Why Your Churn Rate Is a Lagging Indicator (And What to Watch Instead)

Your monthly churn rate tells you how many customers left. It does not tell you why, and it definitely does not tell you who is about to leave. By the time that number ticks up from 3% to 5%, the damage is already done.

Churn is an autopsy report. You need vital signs.

Why churn is always late

A customer who cancels in March probably made that decision in January. Something broke — they stopped getting value, a competitor showed up, their champion left the company, or they simply forgot you existed.

The cancellation is the final act. The leading indicators showed up weeks or months earlier, and most companies never tracked them.

The five leading indicators worth watching

### 1. Login frequency drop

If a daily user becomes a weekly user, something changed. Track the 7-day and 30-day rolling average of logins per account. A 40% drop in engagement over two weeks is a reliable early warning. At one B2B SaaS company we worked with, accounts that dropped below 3 logins per week had a 4x higher churn rate within 90 days.

### 2. Feature adoption depth

Accounts that use only one feature churn at 2-3x the rate of accounts using three or more. Track the number of distinct features used per account per month. If an account has been live for 60 days and still only uses the core feature, that is a red flag — not a success story.

### 3. Support ticket sentiment and frequency

A spike in support tickets is obvious. What is less obvious: the customer who stops submitting tickets entirely. They have not figured it out — they have given up. Track both the volume trend and the silence. Accounts that go from active support conversations to zero contact in a 30-day window deserve a proactive check-in.

### 4. Time to value (TTV)

How long does it take a new customer to reach the moment where they say "this is worth it"? If your average TTV is 14 days but some accounts are still not activated after 30, those accounts are at serious risk. Every day past your benchmark TTV, churn probability increases.

### 5. Expansion signals (or lack thereof)

Healthy accounts grow. They add users, upgrade plans, or increase usage. Accounts that flatline — same seat count, same plan, no usage growth for six months — are coasting. Coasting accounts are one budget review away from cancellation.

How to build a churn risk score

You do not need a machine learning model to start. A simple weighted score works:

• Login frequency below threshold: +2 points
• Feature adoption below 3 features: +2 points
• No support contact in 30 days: +1 point
• Past TTV benchmark with low activation: +3 points
• No expansion activity in 90 days: +1 point

Any account scoring 5 or above gets flagged for immediate outreach. This scoring model can live in a spreadsheet, but it works far better as an automated alert in your BI layer.

The shift that matters

Stop reviewing churn monthly. Start reviewing churn risk weekly. The companies that hold churn below 2% are not better at win-back campaigns. They are better at catching problems before the customer ever thinks about leaving.