a one-time investigation.

Monthly: calculate and segment the core numbers. Logo churn, revenue churn, and net revenue churn, broken out by customer segment (SMB, mid-market, enterprise) and by product tier if you have multiple.

Quarterly: run cohort analysis and compare trends. Are cohorts improving over time? Is the pattern of early versus late-stage churn shifting? This cadence gives enough new data to spot real trends without overreacting to a single noisy month.

Ongoing: feed cancellation and usage-decline data into a shared repository. Exit survey responses, usage decline flags, and support ticket patterns should land somewhere product, customer success, and leadership can all review together, rather than living in separate tools that never get cross-referenced.

After every major product or pricing change: watch for shifts in the churn pattern. A pricing change, a feature deprecation, or a major UI overhaul can shift churn behavior within weeks. Reviewing churn data specifically after these changes catches problems faster than waiting for the next scheduled quarterly review.

The output of good churn analysis should be specific and actionable: not "churn is 8%" but "SMB accounts that never complete onboarding within the first two weeks churn at three times the rate of accounts that do, and that gap has widened over the last two quarters." That level of specificity is what turns analysis into an actual retention program.

Frequently Asked Questions About Churn Analysis

What's the difference between gross churn and net revenue churn?

Gross churn measures revenue lost from cancellations and downgrades on its own. Net revenue churn nets that loss against expansion revenue from existing customers, so a company can show negative net revenue churn (meaning expansion outpaced losses) while still having a real gross churn problem worth investigating separately.

Why does customer segment matter so much in churn analysis?

Churn rates differ dramatically by segment. SMB accounts often see monthly logo churn several times higher than enterprise accounts, because enterprise deals typically involve longer contracts, higher switching costs, and deeper integrations. Blending segments into one churn number can hide a serious problem in one group behind a healthy number in another.

Should I trust the reasons customers give when they cancel?

Treat stated cancellation reasons as one data point, not the full picture. Customers frequently cite price because it's an easy, non-confrontational answer, even when the underlying issue was that they never reached enough value to justify paying anything. Cross-reference stated reasons against usage decline data for a more complete picture.

How often should a SaaS company run churn analysis?

Calculate and segment core churn numbers monthly, and run deeper cohort analysis quarterly to spot trends that a single month of data can't reveal. Additionally review churn patterns after any major pricing or product change, since those can shift churn behavior within weeks.

What's the most useful output of a churn analysis process?

A specific, actionable finding tied to a segment and a behavior, such as identifying that accounts failing to complete onboarding within two weeks churn at a much higher rate than accounts that do. That level of specificity is what lets a churn reduction program target the actual cause instead of guessing.

Related resources:

About the author

Tara Minh

Senior Operations & Growth Strategist

Tara Minh is Senior Operations & Growth Strategist at Rework, helping B2B SaaS leaders scale without breaking their teams. With 8+ years in revenue operations and process optimization, Tara turns messy workflows into systems people actually follow. Readers get practical frameworks they can use to cut waste, align teams, and grow on purpose.

View full profile LinkedIn