Health Scoring

Health Scoring is a composite score that predicts the likelihood of renewal, churn, or expansion for each customer, built from behavioral, sentiment, and relational signals. A useful health score is not a vanity number. It is a decision aid that tells marketing, CS, and sales what to do next with a given account, and when to do it.

Why Health Scoring Has Become Table Stakes

Companies that run a serious health-scoring model see NRR lift of 6 to 12 points, especially in mid-market SaaS (Benchmarkit). In B2B SaaS, where top-quartile NRR sits above 120 percent and the median hovers near 106 percent (Wudpecker, 2025), a 6-point lift is material. It is the difference between a business that compounds and one that treads water.

The mechanism is straightforward. When teams can see risk and opportunity at the account level, they intervene earlier, with more context, and with less wasted motion. Without a health score, CS ends up firefighting the loudest accounts while quieter churn risks drift out. Advocacy and expansion opportunities get missed because no one sees them in time.

What Goes Into a Credible Health Score

A good health score blends four signal families, each weighted to the customer's stage and segment:

• Product engagement: login frequency, feature adoption, breadth of use across the account, time-to-value on new capabilities. The largest single input.
• Sentiment: NPS, CSAT, survey responses, review tone, sales call sentiment. Always paired with behavioral data, never used alone.
• Relational: QBR attendance, executive touchpoints, champion turnover, support ticket sentiment. Catches the signals that product data misses.
• Commercial: renewal date proximity, payment history, contract terms, expansion signals. The closing context that turns the score into a revenue prediction.

The weights change by segment. Enterprise health leans harder on relational and commercial inputs. SMB health leans harder on pure product engagement. One model does not fit both, and trying to force it produces scores that are technically consistent and practically useless.

Where Health Scores Go Wrong

• Too many inputs. A 30-component score is impossible to debug and impossible to trust. Fewer, higher-signal inputs produce more actionable scores.
• Scoring without routing. A health score that sits in a dashboard no one looks at changes nothing. The score has to trigger plays: CS intervention, marketing re-engagement, advocacy invitation, expansion outreach.
• Static weights. A health model that never gets recalibrated against actual churn and expansion outcomes drifts into irrelevance. Quarterly recalibration is a minimum.
• Sentiment over behavior. When sentiment dominates the score, the model misses the quietly disengaged customer who will churn politely. Behavior beats stated sentiment every time.

How Base Runs Health Scoring

Base builds health scores from a unified signal layer that spans product telemetry, community activity, support conversations, sentiment data, and advocacy participation. Scores are segment-aware, recalibrated against actual retention and expansion outcomes, and routed directly into the playbooks that CS, marketing, and sales already run. A score crossing a churn-risk threshold triggers a specific intervention. A score crossing an advocacy-ready threshold triggers an invitation. The score drives the motion, not a report.