Churn Analysis: How to Use Analytics to Keep More Customers

By Emily Redmond, Data Analyst at Emilytics · April 2026

TL;DR: Analyze churn by cohort and segment. Track activation, feature usage, and logins. Identify early warning signals. Healthy SaaS has 3–5% monthly churn.

Churn is the silent killer of SaaS companies. You can grow MRR 20% month-over-month and still die if your churn is 22%.

Most founders can't see churn coming. Their analytics dashboard shows "5% monthly churn" but hides the real story: week 1 cohort churns at 25%, but month 6 cohort churns at 2%. Or: free plan users churn at 12%, paid users at 3%. The blended number lies.

This guide shows you how to see churn actually coming, which segments are at risk, and what to do about it.

The Churn Formula

Monthly churn % = Customers lost in month ÷ Customers at month start × 100

Example: 100 customers on Jan 1, 5 cancel in January = 5% monthly churn.

How to interpret it:

• 5% monthly = 60% annual churn (unsustainable for early-stage SaaS)
• 3% monthly = 35% annual churn (acceptable for mature SaaS)
• 1% monthly = 11% annual churn (enterprise SaaS target)

But here's the catch: Blended churn hides your real problems.

Cohort Churn: Where the Real Story Lives

Blended churn masks cohort decay. You need to see churn by signup cohort.

Set up in GA4:

1. Go to Explore → Retention
2. Select Daily active users or create a custom metric for "active accounts"
3. Cohort: Acquisition date (weekly or monthly)
4. Retention: Days or weeks since acquisition

You'll see a matrix like this:

What to look for:

• Is each new cohort retaining better than the last? (Improvement = product getting better)
• Or worse? (Decline = onboarding getting worse or product diluting)
• Where's the cliff? (Week 1 drop of 50% suggests onboarding failure; week 4 drop suggests activation issue)

💡 Emily's take: I analyzed a SaaS with "5% blended churn" and found month-1 users churning at 22%. Month 6 users: 1%. The company thought they had a product problem. Actually, they had an onboarding problem—users were getting better at activating. Totally different fix.

Segment Churn by Feature Usage

High-level churn numbers hide what's actually killing retention.

In GA4, track these behaviors:

1. Logins in last 7 days - Users who log in regularly rarely churn
2. Key feature usage - Users who use your core feature stay; users who don't, leave
3. Time to activation - Users who activate in day 1 vs day 14 have different churn rates
4. Team size (if B2B) - Users who invite teammates churn less
5. Data migration (if applicable) - Users who import historical data stay longer

Create custom events in GA4 for each:

// User logged in today
gtag('event', 'user_login');

// User used core feature (e.g., ran a report)
gtag('event', 'core_feature_used', {
  'feature_name': 'report_generation'
});

// User invited a teammate
gtag('event', 'team_member_invited');

Then build a custom report:

This tells you:

• Logins are the best churn predictor. Users who stop logging in are churning.
• Feature usage is the second predictor. Users who ignore your core feature don't see value.
• Activation speed matters. Faster activation = higher retention.

Early Warning Signals: Predict Churn Before It Happens

Don't wait until someone cancels. Catch churn signals 4 weeks before.

Track these leading indicators:

1. Declining logins - User logged in 3x last week, 1x this week
2. Feature usage drop - User ran 5 reports last month, 1 this month
3. Support ticket tone - User submitted complaint/question (check support data)
4. Incomplete onboarding - User created account but never completed setup
5. Free plan user who didn't upgrade after trial - 30 days post-trial, still free

Set up alerts in GA4:

Create a custom report with dimensions and metrics like:

• User ID
• Days since last login
• Last 7-day login count
• Last 30-day feature usage
• Activation status (yes/no)

Export to a sheet weekly. Flag users with:

• Zero logins in 14 days
• Feature usage dropped 80%+
• Activation = false after 30 days

Send your CS team a list of at-risk users to proactively reach out.

💡 Emily's take: Proactive outreach to at-risk users reduces churn by 20–30% on average. But you have to find them first. Most founders wait for the cancellation request.

Calculate Cohort Churn Rate (Precisely)

Here's how to calculate churn rate for each cohort scientifically.

Month-over-month cohort churn:

1. Take the cohort of users who signed up in January
2. How many were still active in February? (That's your January cohort February retention)
3. How many were active in March? (January cohort March retention)
4. Month-over-month churn = 100% - retention rate

Example:

• January cohort: 500 users sign up
• End of February: 475 still active (95% retention, 5% churn)
• End of March: 450 still active (90% retention, 10% churn)
• January's March churn = 10%

Plot over time:

• January cohort churn: 5% month 1, 10% month 2, 14% month 3
• February cohort churn: 6% month 1, 11% month 2, 15% month 3
• March cohort churn: 7% month 1, 12% month 2, 16% month 3

If each cohort is churning worse than the previous cohort, your product is getting less sticky. Fix it immediately.

Common Churn Mistakes

Mistake 1: Only tracking blended churn

You hide the real problems when you blend. Always segment by cohort, plan type, and usage behavior.

Mistake 2: Not separating voluntary from involuntary churn

Payment failures (involuntary) are different from cancellations (voluntary). Track separately. Involuntary churn is usually fixable with better billing retry logic.

Mistake 3: Confusing free user churn with paid user churn

Free users churn at 15%. Paid users churn at 3%. The blend is useless. Track them separately.

Mistake 4: Not looking for the activation-churn correlation

Activation is the strongest predictor of churn. Users who activate in-trial have 3x lower churn than those who don't. Make this connection in your analytics.

Mistake 5: Waiting for monthly data to take action

Check churn weekly. A cohort that looks good on day 7 might be churning hard by day 21. Early detection = earlier action.

Benchmarks by SaaS Type

Frequently Asked Questions

Q: What's the biggest churn driver?

A: Activation. Users who don't activate in the first week churn at 15x the rate of those who do. Fix activation first.

Q: Should I segment churn by pricing tier?

A: Absolutely. Pro plan users usually churn less than Starter plan. That tells you your higher tiers are stickier (good for retention, but might mean Starter is overselling or serving the wrong audience).

Q: How do I know if my churn is improving?

A: Cohort analysis. If January's cohort churns at 8% month-over-month and April's cohort churns at 5%, you're improving. If it's the opposite, you're degrading.

Q: What should I do if month-1 churn is really high?

A: Look at activation and feature usage. 70% of new users not logging in after week 1? Your onboarding is broken. Users logging in but not using core feature? Product clarity issue.

Q: How often should I check churn?

A: MRR weekly (impacts cash flow). Churn monthly (it's a trend metric). Cohort retention quarterly (it's slower to move).

The Bottom Line

Churn is the ceiling on your growth. Blended churn hides problems. Segment by cohort, plan type, and usage behavior. Use early warning signals (declining logins, feature usage drops) to reach out before users leave.

Fix activation and onboarding. That's where 70% of SaaS churn happens.

Emily Redmond is a data analyst at Emilytics — AI analytics agent watching your GA4, Search Console, and Bing data around the clock. 8 years experience. Say hi →