Indicator Decay Explained: 4 Stages of Metric Staleness

Indicator decay is what happens when a metric still updates, but the work behind it has already changed. It matters because leaders can keep trusting the number after the field has moved on.

Indicator decay is the point at which a metric still looks current but no longer describes the risk it was meant to track. It usually starts when collection lag, proxy drift, or slow decisions break the link between the dashboard and the jobsite, which is why the chart can feel accurate and still mislead the review.

Definition

A live indicator is not just a number with a recent timestamp. It is a signal whose definition, collection rhythm, and decision rule still match the work. When one of those three pieces changes, the metric can remain clean while the meaning begins to slip, and the slip is what this article calls indicator decay.

The 4 stages of indicator decay

Stage 1: Collection delay

Collection delay appears when the source closes after the work has already moved on, which means the team is reading yesterday's job through today's review. Across 25+ years leading EHS at multinationals, Andreza Araujo has seen that delay create false calm, because the dashboard looks current while the field is already different.

Stage 2: Proxy drift

Proxy drift starts when the measure still tracks a valid proxy, although the proxy is no longer tight enough for the risk you want to manage. That is why leading indicator myths keep coming back, because teams defend the number while the work underneath it quietly changes.

Stage 3: Action lag

Action lag is the gap between seeing the number and changing the work, which is often where the real loss begins. In more than 250 cultural-transformation projects, Andreza Araujo has seen that a metric can be technically fine and still fail if the review rhythm waits longer than the field can tolerate, a pattern that also shows up in safety decision latency.

Stage 4: Comfort bias

Comfort bias is the final stage, where a stable line becomes proof in the mind of the reviewer even though the system has moved. James Reason's latent-failure lens is useful here, because the problem sits in the reporting chain, which is why risk register decay and barrier assurance belong in the same conversation.

How to differentiate in practice

The fastest test is simple. Ask what changed first: the work, the proxy, the collection rhythm, or the decision rule. If you can identify only the date of the chart but not the change in the field, the indicator is already starting to decay.

Indicator decay vs barrier assurance

Indicator decay asks whether the number still tells the truth. Barrier assurance asks whether the control still holds in the field. The first is a data question, the second is a control question, and they only overlap when the metric is built from direct field proof.

What to do this week

Start with the last three indicators that reached the leadership review, then ask whether each one still changes a decision or only decorates the dashboard. As Andreza Araujo argues in Safety Culture: From Theory to Practice, a useful number is the one that still changes behavior before the work hardens into routine.

Keep following Headline Podcast for more field tests that separate a live signal from a polite number.