A few years ago, we rolled out our first customer health score system.
It was supposed to be a game changer. Finally, we have a scalable way to understand customer sentiment, anticipate risks and drive renewals with confidence.
But it didn’t work.
Our team didn’t trust the data. Accounts would churn while still showing up as “green.” CSMs were frustrated. Leadership was blindsided. And our renewal forecast was way off.
That’s when we realized we were facing what’s known as the “watermelon effect.” 🍉
On the outside, everything looks green. But inside—where it really matters—it’s red.
Customers were logging in, opening emails, maybe even engaging in QBRs. But behind the scenes? Their champion had left. Or they were shopping for alternatives. Or worse, our solution was no longer aligned with their goals.
That’s when we knew we needed something more. So we built our Churn Risk Indicator system.
It’s complements our health score but serves a different purpose. While health scores help us scale, automate and spot trends across the customer base, churn risk indicators help us be laser-focused on high-risk accounts, guide save plans and improve forecasting accuracy.
The result? Our team now has a clearer picture of risk, a better handle on forecasting and stronger alignment on where to focus.
Let’s break down the difference.
Health Scores: The Customer’s “Vitals”
Think of your health score as your customer’s check-up report.
It gives you a quick, standardized snapshot of how they’re doing—at least on paper. Health scores are like vitals: they won’t tell you everything, but they’re essential for spotting early signals and tracking trends at scale.
Most health scores are calculated by aggregating a mix of data points, such as:
- Product Usage: How often do they log in? Are they adopting high-value features?
- Customer Feedback: NPS, CSAT, onboarding surveys and qualitative feedback.
- Support Interactions: Ticket volume, resolution time, sentiment and escalation history.
- Billing Behavior: Timely payments, downgrades, contract pauses or renewals.
- Lifecycle Progress: Onboarding milestones, engagement in QBRs, completion of success plans.
A good health score helps CS teams:
- Spot leading indicators of risk early – So you can course-correct before churn becomes inevitable
- Automate proactive playbooks – Trigger timely interventions without adding manual work
- Guide success planning and QBR discussions – Back your customer conversations with data
- Prioritize where to spend time and effort – Focus on the accounts that need you most
- Track changes in customer behavior over time – See which accounts are improving, declining or stagnating
We use health scores as a starting point to guide internal conversations and customer strategy. For example, if a health score is trending down, we’ll dig into what’s behind it and make it a focal point in upcoming QBRs or success planning sessions.
But as helpful as health scores are, they don’t always tell the whole story.
Health scores can miss what the system doesn’t track. Like a champion resigning, a budget freeze or shifting internal politics. And when that happens, your customer might still show up as “green” right before they churn.
That’s where churn risk indicators come in.
Churn Risk Indicators: Your Early Warning System
Health scores give you scale. Churn risk indicators give you precision.
While health scores reflect general engagement and satisfaction, churn risk indicators zero in on the specific signals that suggest a customer might not renew. They’re not always visible in your standard dashboard but they’re critical for accurate forecasting and proactive save plans.
In our team, we track churn risk using two categories:
- At-Risk → There’s still a window to turn things around
- Confirmed Risk → Churn is highly likely or already communicated
This simple distinction helps us triage and prioritize save plans.
To build our Churn Risk Indicator system, we started by doing a historical churn analysis. We looked at every possible variable—from product usage to support tickets to engagement.
And for our business, one insight stood out: Declining product usage was the strongest leading indicator of churn.
So we made it central to our model. But we didn’t stop there.
We layered in qualitative insights from our CSMs—the real-world context that data alone often misses:
❌ Budget cuts or company-wide cost-saving initiatives
❌ Departure of a key champion or executive sponsor
❌ Shifts in product strategy or tech consolidation
❌ Unmet ROI expectations or misalignment with business goals
❌ Negative sentiment from decision-makers around renewal
These are the signals that drive churn but they don’t show up in a health score. They show up in conversations, strategy reviews and renewal negotiations. And when we track them intentionally, we forecast with far more confidence.
Better Churn Prediction Starts with Smarter Systems
You can’t prevent churn if you don’t see it coming.
Health scores help you scale. Churn risk indicators help you get specific. Used together, they give you a fuller, more accurate picture of where risk really lies.
And the good news? You don’t have to build this from scratch.
There are now plenty of AI-powered tools that help you build dynamic health scores and surface churn risks automatically. These tools analyze behavioral trends, usage patterns and engagement signals to give you a data-backed view of health and risk.
But even the best tools can’t replace human insight. You still need CSMs to catch context, nuance and sentiment that data doesn’t fully capture.
The magic happens when you combine both:
- Scalable, repeatable systems driven by AI
- And strategic judgment from your team on the ground
So if you’re relying on health scores alone—it’s time to level up.