Correlations

Insights → Correlations surfaces statistical relationships in your data — which metrics move together, and which might lead or lag others.

What it is for

• Leading indicators — find metrics that tend to move before an outcome you care about.
• Hypotheses — spot relationships worth investigating or testing.
• Diagnosis — understand what tends to accompany good and bad periods.

A caution

Correlation is not causation. A relationship in the data is a lead for investigation, not proof of a lever. Use correlations to generate hypotheses, then prove the promising ones with an experiment or the Causal Lab.

How it works

The correlation engine computes relationships across your KPI and performance series. It needs enough history to be meaningful, so correlations become richer as the platform accumulates measurements for your organisation.