Customer Health Score Benchmarks: How to Know If Your Score Is Actually Good

A health score only becomes useful once you know what "good" looks like. A score of 68 means nothing on its own — is that strong, mediocre, or a red flag? This is the piece most teams skip after building their formula (part 1), wiring up alerts (part 3), and launching playbooks (part 4). Benchmarking closes the loop.

Why Raw Scores Are Meaningless Without Context

Two companies can both report an "average health score of 70" and be in completely different situations. One might have a tight distribution where 70 is genuinely mid-pack. The other might have a bimodal book — a cluster of accounts near 95 and another near 40 — where the average of 70 hides serious risk concentration.

What actually matters is three things:

• Distribution, not just the average
• Correlation with real outcomes (churn, expansion, NPS)
• Trend over time, since a stable 68 is very different from a 68 that was 82 last quarter

A Practical Example: Benchmarking a 500-Account Book

Here's a simplified example based on a mid-market B2B portfolio of 500 accounts, scored on a 0-100 scale.

Benchmark snapshot

Health Score Distribution — Sample Book of 500 Accounts

Clarivoxx analysis

Average score68

Accounts in Red (<50)14%

Accounts in Green (>75)41%

Churn rate, Red band22%

Red 0-49
14%

Yellow 50-74
45%

Green 75-100
41%

Distribution across a sample book of 500 B2B accounts, illustrative

In this sample, the Red band (score under 50) represents only 14% of accounts by count — but because those accounts churn at roughly 22%, they account for a disproportionate share of revenue risk. Meanwhile the Green band (75+) churns below 3%, confirming the score is actually predictive rather than just decorative.

This is the real benchmark test: does your score separate accounts that churn from accounts that don't? If your Red and Green bands show similar churn rates, the formula needs work before the number can be trusted.

Setting Score Bands That Mean Something

Instead of picking arbitrary cutoffs (50, 75, 90), anchor your bands to actual churn and expansion data from the last 2-4 quarters.

Band	Score Range	Typical Churn Rate	Recommended Action
Red	0-49	15-25%	Executive escalation, save plan
Yellow	50-74	5-10%	CSM check-in, adoption push
Green	75-89	2-4%	Standard cadence, upsell watch
Champion	90-100	Under 2%	Advocacy, case study, referral ask

Notice the "Champion" tier above Green — this is often skipped, but it matters. Accounts scoring 90+ aren't just safe, they're your best source of references, case studies, and expansion revenue. Treating them the same as a 76 wastes the opportunity.

Re-Benchmarking as You Scale

A benchmark set when you had 50 customers won't hold at 500. As the book grows:

• Re-pull churn correlation every quarter, not once a year
• Watch for score inflation — if your average creeps up but churn doesn't improve, the formula is drifting, not the customers
• Segment benchmarks by customer size or industry once you have enough volume, since a 65 might be excellent for enterprise and mediocre for SMB

The goal isn't a nicer-looking dashboard. It's a score where the number on the screen and the renewal risk in reality actually agree with each other.

Benchmarking is what turns a health score from an internal vanity metric into something finance, leadership, and the CS team can all trust — and act on with confidence.