CHURN IS DEAD
Health Scores Are Astrology for CS Teams
10 min read · Predictability & Risk
# CHURN IS DEAD
*Issue: Health Scores Are Astrology for CS Teams*
Green, Green, Green, Gone
Last year, a CS leader at an infrastructure monitoring company shared a story that stuck with me.
Their biggest customer — $2.4M ARR, four-year relationship, executive sponsor on speed dial — had a health score of 94. Ninety-four. Usage was strong. NPS was 9. Support tickets were low. QBR attendance was consistent. Every indicator in their platform said this account was thriving.
The customer churned.
Not downsized. Not consolidated. Churned entirely. Ripped out the platform. Gone.
When the post-mortem happened, the picture was devastating. The customer's new CTO — who'd started seven months earlier — had never been a user of the product at a previous company and didn't believe in the monitoring approach. He'd been quietly evaluating competitors since month two. By month five, the migration plan was approved. By month seven, the procurement team was negotiating exit terms.
The health score was green the entire time. Because health scores measure what the product sees, not what the customer's organization is actually thinking.
The CSM found out two weeks before the contract expired. By then, the decision wasn't just made — it was implemented.
Ninety-four. Green. Gone.
The Astrology Problem
Health scores in Customer Success work exactly like horoscopes.
They feel accurate because they're vague enough to match whatever narrative you want. A green score confirms that your customer is happy. A yellow score confirms you need to "keep an eye on it." A red score confirms what you already knew — the account is in trouble.
But here's what horoscopes and health scores have in common: they describe the present in terms generic enough to feel true while predicting absolutely nothing about the future.
Let me prove it.
Pull up your health score dashboard right now. Look at every account scored green — say, above 80. Now ask yourself: can you guarantee that every single one of those accounts will renew? Of course not. Some of them are actively evaluating competitors and you don't know it. Some of them have a new VP who's skeptical. Some of them are "happy" but their budget authority just shifted to someone you've never met.
Now look at your red accounts — below 40. Can you guarantee they'll all churn? Also no. Some of them are messy but deeply dependent on your product. Some of them complain loudly but would never actually migrate because the switching cost is astronomical. Some of them score red because a single power user left and usage dipped — but the account is structurally fine.
Your health score is measuring the symptoms. It has no idea about the disease. And worse: it gives you confidence about accounts that should terrify you, and anxiety about accounts that are actually solid.
Why Health Scores Fail: The Five Blind Spots
Blind Spot 1: They Measure Product Activity, Not Business Criticality
A customer logging in daily isn't necessarily healthy. They might be logging in because they have to — because the product is embedded in a workflow — but they hate the experience and are counting the days until their contract allows them to switch. High usage and high satisfaction are not the same thing. And health scores treat them as if they are.
Conversely, a customer with lower usage might have your product running in the background as critical infrastructure that they rarely touch because it just works. Low login frequency. Green health score says "at risk." Reality says "deeply embedded."
Blind Spot 2: They Can't See Political Shifts
The number one predictor of enterprise churn isn't product usage. It's organizational change. New executives. Restructured teams. Shifted budgets. Strategic pivots. And health scores are completely blind to all of it because none of this shows up in product telemetry.
A new CTO arrives with a mandate to consolidate vendors. A new CFO launches a cost reduction initiative. The champion who loved your product gets promoted to a role where they no longer influence the buying decision. These are the events that actually determine whether a customer stays or goes, and your health score dashboard knows nothing about any of them.
Blind Spot 3: They Reward Recency, Not Trajectory
Most health scores are point-in-time snapshots. Usage this month. NPS this quarter. Support tickets this period. They don't weight trajectory — the direction things are moving.
An account that was 95 six months ago and is now 82 is still "green" in most systems. But the trajectory is terrifying. Something changed. Something is eroding. A point-in-time score says everything's fine. A trajectory analysis says something is very wrong.
Blind Spot 4: They Treat All Green as Equal
A $3M account with a health score of 85 and a $50K account with a health score of 85 get the same color on your dashboard. But the risk profiles are wildly different. The consequences of getting it wrong are wildly different. The level of attention required is wildly different.
Health scores create a false equivalence that distorts resource allocation. Your team spends equal energy on accounts that are not equal in importance, because the dashboard told them the scores were the same.
Blind Spot 5: They Create Comfortable Ignorance
This is the most dangerous blind spot. When the dashboard says green, the CSM relaxes. The cadence slows. The depth of engagement decreases. "They're healthy — I'll focus on my red accounts."
Meanwhile, the green account is making decisions that will determine their renewal without any CS involvement. The health score didn't just fail to predict the churn — it actively caused the CSM to disengage at the exact moment they should have been digging deeper.
From Health Scores to Predictability Scores
Health scores answer the question: "How does this account look right now?"
The wrong question. The right question is: "What is this account likely to do next, and how confident am I in that prediction?"
I've been developing what I call the Customer Predictability Index (CPI) — a replacement framework that focuses on predicting behavior instead of measuring sentiment. The difference isn't semantic. It changes what you track, how you allocate resources, and when you intervene.
The 7 CPI Lenses
Instead of a single composite score, CPI evaluates each account through seven distinct lenses — each measuring a different dimension of predictability.
1. Stakeholder Stability — Have the key decision-makers changed in the last 6 months? Is the executive sponsor still in role? Have new influencers entered the picture? Organizational change is the #1 churn predictor, and this lens catches it.
2. Business Outcome Delivery — Not "are they using the product?" but "has the product delivered the specific business outcome that justified the purchase?" If they bought your security platform to reduce incident response time and it hasn't measurably done that, usage metrics are meaningless.
3. Competitive Exposure — Are they being actively sold to by competitors? Are they evaluating alternatives? Has their industry seen a wave of migration to a competing solution? This requires intelligence gathering, not product telemetry.
4. Budget Authority Alignment — Is the person who controls the budget still the person who values your product? Budget authority shifts are invisible to health scores but devastating to retention.
5. Expansion Momentum — Is the account growing its usage into new teams, new use cases, new departments? Stagnant accounts are vulnerable accounts, even if current usage is strong.
6. Dependency Depth — How painful would migration actually be? How deeply is your product embedded in their operational workflows? Products that are hard to rip out don't churn. Products that are easy to swap do.
7. Engagement Quality — Not frequency of meetings, but quality. Are the right stakeholders attending? Are discussions strategic or administrative? Is the customer bringing problems to you proactively, or are you chasing them for time?
Each lens gets scored 1–5. But unlike a health score, the lenses are weighted differently per account based on what matters most for that customer's specific risk profile. A government account where procurement cycles are rigid weights Budget Authority Alignment higher. A fast-moving startup account weights Competitive Exposure higher.
Your Customer Predictability Playbook
I've turned the CPI framework into a complete implementation guide that replaces your health score with something that actually predicts behavior.
The Customer Predictability Audit includes:
- The 7-Lens Scoring Rubric — Detailed criteria for scoring each lens with enterprise-specific calibration guidance
- The Health Score vs. CPI Comparison — A side-by-side diagnostic showing where your current health scores are lying to you
- The Stakeholder Stability Tracker — A template for monitoring organizational changes that health scores miss entirely
- The Predictability Dashboard Template — How to visualize CPI across your book of business
- The Intervention Playbook — Specific actions to take for each CPI risk pattern (low stakeholder stability + high dependency depth = different play than low business outcome + high competitive exposure)
*Your health score is a comfort blanket. Your CPI is a radar system. One helps you sleep. The other helps you survive.*
*— Kuber*
*P.S. — Here's a challenge. Pick your three "healthiest" accounts — the ones with the highest health scores — and score them across the 7 CPI lenses. If even one of them drops below a 3 on Stakeholder Stability or Competitive Exposure, you have a green account that's about to go red and your dashboard will never tell you. Hit reply with what you found. I'm tracking how often CPI and traditional health scores disagree — and the early data is alarming.*
Share this newsletter with a CS leader who trusts their health score dashboard. They need to understand what it can't see.
Subscribe to Churn Is Dead at churnisdead.com for weekly enterprise customer success frameworks that replace hope with strategy.
By Kuber Sethi · All issues · Subscribe