A behaviour based safety observation is a few minutes spent watching one task being done, recording what you see, and then talking to the person about it. It is the only inspection in this library that looks at people rather than equipment, and that makes it the easiest one to get badly wrong — so this form is built around the conversation, records no names, and asks what made the unsafe way the easy way.
No names, and here is why that is not squeamishness
There is no field on this form for the name of the person observed. That is deliberate and it is load-bearing.
The moment a BBS record can be traced to an individual, two things happen. People work differently when they see the clipboard, so your observations stop describing normal work. And nobody tells you the real reason they took the shortcut, so you lose the only information the exercise was ever going to produce.
You lose nothing by anonymising. The purpose is to find out which unsafe practices are common and why — and both of those are answered by a pile of anonymous observations better than by a named one.
The conversation is the product
The scoring is scaffolding. The thing that changes behaviour is a short conversation at the end, and the form requires it before you can finish.
The shape that works: say what you saw without judgement, ask how they normally do it, ask what makes the safe way harder, agree one thing, and thank them. It takes two minutes. Opening with a correction ends the exercise before it starts, because the next person has already heard about it.
Line of fire, and why it gets its own section
Most serious injuries are not caused by people doing the wrong thing — they are caused by people standing in the wrong place while something else goes wrong. Under a load. In front of a panel. Behind a reversing vehicle. Beside a pressurised line.
It is the hardest thing to see, because nothing is visibly happening. The question to hold in your head is simple: if this failed right now, where would the energy go, and is anyone standing there?
A falling score is usually good news
Programmes that report 95% safe observations from the first week are measuring their own visibility, not their workplace. The number people quietly expect is uncomfortable: honest early observation rates often sit well below that, and they fall further before they rise, because people stop performing for the observer.
If your rate goes down after the first month, the programme is working. If it starts high and stays high, nobody is telling you anything.
Speak while you walk, and the document the law or the client wants writes itself — which matters more here than anywhere, because standing at a distance typing is exactly the posture that makes people change what they are doing.
Common questions
What is behaviour based safety?
An approach that focuses on what people actually do at the point of work, on the basis that most incidents involve a behaviour that had happened safely many times before. It works by observing, giving immediate feedback, and — the part most often skipped — removing whatever made the unsafe way the convenient way.
How long should an observation take?
Five to ten minutes of watching, then two or three minutes of conversation. Longer than that and you are auditing, not observing, and the person has started performing for you.
Should the person know they are being observed?
Yes. Covert observation destroys trust permanently and buys very little, because you have to talk to them afterwards anyway. Tell them what you are doing and why before you start.
Can behaviour based safety be used for discipline?
It cannot, and a programme that is used that way is finished. The moment observations feed into performance management, people work to the observer rather than to the task, and the data becomes worthless. Keep the two processes completely separate and be visibly seen to do so.
What is the most common mistake in a BBS programme?
Recording the behaviour and stopping there. If somebody is reaching over a guard, the finding is not that they reached over the guard — it is whatever made reaching over faster than the proper way. Fix that and the behaviour changes on its own; correct the person and it comes back next shift.