Okay, can we talk about customer health scores for a second? Because I swear, half the time I hear people talk about them, they sound like they’re quoting some SaaS handbook they barely read. Like, “we need to track engagement metrics and assign a health rating based on feature usage and blah blah blah.” Cool. But what does that mean in the day-to-day chaos of trying to keep your clients from ghosting you?
Here’s the thing no one says out loud: churn doesn’t always come with a warning.
Sometimes, you wake up, check your dashboard, and the account is canceled. There is no explanation. Just a ticket from support and a Slack message from someone panicking. And you’re left digging through emails, wondering what you missed.
You didn’t miss anything obvious. But that’s the problem. The signs weren’t loud; they were quiet. Fewer logins. Delayed replies. A skipped QBR. That one weird comment during a call that felt… off.
Health scores, when done right, not textbook right, but reality-right, help you catch those whispers before they become exits. But they have to be real. Not vanity metrics. Not just “number of logins per week” or “open rate on the last email blast.” Those help, sure. But they don’t tell the whole story.
A real health score? It blends numbers with gut. Data with judgment. Structure with common sense.
I used to rely 100% on usage stats. That was the model. Login frequency, feature activity, and billing consistency. If those numbers dropped, we flagged the account.
Except for one of our best clients (and I mean, top tier)—they barely logged in once a month, maybe. But they loved us. They were just… hands-off. Their team was busy. They used our stuff when they needed to and never complained.
Meanwhile, another account looked healthy, tons of usage, feedback, open support tickets… and they churned with almost no warning. Said they felt overwhelmed. Said we weren’t listening. Said they needed to simplify.
That’s when I started blending qualitative check-ins into our health model. How responsive are they in meetings? What’s their tone like in Slack or email? Do they ask about long-term planning or complain about bugs?
We started scoring all of it. And yes, it got messy. Sometimes subjective. But better messy and early than perfect and too late.
Health Scores Are Expansion Gold
Here’s the kicker: health scores aren’t just for preventing churn. They’re expansion gold.
Think about it: your happiest customers? The ones giving good signals engaged, asking more profound questions, trying new features, and looping in other teams? Those are your expansion targets. Please don’t wait for them to raise their hand. You raise yours.
We built a “green zone” list of clients with high health scores and strategic potential. We review it every quarter. We look for signs of new hires, new initiatives, anything that says, “They’re growing, and we can grow with them.”
I’ll give you an example. One client had been stable for months, nothing exciting. But then they started forwarding our product updates to other departments. We flagged this. I reached out not to sell but to say, “Hey, I noticed you’ve been sharing stuff. Want to talk about where this might go?”
Two weeks later? Multi-team rollout. Doubled the contract.
But again, this only worked because we weren’t just staring at usage charts. We were watching behavior, conversations, patterns, and people.
How to Get Started
Let me be real for a second: building a health score model isn’t fun initially. You’ll argue about weighting. You’ll obsess over scoring logic. You’ll probably throw out your first version after three months. That’s fine. That’s normal. What matters is starting, even if it’s scrappy.
And if you’re not sure where to start, here’s what helped me:
- Quantitative Inputs: logins, feature usage depth, support ticket velocity, NPS, billing status
- Qualitative Inputs: responsiveness, attitude in meetings, tone in messages, executive engagement
- Contextual Signals: org changes, layoffs, new hires, budget reviews, tech stack changes
Not every score is perfect. But they’re directional. That’s the key. It’s not about predicting the future perfectly. It’s about not being surprised when the future shows up.
Make It Collaborative
Oh, and one more thing. Share health scores with your team. Like, openly. I used to guard that stuff like it was a secret. Dumb move. Now it’s part of our client review rhythm. Everyone sees it. Everyone contributes to it. Support flags sentiment shifts. Sales flags upsell opportunities. Success flags blockers. It’s collaborative now, not just one person pulling strings in a dark room.
If you get this right, churn starts feeling less like a shock and more like a signal you caught early. And expansion? It becomes proactive, not pushy, just aligned.
When you see your clients and tune into their actions, not just what the metrics say, they tend to stick around. And they tend to grow.
You don’t need a perfect score. You need a real one.
About The Author
