Heinrich Pyramid Explained: 3 Levels of Precursor Evidence

When a plant records many minor events and few serious injuries, leaders often reach for Heinrich's pyramid. The model becomes useful only when it turns those events into questions about exposure, controls, and decisions.

Heinrich's pyramid is not a countdown to a fatality. It is a way to organize precursor evidence, which means the evidence that appears before serious harm and shows where the system is losing control.

Heinrich's pyramid is a historical safety model that arranges outcomes from lower-severity events to serious injury. Its practical value is diagnostic rather than predictive. Leaders can use it to examine patterns, improve reporting, and test controls, but they should never assume that a fixed ratio guarantees when or whether a serious event will occur.

Definition

H. W. Heinrich introduced the triangle to show that visible injuries sit above a much larger population of less severe events. Herbert William Heinrich's original work is the named source for the classic model, while later safety practice often associates the lower layers with near misses, property damage, and unsafe conditions.

The key distinction is between an outcome and an exposure. A minor event tells you what happened on one occasion. A precursor signal can reveal what the operation repeatedly permits, normalizes, or fails to detect, which is why a small event can deserve executive attention even when nobody needed medical treatment.

Three levels of the pyramid

Level 1: Serious harm

The top level contains fatal and severe outcomes. These events are rare, but their consequence is high, which means a mature investigation must examine the decisions and barriers that existed before the event.

Level 2: Recordable or lower-severity injury

The middle level contains injuries that are more visible than near misses, yet still represent only a portion of the risk picture. A low injury count can reflect strong controls, limited exposure, or weak reporting, so the number requires context.

Level 3: Near misses, damage, and unsafe conditions

The broad base contains events and conditions that did not produce serious harm on that occasion. This is where supervisors can find early evidence, provided the organization makes reporting practical and responds without turning every report into a disciplinary exercise.

How to read the pyramid without treating it as a law

The familiar 300 to 29 to 1 relationship is a historical ratio attributed to Heinrich, not a universal law of nature. Different hazards, industries, tasks, exposure patterns, and reporting systems produce different distributions. A chemical release, a fall from height, and an arc-flash exposure do not share one risk profile simply because they can be placed in the same visual triangle.

Use the model to ask whether the base is visible and representative. If workers report only what causes damage, the pyramid is distorted before the analysis begins. If supervisors classify every event differently, the counts cannot support a meaningful comparison.

Andreza Araujo makes a related point in Much Beyond Zero, the English framing of Muito Além do Zero. The absence of an accident does not prove capability, because a quiet record may reflect luck or underreporting. The model therefore needs honest visibility before it can support prevention.

How Heinrich differs from James Reason's barrier model

Heinrich's pyramid organizes event severity and frequency. James Reason's Swiss cheese model examines how multiple barriers can align to allow harm. The first helps leaders see the breadth of precursor evidence, while the second helps them ask which defenses were weak, absent, or bypassed.

These models answer different questions, which makes them stronger when used together. The pyramid asks, “What signals are accumulating?” Reason's model asks, “Which layers failed or were never designed well?” On the Headline Podcast, Dr. Thomas Krause described how an event can look like an employee failure until earlier system decisions make the procedure difficult to follow. That perspective keeps the investigation focused on the work system rather than on a single person.

How to use the pyramid in prevention

Start with a stable classification method, then review reports by task, exposure, control, and decision owner. A near miss involving stored energy should not disappear into the same bucket as a housekeeping issue, because the potential severity and required controls differ.

Connect the review to the evidence tests in energy isolation verification when the pattern involves unexpected movement or release. Use near-miss evidence quality to test whether the report contains enough detail for action. When a signal points to a weak barrier, compare it with the practical distinction between lockout, guarding, and interlocked access.

The final step is ownership. Each recurring precursor needs a named person who can change the condition, approve resources, or escalate a decision. Otherwise the pyramid becomes a monthly chart that describes exposure without reducing it.

What the pyramid cannot tell leaders

The model cannot calculate the probability of a fatality from a count of minor events. It cannot prove that a team is safe because the base is large, and it cannot prove that a team is unsafe because the base is small. Reporting behavior, task mix, exposure time, control quality, and classification discipline all influence the visible shape.

It also cannot replace a serious-event investigation. When a high-consequence exposure occurs, leaders need a barrier review, evidence preservation, and a clear examination of decisions that shaped the work. The pyramid is a signal system, not an investigation method by itself.

Start with precursor evidence

Heinrich's pyramid earns its place when it helps leaders notice weak signals early and assign action before harm occurs. It fails when the organization treats the ratio as a prediction, rewards a clean record, or counts reports without examining the controls behind them.

Use the three levels as a disciplined vocabulary for visibility. Then test each meaningful signal against the work design, the barrier, and the decision that allowed exposure to remain. That is how a historical diagram becomes a practical incident-prevention tool.