Topic 7 — The metrics of Customer Success

Goal: Know the numbers CSMs live by and how to read them.

Lesson 7.1 — Churn and retention: the leak in the bucket

Imani's first quarterly review at Cadence opens with a slide she doesn't fully understand. Lena, her manager, points at one line: "Logo churn: 5%." Everyone nods like it's obvious. Imani writes the number down and resolves to never feel that lost again.

Here's what the line meant. Cadence started the quarter with 100 active clinics. Five cancelled. That's churn rate — the percentage of customers (or revenue) you lose in a period. Five out of a hundred is 5% customer churn. Lower is better; churn is the headline measure of a leak.

Flip the same fact over and you get its twin. If 5 clinics left, 95 stayed, so the retention rate is 95%. They're two views of one number: 95% retention is 5% churn. You'll hear both, sometimes in the same meeting, and a CSM who can switch between them on the fly never looks caught out.

There's a sharper version of churn that talks about money instead of logos. Gross Revenue Retention (GRR) is the percentage of recurring revenue you keep from your existing customers, ignoring any expansion. Because it ignores upsells, GRR caps at 100% — you can't keep more revenue than you started with. It's the purest picture of how leaky the bucket really is. A clinic that downgrades from the $400 plan to the $200 plan dents GRR even if it never cancels.

Churn counts who left. GRR counts how much money walked out with them.

Imani's whole job, boiled to one sentence, is keeping these numbers healthy. Industry-wide, a median GRR sits around 90%, and the best B2B SaaS companies hold above 95%. When BrightSmile Dental renews without shrinking, Imani has personally defended a slice of that number.

Lesson 7.2 — Net Revenue Retention: the metric you name in the interview

Marcus, the account executive, closes new clinics all day. But there's a number that goes up without Marcus closing anything — and it's the one Cadence's executives stare at most.

Net Revenue Retention (NRR), also called Net Dollar Retention, measures how the recurring revenue from your existing customers changes over a year. It starts where GRR starts, then adds the good news GRR ignores: expansion. Upsells, cross-sells, clinics adding seats or upgrading tiers — all of it counts. Then you subtract churn and downgrades. What's left tells you whether your current customer base, untouched by any new sale, is growing or shrinking.

The reason executives love NRR is that it can climb above 100%. If expansion from existing customers outweighs everything lost to churn and downgrades, NRR breaks past 100% — which means Cadence would keep growing even if Marcus never signed another clinic. That's the holy grail of a subscription business: revenue that compounds on its own.

A quick feel for the benchmarks. The median NRR across SaaS lands around 106%. Roughly 100–120% is good, 120%+ is best-in-class, and big enterprise companies push past 130%. Anything under 100% means the base is quietly shrinking and new sales are just refilling a bucket that leaks faster than it should.

Here's why this lesson matters for your career. NRR is the single number that captures both halves of the CSM job at once: retention (don't lose them) and expansion (grow them). When an interviewer asks which metric a Customer Success Manager actually moves, you name NRR — and then explain that you influence it from both sides. Imani rehearsed exactly that answer before her Cadence interview, and Lena later told her it's the line that landed her the offer.

Lesson 7.3 — The health score: an early-warning radar

Three weeks before a clinic called Westfield Physio cancelled, the signs were all there. Logins had dropped from daily to twice a week. They'd stopped using the automated-reminders feature. A support ticket sat open for nine days because nobody on their side followed up. Nobody at Cadence connected the dots until the cancellation email arrived.

That's the problem a customer health score exists to solve. It rolls several signals — product usage, support tickets, engagement, survey results — into one indicator, usually shown as red / amber / green. Green means thriving, amber means watch closely, red means act now. A score sliding from green toward red is an early warning to intervene before the cancellation, not after.

Of all the signals feeding a health score, usage is the one to watch hardest. A customer who's quietly logging in less, touching fewer features, going quiet — that customer is often drifting toward the exit long before they say a word. Declining usage is the smoke that comes before the fire.

After Westfield, Lena turned on health scores for Imani's whole book of accounts. Now when BrightSmile slips from green to amber because Dr. Raman's front-desk team stopped using the scheduling sync, Imani sees it Monday morning and books a call by Tuesday. The score didn't fix anything by itself. It just told her where to point her attention while there was still time.

Lesson 7.4 — NPS, CSAT, and CES: three ways to ask "how are we doing?"

A health score tells Imani what customers do. The next three metrics tell her what customers feel — and each asks a different question.

NPS (Net Promoter Score) comes from one survey question: "How likely are you to recommend us?" answered on a 0–10 scale. The math sorts everyone into three buckets:

• 9–10 = promoters (they love you)
• 7–8 = passives (they're fine, not loud)
• 0–6 = detractors (they're unhappy)

You then take % promoters − % detractors. Passives are ignored entirely. The result runs from −100 (everyone's a detractor) to +100 (everyone's a promoter). In B2B SaaS, anything above 40 is strong. NPS measures overall loyalty across your whole customer base — the big-picture sentiment.

CSAT (Customer Satisfaction) zooms in. It rates satisfaction with a specific interaction or with the product overall, often on a 1–5 scale, reported as the percentage who gave a 4 or 5. You'd send a CSAT survey right after Theo's support team closes a ticket: did that go well?

CES (Customer Effort Score) asks the sneaky-important question: how easy was it to get this done? High effort is a red flag — a customer who had to fight the product to accomplish something simple is a customer drifting toward churn, even if they're polite about it. Low effort tends to predict loyalty better than delight does.

NPS: would you recommend us? CSAT: were you satisfied? CES: was it easy? Three questions, three different early warnings.

Lesson 7.5 — The real skill: reading the numbers and acting

None of this requires Imani to do math by hand. Cadence's tools compute every figure automatically — the dashboard shows NRR, the health scores glow red or green on their own, the survey scores roll in. Imani came from hotel guest-relations with no coding background, and she has never once calculated NRR with a spreadsheet formula. She doesn't need to.

The CSM skill is reading these numbers and knowing what each one prompts you to do.

A red health score is not a verdict. It is a prompt to reach out and find out what's wrong. A low CES on the onboarding flow is a prompt to check whether new clinics are getting stuck. A single detractor on the NPS survey is a prompt to pick up the phone, not to update a chart. The number tells Imani where to spend her human attention; the relationship is what actually saves the account.

This is the whole posture of a modern CSM. The data is an early-warning radar, and the warm, organized, people-first work Imani is great at is the response. A great CSM lets the numbers point, then leads with the relationship. Software flags the risk. A human prevents the churn.

Worked example — Imani reads BrightSmile's quarter

It's Monday, and Imani opens her dashboard to check on BrightSmile Dental, one of Cadence's flagship accounts and Dr. Raman's clinic.

The health score has slipped from green to amber. She digs in. Usage of the scheduling-sync feature has fallen by half over three weeks — the early-warning signal from Lesson 7.3. Nobody's cancelled, nobody's complained. But the radar is blinking.

The NPS survey from last month shows Dr. Raman scored Cadence a 7. A passive — not a detractor, but not a fan either. A few months ago she'd given a 9. Loyalty is cooling.

The latest CSAT, from a support ticket Theo's team handled, came back at 3 out of 5. Someone on Dr. Raman's staff found the resolution underwhelming.

Now Imani reads the story these numbers tell together: a once-enthusiastic champion whose team has stopped using a core feature, after a support experience that didn't land. That's not a renewal she can take for granted. None of the numbers told her why — they told her where to look and that the clock is ticking.

So she acts. She books a call with Dr. Raman, leads with the relationship — "I noticed the scheduling sync isn't getting much use lately, is something not working for the front desk?" — and discovers the team never learned a shortcut that would've saved them ten minutes a day. One training session later, usage climbs, the health score creeps back to green, and the renewal that was wobbling holds.

At the next quarterly review, BrightSmile didn't shrink. That steady line is Imani's contribution to Cadence's GRR. And because she also nudged them onto a higher tier during that call, she pushed the NRR needle too — both halves of the job, in one save.