Dashboard overview

The dashboard is where raw events become a churn signal you can act on. This page explains the moving parts; the default components page documents each built-in signal.

The pieces

• Event — a single thing a customer did (feature_used, login, …), tied to an externalAccountId.
• Event map — rules that translate your raw event names into the canonical metrics the engine understands. One raw type can fan out to several metrics via conditions.
• Component — a named signal with a metric spec (how to compute it) and score bands (what's good vs. bad). Components are the heart of the system.
• Widget — a visualization of a component or metric placed on a dashboard.
• Dashboard — a collection of widgets. New projects start from the Product Health template.

How the health score works

Every scoring component evaluates to 0–100 for an account by comparing its metric to two thresholds — good and warn:

• A normal signal scores higher as the value rises (more logins → higher score).
• An inverse signal flips that (more days of silence → lower score).

Each component carries a weight; the weighted average across all scoring components is the account's health score, banded as:

Components vs. widgets

A component decides what a signal means and how heavily it counts. A widget decides how that signal is drawn on a dashboard. The same component can power a number tile, a trend line, and the score — change the component and every widget and the score update together.

A component's kind controls where it's consumed:

Customizing

• Settings → Event map — map your raw event names to metrics.
• Settings → Components — adjust weights, thresholds, and metric specs, or build your own from the metric source picker.
• Settings → Widgets — arrange how signals are visualized.

For a step-by-step walkthrough, see Build a custom dashboard — how to create your own scoring components and visual widgets from your events.

Continue to default components.