CHURN IS DEAD

Your Health Score Won’t Save You

4 minutes · Trust

Archive note: This issue predates the evidence ledger introduced in August 2026. Treat uncited benchmarks and examples as editorial analysis, not independently verified findings.

I’ve sat in too many QBRs where the slide said Green and the room felt Red.

Logins looked fine.

Feature adoption was “on track”.

NPS was a cute little 8.

And yet the exec sponsor was cold.

Short answers.

Camera off.

“Let’s keep this one tight, I’ve got another call.”

Three months later:

“We’ve decided not to renew.”

The health score didn’t miss it.

It was never designed to see it.

The problem: your health score is scoring the wrong thing

Most health scores were built like this:

Logins? +10
Number of features used? +5
CSAT positive? +10
Last QBR in the quarter? +5

Congratulations.

Your customer is “healthy”.

But here’s what that score never sees:

The new CFO who hates your price.
The program sponsor who just left.
The quiet side-deal your competitor is running with procurement.
The fact that your product is now “just good enough” and no one is emotionally invested in it.

Usage is a signal.

It’s not the story.

And when the board asks,

“Which customers are truly safe and which ones are volatile?”

your health score can’t answer.

Because it was built to measure activity,

not predictability.

Predictability > Health

When I talk about Customer Predictability, I’m not asking:

“Is this customer happy right now?”

I’m asking:

“How likely is this customer to behave the way we expect in the next 3–6–12 months?”

Will they:

Renew at the current level?
Expand into the next product or use case?
Stall, regress, or quietly exit?

That’s a very different question.

To get there, you need to stop obsessing over whether the account is Green today

and start understanding its trajectory.

I look at three big lenses:

1. Direction – Is the account moving toward growth, stability, or risk?

2. Depth – How deep is our value + relationship beyond 1–2 champions?

3. Dynamics – What’s changing in their world that could flip the table on us?

If you only track logins and license consumption,

you miss all three.

The three lenses of a predictable customer

Let’s break those down into something you can actually use.

1. Direction: are we drifting or compounding?

Ask:

Are new use cases being added, or are we still talking about the same two dashboards from go-live?
Are we getting invited earlier into conversations, or only when things break?
Are execs more present in our cadence, or are we slowly sliding down to “just the admin”?

Direction shows up in momentum:

Expanding stakeholders.
Bigger problems.
Broader surface area of impact.

If your deck looks the same every quarter,

the direction is not growth.

It’s stagnation with good manners.

2. Depth: how replaceable are we?

Depth is about how hard it would be to rip you out.

It’s not:

“They like us.”
“We’ve been there for years.”

It’s:

How many teams would scream if we were switched off?
What critical decisions depend on our data, alerts, or workflows?
If our champion left tomorrow, would the account survive on autopilot?

Shallow depth is a single power user and a one-page renewal.

Real depth is when your product is wired into:

Business reviews
Risk conversations
Board decks
Incident process
Forecasting

Depth is where expansion starts feeling obvious, not forced.

3. Dynamics: what’s about to hit this account?

This is where almost every health score is blind.

Dynamics = change inside the customer:

New leadership (CFO/CIO/CTO/Head of Ops)
M&A events or divestments
Budget resets and “tool consolidation” programs
Strategic pivots (“we’re going upmarket”, “we’re cutting SMB”)
Vendor standardisation

These are the moments when a “healthy” customer suddenly becomes a churn story.

If you’re not tracking internal and external dynamics,

you’re not doing health scoring.

You’re doing weather reporting.

How to rebuild your health model in 30 days

You don’t need a 12-month project to fix this.

Here’s a simple path:

Step 1: Kill the vanity score

Don’t overcomplicate it.

For every account in your top segment, answer:

Direction: Growing / Flat / Shrinking
Depth: Shallow / Medium / Deep
Dynamics: Stable / Shifting / Volatile

You can literally do this in a spreadsheet or a simple table in your CRM.

The important part is the conversation, not the UI.

Step 2: Add real human signals

Update your account views to include:

Last meaningful exec interaction (not just “met at conference”)
Org changes you know about (promotions, new hires, exits)
Political and vendor landscape (who else is in the room and why)
Silence patterns: when was the last time they proactively reached out?

Silence is a signal.

Lack of complaints is not safety.

It’s often the pre-read to an RFP.

Step 3: Tie everything back to 3 business outcomes

Pick three outcomes that matter most for your product.

Example:

1. Reduced incident or outage impact

2. Faster project delivery / change cycles

3. Revenue or margin protection

Then ask, account by account:

Where have we proved one of these?
Where are we still promising it?
Where have we lost the narrative?

If you can’t map your work to any outcome,

that account is not predictable.

It’s fragile, no matter how many features they click.

Step 4: Build a single “Predictability View” per account

For each strategic customer, I’d rather see one page that answers:

1. What direction are they heading?

2. How deep is our value + relationship?

3. What dynamics could flip this deal in the next 6–12 months?

4. What are the next two moves we will make?

That’s it.

Not 17 tabs, 4 disconnected tools, and a rainbow of scores.

One view.

One narrative.

That’s what your ELT cares about.

That’s what your CSMs can actually use.

Step 5: Run a weekly Predictability Review (not just pipeline)

For your top 10–20 accounts:

Stop doing meetings where everyone just reads out “green / yellow / red”.
Start doing meetings where you ask:

“What changed this week that affects predictability?”

Talk about:

New risks, new leverage, new dynamics.
Moments of proof you can amplify.
Places where you’ve drifted into “vendor” instead of “partner”.

You’re training the team to think like portfolio managers,

not ticket responders.

5 questions to ask your team this week

Drop these into your next CS sync or leadership meeting:

1. Which “green” account makes you the most nervous, and why?

2. Which account looks messy in the tools but you feel strangely calm about? What does your gut know that the score doesn’t?

3. Where have we proven a real business outcome in the last 90 days, and did we document it anywhere an exec would care about?

4. Which customer has the most change happening internally right now, and how does that change our risk or upside?

5. If you could only keep 5 customers from your portfolio, which ones would you fight hardest for — and what do they have in common?

The answers will tell you more about your real risk posture

than any single health score ever will.

Bring this back to your world

If your dashboard is telling you everything is “healthy”

but your renewal anxiety doesn’t match,

trust your instincts.

Your instincts are picking up on:

Direction
Depth
Dynamics

Your tools just aren’t wired to see it yet.

Start small:

Pick 10 accounts.
Score them on those three lenses.
Build one Predictability View per account.
Run a weekly 30-minute review for a month.

You’ll be shocked how quickly patterns emerge.

If this hit a nerve, forward it to the person in your company who owns “health scores”.

Ask them a simple question:

“Are we measuring comfort, or predictability?”

Because churn isn’t where the story ends.

It’s just where the metrics finally caught up.

**Churn is dead.

Predictability is the new moat.**

By Kuber Sethi · All issues · Subscribe