What does good actually look like?

Most organisations can describe a disaster. Almost none can describe excellent routine performance. That gap is the problem

Ask a room full of safety professionals to describe a major incident from their industry and you will typically receive a detailed, confident account. The event sequence. The contributing factors. The failures of process, communication, or equipment. The investigation findings. Serious incidents generate serious documentation, and that documentation gets studied, referenced, and circulated for years afterwards.

Now ask the same room to describe what excellent routine performance looks like in their organisation — not in principle, not from the values statement, but in operational detail. What does a well-functioning team actually do? How do experienced workers manage uncertainty? What does good decision-making under pressure look like in your specific context? The answers are usually slower, vaguer, and less confident. Most organisations have a much richer vocabulary for failure than for success.

This asymmetry is not incidental. It reflects a foundational bias in how most organisations think about safety. The implicit model is that normal operations are fine unless an incident demonstrates otherwise. The energy goes into understanding what went wrong. Very little goes into understanding what goes right — and specifically, why it goes right when it does.

If you cannot describe what excellent looks like, you cannot build toward it, protect it, or recognise when it starts to erode.

Erik Hollnagel's Safety-II framework makes this point with considerable force. The overwhelming majority of operations in any complex system end without incident. This is not accidental. It is the result of people continuously adjusting, compensating, and absorbing variability in ways that procedures cannot fully specify and management systems rarely capture. These adaptations represent the real intelligence of the system — the accumulated knowledge of how to make things work under real conditions. If you do not understand them, you cannot protect them. You can, inadvertently, engineer them out of existence.

This happens more often than organisations realise. A new procedure, designed to standardise a process, removes the flexibility that experienced workers were using to manage variability. A headcount reduction eliminates the informal knowledge that existed in the people who left. A technology upgrade removes a manual step that was also, quietly, a check. The system appears to be running as designed. What has actually been lost is the adaptive capacity that was keeping it functioning.

Understanding what good looks like requires a different kind of observation than most organisations practise. It means going to where work is done and watching — not to audit compliance, but to understand the actual texture of skilled performance. It means asking experienced workers to talk through what they are doing and why. It means debriefing successful operations with the same rigour typically reserved for failures. It means creating documentation of effective practice that has the same status and accessibility as incident reports.

Some organisations have begun to develop formal processes for this. Learning from Excellence programmes, originally developed in healthcare, specifically elicit accounts of what went well and why — not to congratulate, but to understand and replicate the conditions that produce good outcomes. The evidence base for these programmes is building, and the consistent finding is that they surface operational knowledge that would otherwise remain invisible, tacit, and vulnerable to loss.

The practical implication for any organisation is straightforward, even if the execution requires commitment. Start building a vocabulary for success with the same seriousness you apply to failure. Document what your best teams do differently. Find out what experienced workers know that is not written down anywhere. Understand the conditions that allow good performance to occur and protect those conditions actively. Measure not just what goes wrong but what continues to go right and why.

You cannot build toward a standard you cannot describe. Good is not just the absence of bad. It is something specific, knowable, and worth understanding on its own terms.