Skip to main content

Signal key

risk_score

Category

Composite

Source

Composite rollup of all negative signals, recomputed nightly.
A nightly composite score ranking a company’s decline and risk.

​
When it fires

Trigger condition

Mirror of momentum_score computed over decline signals and negative people signals (departures, churn).

​
Stored fields

This signal persists the following metadata you can read downstream:

​
Magnitude

Continuous score rather than a bucketed event.
For the full bucket definitions used across percentage-based signals, see Magnitude buckets.

​
Why it matters

Together, the two scores let you sort accounts by opportunity and by risk in a single view. risk_score is the inverse of momentum, it rolls decline and negative people signals into one number so you can spot at-risk accounts and protect existing customers before churn hits.

​
How to use Risk Score?

Where you sit. You lead CS operations at a subscription-software company. Your health-score model is honest about what it can see, product usage, support tickets, NPS, and blind to what it cannot: the customer’s own business condition. Which is why the postmortem on every shock churn says the same thing: usage looked fine, and then the company itself hit the wall. The mission. Add an outside-in dimension to customer health, so the model catches customers whose companies are declining even while their logins look healthy. The signal fires. As a nightly ranking rather than an event: risk_score aggregates each company’s decline-side signals, shrinking teams, collapsed hiring, audience decay, leadership exits, into a single comparable number. You pull it across your customer base and join it to health scores. The intersection tells the real story: fourteen accounts sit in the healthy-usage, high-risk quadrant, the exact quadrant your current model cannot see. Reading it. Usage measures whether users like the product; risk measures whether the company writing the check is okay. A power user at a sinking company still churns, through budget cuts, headcount loss, or acquisition, and no in-app metric warns you. The two-axis view separates four different renewals that your one-axis model treats as two. The play.
  1. Blend risk into the health score with real weight, and label the quadrants so CSMs act differently in each.
  2. For healthy-usage, high-risk accounts, shift the conversation from adoption to value defense: multi-year pricing, right-sizing before they ask, executive alignment on ROI.
  3. Stop over-investing rescue effort in low-usage accounts at healthy companies, those are adoption problems, not survival problems, and they respond to different plays.
  4. Feed both scores to finance for renewals forecasting; the blended number forecasts materially better than either alone.
Automate it. A nightly Company Signals API sync of risk_score across the book, joined to health data in your CS platform, runs the whole system unattended. Why it lands. Churn models built only on product data can watch the ship’s instruments while missing the iceberg. This score is the lookout, watching the water.

​
How to read it

Single risk view

One number summarizes all decline signals.

Churn early warning

Rising risk on a customer flags retention work.

Pair with momentum

Together they give a full opportunity-and-risk picture.

​
Outreach playbook

Use for retention. Rising risk on a customer account triggers a save play.

Momentum Score

A nightly composite score ranking a company’s upward trajectory.

Decline Signal

A company shows a clear distress pattern.

Restructuring Signal

A company is undergoing major restructuring.

Leadership Churn Spike

Multiple senior leaders departed in a short window.

Track this signal with CUFinder

Surface risk_score across millions of companies, enriched with verified emails, phones, and full firmographics. Start free, no credit card required.