Understanding signals and insights

Overview

Signals and Insights help you understand what is happening in your product and why it matters. Signals capture raw observations about user behavior and UI states. Insights group related signals into prioritized problems or opportunities you can act on.

Important: Because human–computer interaction is messy and varied, we cannot expect 100% accuracy on signal identification. The system is designed to tolerate noise at the signal level and surface higher‑quality Insights through aggregation.

Key concepts

• Signals are unopinionated evidence of what happened (for example, validation error shown, excessive cursor movement, or repeated back‑and‑forth between screens). Signals are tagged and include an impact level.

• Insights are the core unit of meaning in Adora. Multiple signals that describe the same underlying issue are clustered into a single Insight that persists and evolves over time.

• Impact levels (signals): Information, Minor issue, Issue, Major issue.

• Insight severity: Calculated using impact × frequency, so a few high‑impact signals or many low‑impact signals can both produce a high‑severity Insight.

• Traceability: Every Insight links back to the underlying signals for verification.

• Current availability: Today, Signals and Insights are available for sessions. Support for screens and journeys will be added later.

How signals and insights work together

1. The system detects Signals during user sessions.

2. Signals are tagged and scored with an impact level.

3. Related signals are clustered into an Insight that includes title, description, metadata, severity, and a suggested resolution.

4. As new signals arrive, they attach to existing Insights when relevant, keeping the story continuous.

5. You can push Insights to tools like Linear or Jira for delivery tracking (status remains lightweight in Adora).