Customer Health Scoring: A Simple Framework for Predicting Churn

By the time a customer clicks “Cancel,” it’s usually too late. The decision was made weeks ago. The real question is: can you spot the warning signs early enough to do something about it?

That’s what customer health scoring does. It assigns a 0-100 score to every customer based on behavioral signals, so you can see who’s thriving and who’s about to leave.

The 4-Signal Framework

After researching how enterprise tools like Gainsight approach this (and stripping away the complexity that doesn’t apply to small SaaS), we landed on 4 signals that actually predict churn:

1. Payment Recency (40% weight)

When did the customer last pay you? This is the strongest signal. A customer who paid yesterday is almost certainly still engaged. A customer whose last payment was 60 days ago? Something is wrong.

Paid in the last 7 days: 40/40 points
Paid in the last 30 days: 30/40 points
Paid in the last 60 days: 15/40 points
60+ days ago: 0/40 points

2. Transaction Frequency (25% weight)

How often does revenue come in from this customer? A customer with 12 monthly payments is far more stable than one with a single annual payment — even if the dollar amounts are similar.

10+ transactions: 25/25 points
5-9 transactions: 18/25 points
2-4 transactions: 10/25 points
1 transaction: 5/25 points

3. Revenue Trend (20% weight)

Is this customer spending more or less over time? Compare last 30 days to the 30 days before that. Growing revenue = healthy. Declining revenue = risk.

Revenue growing: 20/20 points
Revenue stable: 15/20 points
Revenue declining: 5/20 points

4. Payment Reliability (15% weight)

Has this customer had any failed payments? Payment failures are one of the strongest churn predictors — both because of involuntary churn (expired cards) and because they signal disengagement.

No failures ever: 15/15 points
1 failure: 8/15 points
2+ failures: 0/15 points

What the Scores Mean

80-100: Healthy. These customers are engaged, paying regularly, and growing. Focus on expansion opportunities.
50-79: Watch list. Something might be off. Worth a proactive check-in.
20-49: At risk. Multiple warning signs. Reach out before they cancel.
0-19: Critical. Likely already decided to leave. Retention effort needed immediately.

Why This Beats Gut Feel

Most solo founders track churn reactively — someone cancels, you look at their history, you think “oh, I should have seen that coming.” Health scoring flips this. Instead of reacting to cancellations, you’re proactively identifying risk.

Even with just 20-30 customers, sorting by health score immediately shows you where to focus your time. That’s the difference between losing a customer and saving one.

Automate It

You can build this in a spreadsheet, but every formula breaks the moment your data changes. RevPane calculates health scores automatically for every Stripe customer, updates them hourly, and flags at-risk customers before they churn.

Try RevPane free — customer health scoring included in the Pro plan.