> ## Documentation Index
> Fetch the complete documentation index at: https://docs.beaconrevenue.io/llms.txt
> Use this file to discover all available pages before exploring further.

# Churn risk

> A 0-100 measure of how likely an account is to churn, its five weighted inputs, and the thresholds that flag it.

## What it is

`churn_risk` is a 0-100 measure of how likely an account is to churn, where a higher number means greater risk. It is measured from signals already present in your connected systems. Alongside it Beacon produces a 12-month and a 24-month retention probability, a churn driver category, and, for accounts at 70 and above, an estimate of days to likely churn bounded between 14 and 180 days.

## How it is calculated

Measured. Five inputs are scored and combined at fixed weights.

| Input                           | Weight | What it reads                                                                                                              |
| ------------------------------- | ------ | -------------------------------------------------------------------------------------------------------------------------- |
| Engagement and adoption decline | 0.30   | Usage intensity, feature adoption, days since the last meaningful interaction, activation gap against the segment baseline |
| Health and support distress     | 0.25   | Ticket volume and severity, escalation rate, satisfaction trajectory, sponsor engagement                                   |
| Commercial signals              | 0.20   | Days to pay, missed payments, billing disputes, downward contract amendments, discount escalation                          |
| Lifecycle economics decline     | 0.15   | Direction and size of the change in the account's own economics                                                            |
| Fit risk                        | 0.10   | The risk dimension of customer fit, which holds the score steady against short-term signal noise                           |

One rule overrides the weighted total: where any single input scores 90 or above, the score is set to at least 70. The weighting varies with the account's lifecycle stage; the scoring logic does not.

| Tier     | Range  |
| -------- | ------ |
| Low      | 0-34   |
| Moderate | 35-54  |
| Elevated | 55-69  |
| High     | 70-84  |
| Critical | 85-100 |

70 is the threshold at which an account is flagged in renewal forecasting. The scale and these thresholds are fixed and do not vary by company or segment.

## Where it comes from

| Source                            | What it supplies                                               |
| --------------------------------- | -------------------------------------------------------------- |
| Billing                           | Required. Payment behaviour, contract amendments, disputes     |
| CRM and customer-success platform | Interaction recency, sponsor engagement, satisfaction          |
| Support or ticketing              | Ticket volume, severity, escalations                           |
| Product usage                     | Usage intensity and feature adoption, the largest single input |
| Finance or ERP                    | Days-to-pay behaviour                                          |

Where an input is unavailable, Beacon substitutes the average for the account's segment on that input and labels the score low-confidence. The score continues to be produced rather than withheld, and the label travels with it. Where a segment holds fewer than 50 accounts or fewer than 12 months of history, the retention probability is calculated from a cross-segment baseline and labelled as such.

## How fresh it is

Recalculated daily on triggering events, and in full on the first business day of each month. That monthly recalculation seals the month-end distribution. The input weights and the retention-probability calculation are re-fitted and sealed quarterly, 10 business days after quarter end. Triggering events are a usage drop of 25% or more week over week, a support escalation, a missed payment, a material satisfaction change, a contract amendment, and a 10-point move in the account's economics.

## Currency and rounding

The score has no unit. The revenue-at-risk figure calculated alongside it is stated in your single reporting currency, and every read of a financial value is recorded in the access log.

## What changes it

Any triggering event above, or any single input moving 10 points. A single input crossing 90 raises the score to at least 70 on its own. A change in the account's lifecycle stage changes which weighting applies. Between quarters the weights are fixed; the quarterly re-fit changes them for the periods that follow.

## What it is not

* Not a forecast produced by a learned model. The score is calculated from the inputs above by a fixed formula, and the same inputs return the same score.
* Not the retention probability. That is a separate value derived from the score, the segment, and the account's position in its lifecycle.
* Not the renewal decision. The score informs which action is offered; the decision follows from the account's economics.
* Not expansion readiness. A high churn risk removes an account from expansion consideration regardless of its expansion signal.
* Not non-renewal. A contract ending without renewal is recorded separately from churn.