Threshold
A Threshold is a defined boundary in the data that separates “watch” from “act differently”. It can be a level, a duration, or a pattern that signals a meaningful change in risk or opportunity.
What we mean by “Threshold”
A Threshold is a decision boundary in a metric: a point or zone beyond which the current playbook no longer applies. It is not just a red or green colour; it is an agreement that, once crossed, the team will treat the situation differently, even if the full root cause is not known yet.
Why this changes how people read a dashboard
Without clear thresholds, dashboards overload people with choice. Every small shift can be argued as “no big deal” or “urgent” depending on who talks.
- Reviews get stuck in long arguments about whether a change is serious instead of what to do about it.
- Teams either under-react, letting problems compound, or overreact, constantly chasing noise.
When you will feel this term in real life
Thresholds are behind one of the most common dashboard frustrations:
- Endless debate “Is this serious?” — the team has no shared view on what counts as a “real” deviation.
- Decision fatigue — people feel they have to judge every small movement from scratch.
Agreeing thresholds is not about predicting the future perfectly. It is about reducing repeated judgment on predictable situations.
See this term in context
Thresholds for Action is dedicated to this topic:
Related terms in this glossary
Thresholds are defined together with:
When teams are clear on thresholds, dashboard reviews move faster: less time on “Is this bad?” and more time on “Given that it is, what now?”.