People Don’t Die Because They Panic: The Truth About Crowd Disasters with Dr G. Keith Still.

When a crowd disaster makes the news, two words usually appear almost immediately: panic and stampede.

Dr G. Keith Still argues that both terms are often wrong. They place the blame on the crowd, when the real failures may have happened months earlier through poor planning, misunderstood capacity limits, weak crowd-flow analysis and safety documents that were never properly tested.

In this episode, David Franklin speaks with one of the world’s leading crowd science specialists about how crowds really move, what happens when density reaches dangerous levels, and why people do not simply behave irrationally in an emergency.

Keith explains the relationship between crowd psychology and physical crowd dynamics, including the point at which a crowd begins to behave more like a physical system than a collection of individuals.

They also discuss:

• Why people often leave through the entrance they already know
• The difference between crowd panic, crowd crush and progressive crowd collapse
• What changes between five and six people per square metre
• Why smooth computer simulations can create false confidence
• How simulations should test failure, not simply demonstrate success
• The DIM-ICE and RAMP crowd-safety models
• Why simple site plans, tracing paper and coloured pencils can outperform expensive software
• The difference between an experienced crowd manager and a competent one
• Common warning signs in generic and copied risk assessments
• How crowd risk should be addressed across ingress, circulation and egress
• Why site-specific diagrams and flowcharts strengthen event safety documentation
• The importance of live, current and venue-specific operational plans
• What Martyn’s Law could mean for event documentation, training and crowd communication

Keith also shares examples from his work on major international events and crowd-safety investigations, including the Hajj, the Olympic Games, the Love Parade disaster, Astroworld and Hillsborough.

This is an essential conversation for event organisers, venue operators, safety professionals, security managers, licensing officers and anyone responsible for planning or managing places where crowds gather.

Guest: Dr G. Keith Still BUY THE BOOK
Host: David Franklin
Podcast: SafetyDocs Podcast FREE TRIAL no card required

Transcript

SPEAKER_00 0:07

Every time you hear a crowd disaster make the news, you hear the same two words stampede and panic. Today's guests have spent thirty years proving that both words are wrong and that they blame the very people who died. The failures that kill people in crowds happen months earlier, in planning meetings, in capacity calculations, and in safety documents that nobody stress tested. Joining us today is Professor G. Keith Still, who has dedicated more than twenty-five years to understanding how crowds move, how they behave, and above all how to keep them safe. Keith comes from a mathematical modelling computer science and operations research. He holds a PhD in crowd dynamics, teaches crowd science at several UK universities, and since 1999 has been a key figure at the UK Cabinet Office Emergency Planning College, where he developed their crowd dynamics and crowd science materials. But Keith is far far more from a theoretician. He has worked as a safety advisor on some of the largest and most demanding events on the planet. The 2011 Royal Wedding, the Sydney and London Olympics, the London New Year's Eve fireworks, and the Jamarat Bridge in Saudi Arabia, a site where up to three million pilgrims perform rituals in a single day. And when things go tragically wrong, Keith is who they call. He has acted as an expert witness on high-profile crowd disasters, including the Love Parade tragedy in Germany, and as an advisor to the Hillsborough Inquiry. Here is what I find most interesting about Keith. He spent years building highly sophisticated computer simulations of then he concluded that expensive software is not what keeps everyday events safe and now teaches event managers to switch their computer off and use simple, low-cost tools instead of his demise and ramp models as a site plan and a pencil. He is a fellow of the Institute of Mathematics and its applications, a fellow of the Institute of Civil Protection and Emergency Management, and a specialist fellow in the International Institute of Risk and Safety Management. Please welcome the author of Introduction to Crowd Science, Dr. Keith Still. So welcome. Sorry for that long introduction, Keith. So just the first sort of question, before we get into the tools and the guidance documents, I want to start with the discipline itself because I think most of our listeners have never heard the phrase crowd science before this show. So the first question for someone who thinks um a crowd is just a lot of people in one place, how would you describe crowd science?

SPEAKER_02 3:03

Yeah, I mean it's it's the how, why, where, and when crowds form and move. So there's a combination of mathematics and psychology. Uh I was once uh uh calling onto a radio program and they wanted to put myself and a psychologist against each other uh to try and define, you know, is is a crowd psychological or is it is it physical? Uh and there's a threshold below one person per square meter, people are influenced more by their surroundings. Um, you know, if you're walking in a park, for instance, it's what you see in the distance that may influence your movement. But as the crowd density starts to increase as you get more and more people per square meter, it starts to take on a more physical form. So crowd science is this combination of maths and psychology, and the threshold there is around about one person per square meter. So it's a combination of understanding movement, density, flow, psychology in terms of the behavior, and unlike other psychologists that maybe study a crowd and try to figure out why they did something, we look at what we need to do to influence the crowd to do the right thing. So signage, information, design systems, uh how you manage a crowd, uh, how you present yourself to a crowd. Uh, for instance, a security guard with a military look uh may well elicit a type of response from a crowd where somebody that's drifted a more pastel colour with the word safety on them may influence them differently. So it's about psychology, it's about behavior, it's about managing the crowd, but understanding those thresholds. At what point does the crowd become a physics, more physics-oriented, where density and flow are related? At what point does it become more psychologically uh influenced? Uh at what point does it reach critical threshold where uh crush density might occur? So, really it's it's this combination of maths and psychology. Around about 70% influencing behaviour and 30% understanding those thresholds.

SPEAKER_00 5:10

Interesting, yeah. And and one thing is that when when something goes wrong in a crowd, um the word used is panic, isn't it? But you spent a career pushing back on that word. So, what actually happens inside a dense crowd when things go wrong?

SPEAKER_02 5:25

Well, first and foremost, um I think one one of the quotes it's attributed to me over the years is that people don't die because they panic, they panic because they're dying. So it was one of the last responses, and panic is generally uh a shutting down of the senses, not understanding what to do next, like a rabbit caught in a headlight. But running away from a threat is a perfectly natural response, a fight or flight syndrome. So as crowd density builds up, and as you start to find yourself caught in uh confined spaces or caught within a crowd, you're now desperately trying to get out. So crowd reaction, crowd behaviour there starts to take on a fight or flight syndrome, but that's after the risks are already manifest. So understanding that risk so um as the crowd reaches uh a threshold where people have a difficulty in breathing, then of course they're starting to push and shove to get out of the way. So it's a it's a natural reaction, a fight or flight syndrome. Uh, it's not panic. Um, panic would be the irrational response to a set of stimuli. And uh getting away from a threat is a perfectly rational response. The other phrase it's often used is stampede. Um, but if you look at things like Love Parade, Ito one disaster, uh Hillsborough, most of the other major disasters, when the press are using the word stampede, it implies the crowd have got space to run. But a progressive crowd collapse, a crowd crush, is they've not got the space to breathe. So majority of people that die in crowd crushes are dying through compressive or restrictive asphyxia, pressure on the ribcage, uh, preventing people from breathing. It's nothing to do with panic, it's nothing to do with stampede. It's a failure to understand that threshold where the volume of people exceeds the safe capacity of a site. So if I have a 100 square meters area, then there's a threshold at which I can put people in that. Once I exceed that threshold, they're at risk of a compressive or restrictive asphyxia. And generally we say once you understand your area, once you could do the calculation, once you understand what that threshold is, that at 80% of that uh maximum value, that's when you should start controlling crowd flow into those areas. It's um in an area or even a street, once the volume of people exceed the safe capacity of the flow down that street, then you're uh you're into a critical situation. So failing to understand the volume of people and the capacity of your site is is fundamental to these things. Once people are in that environment, they're at risk.

SPEAKER_00 8:04

So you talk about well, the industry talks about five people to a square meter. Then when that starts to increase, you've explained it a little bit, but what sort of happens when you move from like five people per square metre to then say six and a half, for instance?

SPEAKER_02 8:22

Well, we do the experiment in uh classrooms where we actually mark out a square meter on the ground, you put two, three, four, five. Um, once you get up to that five, there's still a bit of space between people, but you go up to six people per square metre, and they're now in physical contact, which means that any small movement on the outside can amplify uh and and start to cascade through the crowd. So you get this effect like dominoes, if people in close contact, first their feet get closer together, so the centre of gravity, the balance starts to it's like um uh you know, the kids' toy weebles wobble but they don't fall down, it was the old phrase. Well, you get that sort of uh experiment in uh classrooms where we actually mark out a square meter on the ground, you put two, three, four, five. Um, once you get up to that five, there's still a bit of space between people, but you go up to six people per square meter and they're now in physical contact, which means that any small movement on the outside can amplify uh and and start to cascade through the crowd. So you get this effect like dominoes, if people in close contact, first their feet get closer together, so the centre of gravity, the balance starts to it's like um uh you know, the kids' toy weebles wobble but they don't fall down with the old phrase. Well, you get that sort of uh inverted 10 pin, it's not got a stable configuration, and therefore any movement and people start to stagger and move against each other, pin, it's not got a stable configuration, and therefore any movement and people start to stagger and move against each other. Uh, the second element that comes into play is psychologically, once you're confined, you again you're trying to find a way out, you're starting to feel the claustrophobic effect of crowds around you. Uh seven, eight people per square meter, it just amplifies from that point onwards. So once you reach that five people per square metre, you still have enough space, uh, assuming that everybody's a sort of average build. Obviously, if it's um children, you might have a higher density or uh different countries. India, for instance, has a smaller body frame, so it has a higher threshold, but again, it's that physical contact. We call it the packing coefficient. Once people are packed so that they're touching, uh that's the point at which you then have those psychological and physical thresholds kicking in. Below that, it's reasonably safe because people feel they've got space to move. So, again, it's the psychological limit plus the physical limit when those two start to reach that critical threshold of five people per square metre, that's when the effects start to amplify and people feel constrained and therefore will start to react.

SPEAKER_00 11:07

Yeah, excellent, excellent, yeah. And and I don't know, um uh one thing that I've particularly observed is certainly since COVID, people seem to be struggling a lot more with anxiety, don't they? You know, even just going out shopping or things like that, there's seems to be an increased risk of people with anxiety. So that that's uh that's only going to affect the modelling, isn't it?

SPEAKER_02 11:29

Well, yeah, I mean we we factor that in. In fact, um one one thing that we've seen in uh certain environments is a sense of entitlement. People have been during COVID, you know, you switch on Netflix in, you've got instant gratification. You go to an event and there's queuing, uh there's merchandising, there's toilets, there's things to do, you have to physically move around, uh, and you have weights, so that level of um sense of entitlement of that, you know, I've paid my money, I should be able to have it instantly, right? Starts to kick into. And we saw a distinct behavioural change on that, that still seems to be um uh you know prevalent a couple of years after COVID shut down, and it seems to be a sort of psychological element within crowds, and again, you factor that into your thinking. Um, how might this crowd respond? And if you increase the information, telling people how long the wait is, what's happening, uh, it can reduce that level of anxiety. So, how we inform the crowd is a vital element to keep them uh safe, keep them informed, uh, and keep them behaved.

SPEAKER_00 12:33

Interesting, yeah, yeah, fascinating. So, I think we're kind of building a bit of a picture as to how these questions are going to go because you came to this field as a modeler, you built some of the most sophisticated crowd simulations in the world, which makes your current position all the more striking. So, to ask you about that, then you spent decades developing computer simulations for crowds, and yet in your book you tell people to turn the computer off, get out of sight plan, and use baking um well paper and coloured pencils. So, why did you walk away from this expensive software? And and what makes a simple line drawing a better tool for a crowd manager, for instance?

SPEAKER_02 13:11

Well, you you answered the question there uh when you when you posed it, expensive software. Uh some of these sophisticated tools require um a lot of knowledge of computer software and simulations, um, require a lot of modelling to put in the various parameters that are set there. And whereas somebody like myself who spent decades studying behavior and movement and crowds and different cultures, uh, I find it relatively easy to program those parameters and to turn a psychological model into a set of parameters to feed into a computer program. You know the old phrase garbage in, garbage out. Well, if you don't know how to do that, how can you trust the output? So what I realized is that in all essence there were things that I was seeing in the computer simulation that didn't need a computer after I'd found that relationship anymore. You didn't need it anymore. You know, you could understand if you put for instance, um, a lot of people program intelligence into agents. Well, crowds are not necessarily that smart. I mean, you walk through many environments where you think, why are people going that way? You know, what's influencing uh in those elements? And those are all the things you need to program in, that that level of sophistication. Whereas uh if you look at a map plan or a diagram, you can say I need to put signage here, here, and here in order to tell the crowd to move this way. And the signage needs to be high enough and visible enough. All of that is difficult to code into a computer simulation, but easy if you know the simple rules.

SPEAKER_01 14:48

Yeah.

SPEAKER_02 14:49

What are lines of sight? At what point does density start to affect crowd behavior? How do people know where to go? Uh and that's down to signage and information. So I realized that out of all of the years of building sophisticated modeling tools, there were a simple set of rules. And if I could teach those rules, then uh you know, in the front line, where uh maybe you don't have access to computers, but if you understand what the rules are, what the models are, the fundamentals of it, then you can keep crowds safe. So, for instance, we did a um years and years of research into the disasters, the fundamental causality from a legal perspective, uh, and distill that into a dim ice model. So you look at any section of a site and you say, What's the design, what's its limits, what information is available, and how do we manage the crowd in those spaces? What about ingress, circulation, and egress, dim ice? And then under both normal and emergency conditions, that's the three fundamental distal causes design failure, information failure, or management failure. And it happens during ingress circulation and egress. So if you understand that, then you can look at your site and say, How do we keep this area safe? So those rules can be distilled into a matrix in analyzing your environment, and then color-coding it, green, amber, and red. Green is stuff that we think we do well and therefore never need to scrutinize. Amber are things that we might need to improve for the future, and red are things that we've definitely got to uh you know fix now. So uh, you know, when you go to it's okay, we need monitoring or we need to fix it, and again, it's simple sets of rules that can be applied to any built-in complex environment in order to keep it safe. So, really, it was distillation or modeling simulation, years of research. That was my PhD basically, years of research into the underlying causality, and then how do we turn that into simple sets of rules and training and instructions so that your crowd manager on the front line who maybe doesn't have access to sophisticated computer programs, uh, or or even the the knowledge and experience to program or code that uh how do we turn it into tools that they can use that can keep crowds safe? So, really, that was the flip side of the work as I discovered how to do it and then started to develop teaching programs. So when I started teaching at the Cabinet Office back in 1999, the Emergency Planning College. I came in with a PhD up at this level, and I realized that everybody else was talking about at ground level what do we need to know to keep this crowd safe? So I had to come all the way down and think about how do I take what I know and turn it into teaching methodologies. So demise, crowd dynamics, density and flow, uh ramp analysis, roots, areas, movement profile, break it down into the risks, break it down into the dynamics models, break it down into analyze how crowds can develop into high density environments, look at risk and congestion mapping, which is color coding maps by time. So, what's the congestion during ingress, what's the congestion mid-event, what's the congestion during egress, and then a decision support tool. How do I put that together in order to determine my strategy and tactics? So that was so successful that many of the tools and models are now in UK, European, and Australian guidance and coupled with the work that we're doing out in the Middle East, and it's been so successful that it's mandatory training for all UK public event commanders. And I have for the last eight years been teaching a hundred police officers at a time, four or five times a year in India, just to understand how to apply science to an environment, to a culture. So when I did the workout in Saudi Arabia, I was an 11-year special advisor to the Saudi government for Hajj projects for the Jamara Bridge, uh, three million people per day. Same rules. Whether it's uh 500 people queuing uh at a nightclub or three million people per day uh in complex religious uh festivals and mass gatherings, it's the same rules. The risks are the same. Once density, exposure to high density reaches a certain level, then the risks start to rise exponentially. Now, the simple way to explain it is think of it like a tachometer on a car. You know, there's a red line. Well, what happens if you rev your engine in the red line beyond the red line? Something's gonna go. It's the same with a crowd. If you understand what that red line is, whether it's density or flow or volume and capacity, whether it's a queuing system, whether it's a behavioral change, if you understand where those red lines are, keeping it below the red line keeps it relatively safe. So it's understanding how to apply this red line model to built and complex environments so that during design, during planning phases, if you know how to apply the rules, you know how to keep that crowd safe.

SPEAKER_00 19:48

Interesting, yeah. I think one of the um most memorable things that I've I picked up from a training session that I did was we were we were at a venue simulating an evacuation and it was around a round building with doors all the way around the perimeter. So you could evacuate you know 2,000 people in in 10 seconds, really. But we were we were told that crowds will try and exit the way they come in because they're familiar with that. So there's all of us uh stewards being trained, and this and then they they triggered an evacuation, and I watched all of us go to the main entrance when all the exterior doors had been flung open, and I just thought that's it, we're we're we're kind of experts here and we're still doing it wrong. And it was only when the stewards started saying this way, this way, and directing the crowd through the nearest exit. So that really stuck with me. And I suppose that goes back to what you were saying about simple rules, you know, understanding the basic rules of human behaviour.

SPEAKER_02 20:52

Yeah. I mean that that can be explained um what's called the Ellsberg's paradox. And I can give you an illustration. Imagine a red ball, black ball game. In this hand, I have a bucket that does 500 red balls, five hundred black balls, so fifty-fifty. In this hand, I've got a bucket, same weight, same size, a thousand balls in there, but I don't know what that mix is. Could it be all red, could it be all black? Now, can you afford to lose five thousand pounds?

unknown 21:23

No.

SPEAKER_02 21:24

No? Okay, well let's play for ten thousand pounds. Let's make it interesting. Uh if you pick a red ball, I'll give you ten thousand pounds. If you pick a black ball, you have to give me ten thousand pounds. Now you've got a choice of these two buckets, fifty-fifty or the unknown percentage. So which bucket do you choose?

SPEAKER_00 21:43

Fifty fifty I'd go for.

SPEAKER_02 21:45

Okay, now mathematically they're identical, but that's not how the human mind works.

SPEAKER_01 21:49

Right.

SPEAKER_02 21:50

What we look at is what we feel is a known risk against the unknown risk. Right. And that's why people go out the way they came in. I know where that exit goes. I've got no idea where these other exits lead to. It could be into more danger. I don't know. So you go out the way you came in because it's that familiarity. And it can be negated if you've got the steward saying this way to safety. Now I've changed the information in the environment. I've told you, for instance, this unknown mix is mostly black. So I've changed the information and therefore I've changed your perception, your decision making on it. And that's how the psychology influences crowd behavior. So it's that information at the critical point, steward shouting this way to safety will change that behaviour of that crowd. Because otherwise, all they know is I'm going to take what I perceive to be the lesser risk. By the way, you lost, you owe me 10 grand.

SPEAKER_00 22:47

Excellent. Um, thinking a little bit about venues, you know, and and how they can implement some of these things, they'll often go to a consultant, and some of the people that listen to this podcast that are consultants, and they they will use simulations, won't they, to sort to sort of prove that the system works rather than test how it fails. So, how much false confidence would you say that's been put into that habit into the events industry, you know, relying on this software and I know of two major disasters where the simulation showed that it worked and it was safe.

SPEAKER_02 23:19

So um I that's why I bulk against simulations and models, because unless you know how to test a system and test the thresholds of failure, then all you're doing is producing an animation of this is what it'll look like if everything goes right. So I'd always programmed uh dumb people doing dumb things in order to see what was the worst case scenario, and then realizing once I've figured out what the worst case scenario is, what do I need to do to keep it safe? So when people see simulations and models, they're generally being presented on the basis I get paid a lot of money, and I don't want the client to see something that fails. So, you know, I want the client to be reassured that everything's going to be fine, and of course, that's where the sales pitch and the simulation and recouping your expenses on these uh expensive simulations, that's where it starts to fail. Is that you want to show the best result because you want the client to pay at the end of it for everything's working? Now, unfortunately, again, that's where you need to understand human behavior psychology, how to code the mathematics behind the models and simulations. And I often said that if um if you're presented with a crowd simulation, just ask the guy that's coded it, okay, what were your rules? What did you code into this? Uh, at what point will this system fail? And the faster their jaw hits the ground, the less confidence you should have in their input and analysis. Right. So take everything with a pinch of salt. Um, I've been called into uh several major events where the simulation showed everything was right, but the security staff are saying not comfortable with this because that's not how real crowds move in that environment. And what happens if you know something uh say it changes the weather conditions? How have you programmed that? What happens if uh you know I've got somebody behaving uh in a rather odd and curious way? How would the crowd respond to that? What happens if I've got a transport delay? What happens if I've got a stop and search procedure that changes? What happens if somebody comes into that environment and starts to look as if he's uh you know um uh behaving in an odd way or or with malicious intent? Uh, you know, what happens if I've got uh you know a bad smell? You know, how how might the crowd respond to that? Uh one classic example, I was I noticed somebody quoted me over the weekend, we had a uh a tunnel in Leeds, uh Leeds United, where all of a sudden there was this massive crowd congestion. Uh and the reason for that was that just before the end of the event, the police horses came through the tunnel and and left a rather large deposit, which of course meant there was a huge smell and stuff that was on the ground that you needed to avoid. I'll use the word stuff. Um, so obviously, you know, that changes behaviour. So, how does that factor in? Now, all of these are odd little anomalies you need to code in, in other words, dumb people doing dumb things uh rather than clever people always avoiding risks, and therefore your system's never going to show any of the dangers that might be involved in large, complex, crowded spaces. So, again, unless you're programming all that in, all you get out is a very positive result that all looks shiny and glowing. And that's an animation rather than a simulation of potential failure.

SPEAKER_00 26:35

Yeah. Yeah, it was uh it was us that posted that um your your quote on that. I think we even did a picture of them navigating around this um deposit that the horse had left as well.

SPEAKER_02 26:48

So apologies. I mean, I I got sent articles and details uh all the time, and it was it was fascinating to see that because of course that's what happened at Leeds. Uh large deposit, crowd behaviour. And there was quite a significant uh crowd crash um developing until we realized what what what had happened, and therefore we could then put some uh teams down there to just inform people to say if you just hold a couple of minutes, it's it's gonna be a bit slow ahead, you know. A horse has left a deposit. So again, it's easy to fix if you know what's gone wrong. But that's that would never show up in a simulation unless you coded it.

SPEAKER_01 27:25

No.

SPEAKER_00 27:26

So this is a little bit of a trickier one then. So if the licensing authority turns up to check on the events plans and they're shown this rather smooth simulation of an evacuation, what what should the governing authorities really be asking? What should they be really wanting to see demonstrated? Because the venue won't want to show them a bit of baking paper, will they, with some lines on it?

SPEAKER_02 27:51

You know, there's so where where do they kind of you'd be surprised because we we teach people how to incorporate that into your license submission.

SPEAKER_01 27:59

Right.

SPEAKER_02 28:00

So uh if you've got a uh and again we teach licensing officers to look out for this, uh, it's very difficult for AI to create a realistic site plan or site image. So it demonstrates that somebody has put some genuine thinking into it. I can produce a document and a risk assessment on AI, and the licensing officers would never know because they're now you know that good. However, the diagrams, things like flow diagrams, ingress and egress, a demise model, these things are very difficult for AI to produce accurately, and therefore you can see here and sense the human hand, the thinking that's gone behind it. And therefore, we've got a method where the Institute of Licensing can say, if I've got evidence that there are uh site diagrams within my license submission, I should feel that much more confidence that there's a competent person drawing it up. This is a huge difference between experience and competence. Uh, as you probably know, anybody out there that drives, there's a lot of uh experienced drivers out there, but not really that many competent drivers that fully understand the rules of the road and you know give way and thinking ahead. It's the same with crowd managers, there's a lot of experience out there, and if you take Astro World, uh where ten kids died, um in that I was I was an expert witness. If you want to see the report, Keith Still Astro World, Google that and you can get my expert report on it. It was published by the Houston Chronicle. And what you had there were a lot of people that had experience in events, but didn't know how to apply the fundamentals. So within their reports, they had the line diagrams, just didn't think it through, didn't apply it to crowd movement. Uh, you know, they they knew what the principles that set were, but because they had been formally trained in it, they didn't complete that uh that modelling scenario. So again, uh, you know, insufficient documents, uh training licensed officers what to look for, training police officers in order that when they're involved safety advisory groups in the UK, they can ask that questions. What if everybody goes out the way they came in? What if we've got malicious intent? How might the crowd react? What's your plan for it? And that really helps improve catching those um um plans that might not be competent but may well be experienced, and then just pushing them uh higher up that uh chain of safety understanding by applying crowd science.

SPEAKER_00 30:36

Yeah, yeah, yeah. Interesting. This isn't this in this we didn't discuss about this, but I've just had a thought really. So as you as you know, our safety docs platform, a lot of events use our app to store live plans and data. And just a question, because some people like flow charts, other people don't like flow charts, you know. But thinking about what you've described about what if, do you th what what is for the guy on the ground, you know, that's at uh at an event and they need they're asked that question, what if what's what's a good way to present that from uh you know, like an event manager's point of view?

SPEAKER_02 31:17

Do you agree with flow charts or I s certainly do, and and people uh I mean we teach uh three different methods for risk uh analysis. Uh do it numerically, uh do it by written description, do it by a diagram. Uh a flow chart, for instance, will help uh fundamental understanding, but not everybody sees things in the same way. So do it all three.

SPEAKER_01 31:40

Yeah, right.

SPEAKER_02 31:41

And therefore, whoever's reading it uh can see a numerical analysis, can see a written analysis, can see a diagram analysis, and therefore, you know, whichever one is is their mindset, we'll pick it up that way around. Now, I've I've seen and I've I've taught um we we did a calculation over 10,000 people. I think I've got 4,000 learners online at the moment, so over the the last 25-30 years or so, um, one of the comments that come back is, ah, this is so simple, I can not only apply it, but I can explain it to a client in a diagram. You know the phrase a picture's worth a thousand words, and what I like most about safety docs is that ability to take the best way of presenting information and keep that consistent.

SPEAKER_01 32:27

Yeah.

SPEAKER_02 32:28

So that when people come to a site, a different security company going to the same site, they've got a template that works for that site rather than oh, we did this at one football stadium, and therefore that should be how we apply it to all football stadium. As you know, different ingress, egress routes, different geometry, different nomenclature of north exit, Preston side, east side, you know, all of the phrases there that are unique to the site need to be maintained within the site. But you've got somebody that's got experience with that site, then leaves the job. You don't want that knowledge going with them. Yeah, you want to capture it, and and that's where I really like safety docs as a concept, is that it's site-specific, it's uh you know, it's venue specific. It means that people then have to learn how to communicate in that venues language, uh, and that knowledge is caught and captured. Now that's what I do with modeling, is that I take uh uh the event-specific details and turn it from what's known as tacit information, information that we know from experience, and turn it into codified information. So, how do I turn it into a forum that other people can pick up, understand, and utilize? And I think we've had some fascinating discussions, David, about how we can utilize that modeling technique and incorporate it into safety docs to give it um uh you know uh a more coherent alliance between the teaching and training principles and the documentation element of it. So it's it's a very exciting opportunity to be working with you on that.

SPEAKER_00 34:01

Yeah, yeah, absolutely. And um it's funny, only a couple of weeks ago we were actually um walking our evacuation route at at Leeds Arena that we did that we had a venue at. In our evacuation plan, one of the the so we have a an ERT, so emergency response team north and and one for the south. Emergency team split because you don't know quite where the evacuation will be. Well, the route described going through the building and out through a particular door, that was the easiest way to the muster point. As we're walking it, we're saying, hang on, this is during an evacuation, these corridors are going to be full of people, so we can't take this route, you know, to the muster point. So we changed the documentation that we would actually leave the building at the main door and walk around the outside. Well, in so we changed the documentation, it instantly synced to all the departments, you know, for that. So I think I think you've touched on that, like having live documentation, not a static PDF that is a nightmare to update and multiple versions.

SPEAKER_01 35:13

Absolutely, yeah.

SPEAKER_00 35:15

But just getting into the sort of the next the next sort of section, we um we kind of live it in safety docs, we kind of live at the moment because that's the options. Purple guide, green guide, the e-guide, and then of course the HSE documents, NATO, those sort of documents are all are all we've got to sort of check documentation and things like that. Now you've not always been um positive about these guides, and I and I get a lot of the we spot a lot of inconsistencies, believe it or not. Um but the UK guidance often quotes a capacity of two people per square metre, but but you argue that that number is dangerously oversimplified because it fails to distinguish between a static crowd and a moving one. So, what happens when an organizer applies that figure blindly to a crowd that's moving?

SPEAKER_02 36:06

You you put the crowd at risk. So basically, um you split the space into starting a dynamic. Now I'm currently working with uh various uh government agencies on updating uh guidance. Um but what you need to understand is that guidance and legislation, uh, so things like a statutory instrument like the building code uh will define density and flow, but that's not necessarily the same thing that's in guidance. So uh you can get prosecuted by failing to observe the building codes, but it could still be guidance compliant. And one example would be say the grain guide, guide to safety at sports ground. Excellent document, uh, and it does highlight at the start used by a competent person, you need other understanding, etc. But you just pick it up and you read through it, and it says four people per square meter for queuing areas uh in and around buildings. Now, the statutory instrument, the building codes, is 1.7 people per square metre, so immediately inconsistency. Now, we've advised on various buildings and sites in it and said if you know you're going to exceed the statutory limits for the building code, then you need to put additional safety measures into place to mitigate that risk. So you've recognized the risks, you're mitigating the risk, you realize that, for instance, um uh a stadium that might have uh hoarding around it for remedial uh work or building or construction work might then expose that crowd to a higher density, higher flow during egress. But if you've got monitoring and crowd management on standby for it, you're exposing that crowd to that higher risk for a short duration and it's monitored and you've got a safety or contingency plan in place, then that stands up in a court of law of taking all reasonable precautions. If you don't do that, you've exceeded the safety limits and therefore you can get prosecuted under the Health and Safety at Work Act, Occupiers Liability Act, and the building code would be the thing that people would use to say you've exceeded the safe limits. So understanding what the difference between legal requirements and what would be guidance requirements and it's it's great stuff, but you need to you know be a lot more competent than just to say, well, the green guide says four, therefore we can have four.

SPEAKER_01 38:25

Yeah.

SPEAKER_02 38:26

Under what circumstances is that permissible? At what point does that study break down? Uh, even things like um you read the green guide, it says eight minutes evacuation time. Uh that's what people think. I should be able to evacuate a stadium in eight minutes. No, what the green guide says is eight minutes for people to get to a free-flowing corridor, an area where they can continue to move without restrictions. So that could be getting from your seat to the concourse area. It doesn't necessarily mean to evacuate the building. So you've got to be uh precise about how you're reading and applying the guidance, and that's why in the guide to safety at sports grant in the event guide, it talks about competent uh use of the material, a competent person. And back to our driving analogy, there are a lot of experienced drivers out there, but not necessarily a lot of competent drivers. So competency here comes down to education, training, qualifications.

SPEAKER_01 39:23

Yeah.

SPEAKER_00 39:25

Interesting. So um talking about the sort of sort of crowd risk analysis at the minute, um, I mean it it's hard work for event organisers when you think when they're putting together their um operational plan, they're gonna have lots of documents in there. So safety plan, emergency plan, um well, the uh event-wide risk assessment, the list goes on and on, doesn't it? And then now with Martin's law, there's gonna be hostile vehicle mitigation, there's gonna be a whole raft of you know, what if in situations? The crowd risk analysis kind of stretches through all those documents, doesn't it? You know, there's gonna be elements of crowd risk in every one of those. What what when you've looked at documentation that's been submitted, certainly on these inquiries where things have gone wrong, what are kind of the red flags for you that show they've not really done a proper crowd risk analysis?

SPEAKER_02 40:20

Uh evidence of cut and paste, uh inconsistency in fonts, inconsistency in use in language, uh things that are generic that come straight out the health and safety executive's uh website. Yeah, yeah. Uh things such as uh generic trip slip and falls. A high proportion of accidents uh and incidents are a result of a trip slip and fall in a crowd. So if you just say trip slips and falls, high risk about that, but then don't specify at what location might the trip slips and falls be higher. An entry point, an exit point during entry and exit. But therefore that's a short duration high risk event that should be recognized and identified in a risk assessment, not just trip slips and falls across the site.

SPEAKER_01 41:09

Right.

SPEAKER_02 41:10

So specifying where a trip slip and fall might be at in trip slips and falls is one of the first things I look for is that uh you know, being uh you know one of the highest percentages, I forget what the actual statistics are, but if you Google it, the number of accidents and incidents that are a result of a trip slip and fall in a crowd that can then result in a progressive crowd collapse, but only at entry and exit points during entry and exit. So unless a risk assessment specifies that, it's just a generic cut and paste from the health and safety executive. Also the structure of a document. If it flits around the risk assessment from section to section, but if it's coherent, uh ingress, here are risks associated with ingress, circulation, here are risks associated with circulation, egress, then I see a different structure to what the uh you know the cut and paste from government guidance is, and I see somebody actually then thinking about it and applying it to that specific site. So the structure of your risk assessment is important, and again that's the one of the beauties about AI is that it it gives you a very generic format. If you start to look at okay, how structured is this risk assessment? How much information is put in there, what site specific details are in there, then uh you know when when I do expert witness, it's usually crowd management plan, risk assessment of the first two documents to assess. Did they fully understand the risks at that site? Have they structured the risk assessment to be specific to that site? And again, that's for safety docs are such a huge advantage. You can take site-specific information and say, here are risks known to this specific site. What are you doing to mitigate those risks? Yeah. So, you know, you you give yourself a template which is more site-specific and more structured, and now people have to think about filling the form in rather than just cutting and pasting from other elements.

SPEAKER_00 43:05

Yeah, that's good. So you'd like to sort of see there could be a risk assessment for slips and trips, but that could be by the the um concession booth, which if somebody's slipped there, probably not a big problem. But if somebody slipped by the exit, drawing egress, that's a big risk that needs to be identified in the in the document.

SPEAKER_02 43:27

And if it's not differentiated in your risk assessment, then somebody hasn't thought it through, hasn't uh, you know, a generic application. I'm dealing with a case just now, which is exactly that generic trip, slips and falls, which uses exactly the words for the EHSE uh guidance uh template. Um, you know, it talks about kitchens for an outdoor event. So it talks about slips, trips, and falls in kitchens. And I'm going there are no kitchens at this event. This is an outdoor event, this is uh catering tents, and there Therefore, the language is inconsistent with the actual site. And it's sure enough, it's it's a subcontracted uh standard template for all events rather than a specific template to that event site. And that that's where you know your uh safety docs is so uh unique in the feature. Is look, here's a template that is for our site. Yeah, you need to now think about how you're going to manage those risks for our site. So any uh subcontractor coming in uh now has a site-specific template that you've thought through, modified, and enhanced rather than coming in with their own risk assessment, which might be generic from somewhere else. So Safety Docs is a huge advantage of being able to capture and codify all of that information, that experience, and then turn it into a competency document rather than just uh a generic uh application of a risk assessment.

SPEAKER_00 44:51

Yeah. One major obstacle, you know, that that we're facing with that is um you'll you'll go for a a site tour of a venue and you'll say um what what can we see your documentation? You know, can we see um your evacuation rule or your muster points? Oh no, we can't give you that. That's uh you know, well, hang on, we're here to do a site risk evaluation, you know, to produce the documents that are site specific and venue specific, and yet you're not even giving us any information, you know. And that's surprisingly common.

SPEAKER_02 45:28

I mean, I I've done it for uh um more venues than I'd care to mention, is where we do um uh what would happen if you got litigation? You want us to do uh a competency check of your documentation, sign a various non-disclosure agreements. Obviously, I can't talk about who was involved in that, but uh then going through a review, uh doing a site survey, a site visit, and saying, Do you realize you are now exposed to the following risks because you haven't mentioned them in your documentation? So I approach it from a point of view of saying it's like an insurance policy. If you've got somebody like me checking it, then if those uh and you comply with what we've stated in our response, then in the event of litigation, I have to be defending that. So I often get called in because they don't want me on the prosecution side of things or the plaintiff side of it. They say, well, if we can get our documentation right and it's been checked by an expert, uh therefore we're that much safer if an accident occurs to demonstrate clearly, not only have we thought it through, we've had somebody review our documentation and improve the process procedures. Yeah. So uh, you know, uh safety is like insurance. Yeah, uh it you need to invest in it because if something does go wrong, it's your insurance policy that takes the brunt of that. So reasonable care, reasonable precaution, a higher level of documentation. The unfortunate thing is, as you know, it's it's um there's no profit in in the documentation, it doesn't make money. Exactly. So you need to think of it like an insurance policy. Yeah, yeah, you have to pay that out because you're paying it out for if something goes wrong. So reasonable care, reasonable precautions, adherence to uh uh you know standard operational practices, uh making sure that guidance and documentation, like including Martin's law, is all up to date and up to code, and then being specific about that site's nomenclature for trip, slips, and falls and location, duration, severity. So that's the huge difference between uh you know, again, uh an experienced individual drawing it up rather than a competent person, and because there's no profit in it, it generally tends to get um you know the the uh the the least uh path of resistance. If I can cut and paste from previous documents and I can do it quickly, and if I don't need to think about it, then it doesn't it hasn't lost me money. That that's the wrong attitude in in safety. Now you look at events and compare it to any other safety-related industry, like if the near-miss or close call, what happens? Team of experts go out, a new process procedure's drawn up, they've got legal experts, they'll have aviation experts, they've got pilots, all reviewing what that near-miss was in order to you know to put in a new process procedure to prevent it from happening again. Near miss and close call in the events industry, nobody talks about it. I'm not about to advertise I had six close calls last year. It's not good promotion, it's not good marketing. So it's the documentation that really should be embedding that information in and uh you know, learning from other people's mistakes. I think that the if the common phrase is it's a wise man that learns from his mistakes, it's a wiser man that learns from the mistakes of others. So, how do we capture that knowledge and information in order to keep our environment safe? And that's where training process procedures come into play.

SPEAKER_00 48:55

Yeah, yeah. No, it's uh it's very true, is that, and you know, you've got to think of the people turning up to these uh events as well, you know, exhibitors for instance, and they're thinking about setting the stand-up, are they gonna make any money and uh you know getting where they're gonna stay and all these things, these are all the main things and safety is like at the bottom of the list, really, and they'll they'll often start thinking about it a day or two before the event. So it's kind of just a re a bit of a re-education. Mart thinking about Martin's law now, because I think one there's a lot of talk about it, isn't there, when when a good venue has been doing uh compliant anyway um for a long time. Um but I think one of the big differences, and I don't know if you agree with this, is what will change is the fact that the SIA or other bodies can turn up at an event and either say show us your documentation, your processes, we don't think you're running a safe event, so we're gonna stop it. That's probably the biggest risk, the embarrassment, you know, of all that and the loss of credibility. Um but then there are the fines, and people are saying oh up to 18 million, but that's probably very unlikely that those sorts of fines will go out unless people are persistent. So it's gonna come the documentation's gonna be under more scrutiny scrutiny, ain't it? Is that is that how you see Martin's law?

SPEAKER_02 50:22

Well, two things. It's one that um uh in the event of an emergency, uh it's communication that's key. How do we communicate with a crowd? And unfortunately, things like a fireball just doesn't cut it. If I need to get people to move and to move that way, I need to communicate that uh to it. So again, uh, you know, it's that lack of uh thinking through an emergency scenario. Um documentation, I mean I've I've been involved in risk assessments from from small to very large events, and um it it's essential to consider documentation as part of the training process. So uh what we'll find is that uh if you've got to document Martin's law in a particular way, as as as we know, if that's then going to be checked by independent authorities, the SIA for instance, as we know, then that's just part of the process. I still need to turn that into an actionable plan. So having a document that passes scrutiny as opposed to having a process and procedure that is safe in the built-in complex environment are two very different things. So uh yeah, documentation is the starting point, but then teaching, training, education, communication with the crowd, and emergency, uh that's the things that then follow on from the documentation. So unless you've got a solid, stable base there, the rest of it is got to be built on that foundation. So your documents, your guidance, your details is the start of the process. Uh teaching, training, education, and forwarding. I mean, we we have a a choir. My wife and and uh good friend of ours are the musical directors. And we have a process of procedure, we've got elderly people in the choir, we've got people that are uh um not as mobile. Therefore, we get to make provision for it. We went and inspected the stage, it's the village stage. I I went along and I said look, we need uh footers on that, uh it we need to be visible because it's gonna be dark, uh, there's banisters need to be replaced uh to be stable because we've got people that have got uh lesser mobility impairment. And in the event of an emergency, we can't have those people on stage because they'll then not be able to get off. Uh you know, we start all of uh a hundred people in the audience, but have we got sufficient fire exits? Are they all clear? That's my job to go and check that, and that's just a small venue. Now we don't fall under Martin's law because we're less than 800 people, but we still take the process and procedures very seriously. Why? Because they're our friends and neighbours. You know, if anything were to go wrong uh in the unlikely event, I need to be able to stand up in a court of law and say we took all reasonable care, reasonable precautions. So that's the seriousness of your safety manager. And unfortunately, there's a lot of people in the safety industry that have the same attitude. I want to make sure that my venue, my event is as safe as possible. But we don't have the competency that can then think that through to all of its dimensions. And when you start adding Martin's Law, and well, there's a whole bunch of new stuff I now need to think about and learn and study and understand. And uh one of the biggest elements of that is communicating with a crowd. You know, uh I mean, even back to our simple example, if if we got loss of power during the summer, not so much a problem. Because it's light and people can see the exits during the winter. Yeah, that could be a little bit more serious. So, what's our emergency lighting? What's our backup plan? But more importantly, that power might come back on. We've a number of power outages here that are only for a couple of minutes. So, what's our contingency plan? Okay, the choir goes into a cappella mode. We keep singing, we keep performing. Sorry, the lights are out because we've lost our amplifiers, we've lost our uh keyboard, uh, we've lost all the power to everything that's uh you know musical instruments, we just have a choir on the stage, we've lost the lighting, we lost everything. What we don't want is people to suddenly react to that. So uh announcement from front in a very loud and clear voice. Right, if there's a power outage, if the lights go out, we will continue in a cappella until we can figure out what the situation is and come up with our uh you know contingency plan for it. But the most important thing is to keep that crowd calm until we've evaluated what the risks are. Yeah, you know, we we've uh had outages here that are just transient, but things need to be reset, yeah, things need to be switched back on again. So, what's our contingency plan to keep that crowd informed and calm while we're processing things? Because what we don't want is somebody to suddenly jump up, start running for an exit, start tripping over people, and then we've got uh a less of a manageable situation. I'm not gonna use the word stampede or panic. People are responding to that threat. So if they're warned about it in advance, if they know there's a contingency plan to play it and acquire a rehearsed on that plan, then we're not gonna have a problem. So again, it's understanding what does Martin's law mean for my crowd communication, my communication strategy, my documentation, my uh you know, sudden uh and unexpected check on that, my licensing application. That's where we see, you know, well we're organizations like yourself with safety docs, uh, really comes to the fore, you know, because you can think that through for venues and you can provide guidance for those subcontractors that are coming in to operate within that site.

SPEAKER_00 55:58

Yeah. You mentioned earlier about the cut and paste mentality and risk assessments and things. Do you there's going to be a lot more people having to do assessments from under Martin's law when you I when you think of big weddings, big funerals, you know, the more than 800 people, a lot more organizations will be coming into that. Do you think that Martin's Law could become like the biggest cut and paste exercise the industry's ever seen?

SPEAKER_02 56:25

Interesting. I I think there are uh certainly uh um well, like yourself, you're providing a template that's event specific, they then have to fill in that information to a compliant form. Uh that's not cut and paste. So, but if you are looking at something where you have to do it and you've got no template to work from, then there's a higher risk of thinking, okay, what's the quickest way I can get this done? Because I've got to get up to the license application. I've got to, by law, put in a risk assessment. So what's the quickest way I can do that? So without that compliant format, without uh the background in training, yes, I think there is a huge risk of cutting basis, and unfortunately. Uh well I think possibly, fortunately, um, we're still in a situation where that's a very low-level um element of risk to it. It needs to be done in order to protect people from harm. But uh, you know, the M type bombing, which led to Martin's law, it's not something that happens every venue every day. So uh, you know, again, be proportional in your assessment of risk. I know what I need to do for my local choir with a hundred villagers. It's not the same as that I would do for Hajj with three million people. It's not the same that I would do for a major venue where I've got maybe controversial performers or politicians appearing, where I've got a higher risk that's involved in that. So being you know, being proportional to your venue, your risk element to it. But I think that's where safety docs is a huge advantage because your location specific, whether it's an event or a venue or a building, then that template becomes uh a methodology where people need to follow it and then need to fill it in. But because you've already done the definition of entry exit points, you've walked it through, you've thought it through, and now you've got a template that's unique to that venue or building or event that then forces people into thinking through the risk assessment as opposed to just applying it.

SPEAKER_01 58:33

Yeah, yeah, brilliant, yeah.

SPEAKER_00 58:35

Well, thank well, thank you, Keith. We've we've not got through all the questions, but we'll probably do this again to go through those. Yeah, but a big thank you. It's been great to have you on, and and a lot of the listeners will really uh value this. What we like about the podcasts is you can listen to them as you're driving around, you know, so it's a great way of educating people without them actually being in you know in a more relaxed situation.

SPEAKER_02 58:57

Exactly.

SPEAKER_00 58:58

Yeah, the the book um is Introduction to Crowd Science. We'll put the link and Keith's details on the episode description. Um and if you work anywhere near a crowd, buy it before your next event or afterwards. Would you agree with that?

SPEAKER_02 59:13

Of course, though there was a sequel to it, Applied Crowd Science, which is published in uh 22 or so. Uh, and there's a free download of all of the case studies and examples that we've applied it to on the publisher's website. So uh I'll I'll send you the link to that. You can maybe put it in your podcast.

SPEAKER_00 59:29

Please do, yeah.

SPEAKER_02 59:29

Well, thanks very much for for letting me to uh to bleat on about my subject. You can understand I'm passionate about this. Uh, you know, spent more than half my life uh trying to um you know uh steer people in the right direction and developing tools and models, education and training programs to to inform that. And I and I teach around the world, so um still going strong.

SPEAKER_00 59:50

No, it's it's wonderful. Well, wonderful to have your experience. So thanks again, Keith. So if today's episode's changed how you think about the documents sitting on your shelf or the complicated, expensive software that you're using, we'll share this with the person who signs them off, and we'll see you next time. So thank you, Keith.

SPEAKER_02 1:00:09

Yeah, thank you, David.

SPEAKER_00 1:00:10

Cheers.

SPEAKER_02 1:00:18

If you would like a free trial of safety docs, please visit safetydocs.org.