All glossary termsGlossary

What is Recurrence Risk?

Definition

A score (0–100) reflecting how often a specific table or schema change type has been associated with incidents, failed diffs, or failed expectations within a configurable lookback window (typically 30–90 days). High recurrence risk means 'this table has a pattern of causing problems' — it surfaces at release time so teams can make an informed decision rather than re-learning the same lesson.

Why it matters

Repeat incidents on the same table are preventable. If 'billing.invoice_items' has caused billing incidents three times in the last 90 days — every time someone touched its schema — then touching it again should require explicit sign-off, not just CI green. Recurrence risk makes that pattern visible at release time.

How Well Tested handles it

Well Tested's Memory Graph tracks schema changes, table observations, and incidents over time. When a new release touches a table, Well Tested queries the graph to compute a recurrence risk score: how many times has this exact change pattern (column type change, column removal, etc.) been associated with failures on this table? The result surfaces in the release risk score with a plain-English explanation.

See it in your release workflow
How Well Tested tracks Recurrence Risk
Get a 20-minute demo to see how recurrence risk fits into your release decision workflow.

No commitment. 20 minutes. We review your release workflow and show you what's possible.