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

# Signal

> Tracking user behaviour changes over time.

Signals are a method for tracking your [customers](/core-resources/customer) behaviour over time. These are helpful to identify the drop off in usage of features as potential retention issues or an increase in usage as a signal for expansion.

## The signal model

### Properties

`name` *string*

A unique name representing the signal.

`action_key`  *string*

The [action](/events/action) you want to track.

`metric` *string*

Use `metric` instead of `action_key` to specify a custom `name` and `period` for the two default signals: `active_unique_users` and `user_sessions`.

`outcome`  *string*

Whether the signal is `positive` or `negative` in nature. (Default: `positive`)

`aggregate`  *string*

How to aggregate when comparing the data between the two periods. (Default: `sum`)

`attribute`  *string*

The attribute in the event data to use for calculating the `aggregate=sum`.

`threshold`  *integer*

The percentage difference between the previous and current periods to act as the trigger for the signal. It can be either a positive or negative difference.

`period` *string*

The period lengths to aggregate data over.

`alert_if_no_activity` *boolean*

The signal will activate when there are zero events in the current period. (Default: `false`)