Normalisation of deviance: why safety leaders get blindsided by drift

Table of contents

Click on a specific section below to navigate to that area:

• 1. What is normalisation of deviance?
• 2. How safety complacency lets deviance become “normal”
• 3. From small shortcuts to drift into failure
• 4. Normalisation of deviance is how safety drift begins
• 5. How to detect and reverse drift before it becomes an incident

What is normalisation of deviance?

Normalisation of deviance (NoD) happens when employees repeatedly break from standard procedures without any immediate negative consequences. Over time, they stop seeing the deviation as a risk. The shortcut becomes the new normal, and no one sees a need to question or challenge it anymore. At EcoOnline, we refer to this as Safety Drift.

What is normalisation of deviance?

In January 1986, Space Shuttle Challenger broke apart only 73 seconds after launch, killing all seven crew members. The Presidential Commission that was formed to investigate the disaster eventually traced the technical cause to failed O-ring seals that let hot gas leak from a booster rocket.

A more troubling discovery was found on an organisational level. NASA had known about the O-ring risk for years still kept making the decision to launch new space shuttles. This went on without incidents until they drifted into failure with the fateful Challenger launch. The pattern of accepting known risk simply because nothing had gone wrong yet is what, through Diane Vaughan’s work, gave rise to the term Normalisation of Deviance.

Normalisation of deviance vs. simple rule-breaking

Rule-breaking tends to be micro (focused on the actions of a specific individual or group) whereas normalisation of deviance is what happens when that rule-breaking becomes consistent and widespread.

In a sense, “rule-breaking” is an oversimplification: the conversation around it tends to devolve into finding and assigning blame, because rule-breaking requires a rule-breaker. But once a pattern has drifted into normalisation of deviance, the people involved usually don’t experience it as breaking a rule at all. It’s simply how the work gets done.

Rule-breaking is typically framed as a simple, individual act.

A person or group consciously deviates from a known standard, and blame is assigned accordingly. Normalisation of deviance describes something different: the gradual, often collective shift in what a group starts perceiving as the standard itself.

In the Challenger disaster, NASA had already flown 14 launches with the same flawed O-ring seals. As those O-ring failures were repeatedly accepted without consequence, they stopped being seen as anomalies at all. The deviance had become the norm.

In other words, as deviations repeat without consequence, they stop registering as deviations in the minds of employees altogether. This is what separates normalisation of deviance from simple rule-breaking. With rule-breaking, assigning individual blame is at least possible. With normalisation of deviance, it becomes difficult (and arguably misses the point entirely.)

See Rob Leech’s example from our Safety Drift webinar below:

Safety Drift: Why Your Safety Program Is Working – Until It Isn’t

Watch the full webinar

Discover how senior safety leaders can spot and correct weak signals, contractor risk, and supply chain blind spots before they turn into serious incidents.

Watch now

How safety complacency lets deviance become “normal”

As safety starts to drift, it becomes vital to understand the several root causes of what set everything in motion in the first place, preceding the incident. One of several root causes is safety complacency – the shift in mindset and risk perceptions in the workforce.

Safety complacency is more likely to set in after a period of no serious incidents or fatalities in an organisation. The stronger the safety record (on paper), the more confident (and complacent) the workforce will become. This is what Steve Arendt and Mark Manton refer to as the “Superman complex”.

The role of production pressure and “nothing bad happened… yet”

Production pressure is one of the most common causes of safety drift, especially in high-risk industries. In the case of the Challenger launch, the project had already weathered multiple scrubbed launches, and pressure was mounting to finally get the shuttle off the ground.

Unless carefully managed, production pressure and safety can easily pull against each other, where gains in one come at the expense of the other. A worksite that ignores safety in favour of pure productivity won’t last long. The direct opposite doesn’t work either. The safest possible site is one where no work happens at all. The goal is a middle ground, where risk is reduced to as low as reasonably practicable (ALARP).

A common explanation employees give for why the rules and procedures weren’t followed is a version of:

• “Nothing bad’s happened yet” or
• “We’ve always done it like this!”

Both explanations are clear signs that safety has drifted in the organisation. With both statements, it’s clear that work is being done differently to what was expected. Reasons why this is happening include:

• Risk normalisation – the employees have started taking shortcuts (often in the name of productivity) and the shortcut didn’t immediately result in an incident or other negative event.
• New protective measures or systems being seen as perfect solutions instead of fail-safes to known issues. This can lead to employees thinking that previous safety rules and SOPs don’t apply anymore.

It’s worth considering that risk normalisation and normalisation of deviance can compound. An employee can think of a workaround that saves them time (but endangers safety), which, if doesn’t lead to an incident, becomes normalised. As time goes on, the employee can invent another workaround on top of the previous workaround, which has become normalised (“we’ve always done it like this!”), which increases safety drift even more, and ultimately culminates in a major incident.

The example above was also seen in the Challenger case, where the faulty O-rings had been in use in several previous launches, during which nothing went wrong.  

6 warning signs of safety complacency on your site

• 1. Production pressure is climbing, with a push to “do more with less”
• 2. Safety dashboards look great, with few warning signs
• 3. It’s been a long time since the last (reported) incident
• 4. There are bonuses tied to safety metrics such as number of monthly safety observations
• 5. Departments operate in silos, with little communication between them
• 6. Near-miss and incident reporting is slow and sluggish

From small shortcuts to drift into failure

The core principle behind “drift into failure” is that accidents and disasters never come out of nowhere. They’re always preceded by a period of rising risk within the organisation (sometimes called the “incubation period”) during which that risk may or may not be recognised.

Over this period, several “latent errors” are allowed to accumulate, often taken for granted or simply going unnoticed in the organisation. It’s this slow accumulation, not any single event, that drifts an organisation into failure.

Things go wrong because things are wrong.

Dekker’s “drift into failure”: why good systems fail slowly

One key reason even good safety systems and organisations with a seemingly strong safety culture can drift into failure is the “Superman complex” mentioned earlier. Success in safety (or more precisely, the absence of failure) tends to increase confidence in the organisation, usually without considering whether the underlying risk has gone down or not. This is also why researchers have argued that, in safety specifically, the past doesn’t predict the future.

A long run without incidents tells you far less than it feels like it does.

James Reason’s Swiss Cheese model illustrates this well. All the components needed for an incident to occur can already exist within an organisation, but that doesn’t mean the incident will happen today or tomorrow. The risk grows as more “holes” line up over time, until eventually a trajectory opens right through the system and an incident becomes almost inevitable.

What the research says: normalisation of deviance in high-risk industries

Even high-risk industries aren’t immune to normalisation of deviance or safety drift. Every organisation is a living organism, subject to shifting productivity pressures and budgets, workforce turnover (with the resulting loss of knowledge and experience), and a changing mix of contractors.

Organisations in high-risk industries also tend to be complex by design. As Dekker & Pruchnicki note, this complexity makes it difficult to predict how a change in one system will ripple through the rest of the organisation.

And while these organisations typically have strict safety policies in place, a recognised driver of normalisation of deviance is that shortcuts or altered ways of working often aren’t born from production pressure at all. Instead, they emerge to fill gaps left by vague or unclear procedures.

Normalisation of deviance is how safety drift begins

Safety drift occurs for a number of reasons, with normalisation of deviance being one of the most significant. What’s important to note (as mentioned earlier) is that normalisation of deviance doesn’t need to happen knowingly.

Part of why this drift is so hard to detect is that it rarely happens all at once. It’s a gradual shift, occurring so slowly that even when employees do notice it, they often don’t consider it significant enough to report.

Another reason is that as the risk becomes normalised, employees stop perceiving it as a risk at all. They stop connecting it to the organisation’s safety policies. Over time, this way of working gets passed down from more seasoned employees to new recruits, who rarely have the experience, standing, or authority to question it.

The gap between work-as-written and work-as-done

A useful mindset to adopt is that the gap between work-as-written and work-as-done can be closed entirely. That’s not strictly true. A day without incidents doesn’t mean tomorrow will look the same (the Superman complex), but treating it as an achievable goal is still the most sensible approach to long-term success in safety.

Something that was implied earlier worth stating outright: work-as-written can often be the problem more so than work-as-done. As noted in a recent EcoOnline webinar on Safety Drift, it’s better to ask why the procedure stopped being possible to follow, rather than simply asking why it wasn’t followed.

It’s also common for the nature of the work to change while the standard procedure does not. When that happens, employees will easily start treating safety as a checkbox exercise, not out of carelessness, but because the safety procedure no longer reflects reality.

Why your dashboard won’t show you the drift

The core issue with trusting safety dashboards and taking everything at face value is that the dashboard can’t give insights about information it never received. A common saying in safety circles is that near-misses are gold. They work as great warning flags of incidents and fatalities that could have happened and might happen in the future but haven’t yet (remember Reason’s Swiss Cheese model).

In other words, your safety dashboard is probably not showing you all of the safety drift that is happening on all of your sites and departments, unless you’re getting in 100% of the incidents, near-misses, cases of non-conformance, and safety observations from your organisation.

How to detect and reverse drift before it becomes an incident

There are two parts to preventing safety drift. First, you need to get all the safety data in the system (because you don’t know what you don’t know). Second, you need to implement corrective actions to actually prevent safety from drifting in the future. While this is obviously easier said than done, this type of proactive safety is what most healthy organisations strive for.

Make near-misses easy (and safe) to report

As mentioned earlier, near-misses are gold. Every organisation should want to capture as much of this data as possible in their safety management systems. It’s the difference between preventing incidents before they happen and simply reacting to the fallout once they have.

To make near-misses as easy as possible to report, the reporting process needs to be frictionless, transparent, and risk-free.

In other words: the employee must find reporting easy and intuitive. They also need to see what happens as a result of their reporting. If they get no prompt feedback or sense of next steps from the organisation, they’ll have little reason to report next time.

Lastly, employees often perceive reporting as a negative act, one they fear being punished for, directly or indirectly. Management needs to address that fear often, and clearly.

Turn leading indicators into an early-warning system

As mentioned earlier in this article, things go wrong because things are wrong.

What follows from that is you likely already have plenty of leading indicators sitting inside your organisation, The work is in surfacing and acting on them.

Shifting your safety culture from reactive to proactive, where every department stays on top of potential risks before they materialise, requires a relentless focus on leading indicators that help flag blind spots and early signs of safety drift. While the work on safety is never finished, that’s what keeps organisations ahead of incidents rather than reacting to them.