The scene is some forsaken and forgotten small town in the 1800s, a dirty main street. Two gunslingers face each other, but the dust in the air is so thick we only see one. Suddenly a decision over life and death. The gunslinger draws and shoots from the hip. 

From the dust to the smoke. This is some ugly office-warehouse-mix-up building. I stand at the smoking corner with some logistics or manufacturing managers. If it were tech, I guess, it would be the coffee corner and there would be lots of steam. The people are engulfed in a nicely informal cloud of jokes and opinions. Tons of ideas are exchanged, half of them drowned by the atmosphere. I feel like I’m in one of those cozy business books: informal networking, real talk, authentic leadership. Gladwell and Sinek would probably milk it for a keynote. 

And then the top dog makes a decision, an important one. And I have to swallow hard: There goes 2500 years of methodology. 

The Board Room Version: The gunslinger manager is quite common. He takes decisions without a method other than his gut. But from Hippocrates to Galileo to Shewart we have 2500 years of scientific reasoning which should give us a blueprint for decision making. Because intuition although a real thing, has its limits as Klein and Kahneman show us. Same as our own rationality which is bounded according to Simon. What we need is Deming’s Profound Knowledge: Having an understanding of our company as a system, the data and variations it’s producing, how knowledge is built and the psychology of our people.

Maybe the problem is branding. “Scientific reasoning” sounds rather dry. “Evidence-based leadership journey” would probably sell better. Actually it is a journey: From Hippocrates and Aristotle to Bacon and Galileo we learned what proper scientific reasoning consists of: first you need to observe carefully, then throw in some systematic and logical thinking, testing your conclusions might also be a good idea and then learning from what you find along the way, finally adjust. The statistician Walter A. Shewhart first put this into a business perspective: The PDSA-cycle which we know as PDCA. Plan, Do, Check – or Study – and Act. 

  In the 1960s, Peter Drucker turned decision making into a cornerstone of how managers should work. In The Effective Executive the chapters on it are probably the most thoughtfully elaborated – besides maybe the advice to managers that they should have their secretaries keep a time log of all their wasteful activities. Never a bad idea.

With Drucker, decision making in business becomes a framework: you need to define the problem correctly before going for solutions and you need to identify the boundary condition. So first you need to know the type of problem before you can solve it because it tells you about how to solve it. Is it some unique event, a volcano eruption bringing down air traffic on a whole continent or the steady overcrowding of airports causing a constant stream of delays. That will tell a lot already. Individual emergency planning or structural change. Every quality manager knows that. Secondly you need to know the boundaries, what does your solution need to achieve, what should it not touch. But of course it doesn’t stop with the solution, you need something after the fact: the feedback loop. Otherwise how should you know if your solution or decision worked as intended? This is really subtle but also really important, so listen up: What Drucker added to the system is the nature and the context of issues. That is as important as the thing itself, otherwise you would probably get 42 or some other meaningless answer at the end.

This makes decision making in management not a talent, it’s a process. Something that can be learned and applied. I know, people, that takes the magic out of the whole dazzle of business. But magic isn’t real anyway – you do know that, sorry, Hogwarts not a real thing. That’s especially true when we do look at the moments when decision making seems to be at its most mysterious: The intuition. 

There is a whole school of thought rather fond of that: Naturalistic Decision Making. One of the most learned proponents is the psychologist Gary Klein. He studied it extensively, interviewing hundreds of firefighters, nurses, military personnel, trying to find out how these people make decisions. They cannot follow a long complicated process, right? They don’t have the time. The house might by already a pile of ashes once you got to the Do-Phase. Needs to be shorter, absolutely, but they don’t use magic. 

Klein found something he called Recognition-Primed Decision model in the cases that people think they act on intuition or sixth sense. The trick is that these people use a shortcut. Experienced people don’t compare different options. They recognize patterns out of their experience. The sick babies always show a change in skin color first. And then the expert finds the first working option. So the nurse thinks, I have to give a specific treatment. Third step is that they mentally simulate this option, checking if it would do the job. If given that treatment, the baby usually showed no worse symptoms. 

That is of course a very fast and efficient method to arrive at a solution. This resonated immediately. Because just as I finished reading Klein’s chapter on pattern recognition I went on a walk with my dog, Max. I realized for the first time that I was always fishing for the poop bag before Max actually started doing his business. That’s it, a recognition-primed decision. I was right, because I have seen Max do his little pre-pooping dance hundreds of times before, I recognized it out of my experience. The decision in that case was easy, but nevertheless a decision, scooping up the poop. What do you know. There is even a theory for dog walking.

So it must also be ok if an experienced manager does the same – I mean the intuitive decision making – not the poop thing. Well, not quite, because there are several limitations to that method.

Nobel laureate in economics Daniel Kahneman – a long-time intellectual sparring partner of Klein – describes this in his book Thinking, Fast and Slow. The two disagreed for years, but their exchange produced a useful conclusion: intuition can be real expertise — but only in the right environments. In order for the Recognition-Primed decision model to work you need to have learned from a high repetition of stable pattern that gave you a clear and rapid feedback. I must have observed a lot of pre-pooping dances – the pattern – and seen a lot of pooping – the feedback. If that’s not given, no learning, nothing to recognize, no simulation, no decision making that is any better than wild guessing. The smoking corner manager is not a nurse. They are not confronted with the exact same condition a hundred times. Their decisions do not produce immediate clear feedback. The outcome of their mostly somewhat strategic decisions produces results only months afterwards, most of the time filtered by politics and washed out by reporting structures and intervening events. (If any managers now goes, well I have a lot repetitive decisions with fast, reliable feedback, sorry, then you are micromanaging – taking care of stuff that is below your pay grade.) 

And actually Klein himself sees that too. He studied a worst case scenario when wrong pattern recognition led to the US Navy vessel Vincennes shooting down a passenger plane with 290 people. In the stress of the moment, with conflicting information available and a threat frame already in mind the crew built a selfconfirming story: Attacking fighter jet. An event turned into a terrible tragedy because of intuition. Kahneman says this happens because the associative machine – System 1 as he calls it – tries to form a coherent story even if only thin evidence is available. He calls that the WYSIATI effect – What You See Is All There Is. And I thought WIIFM was a stupid acronym! Joke aside, it is exactly what is happening. You do not take into account what you don’t see. How many times have we seen a team’s performance numbers drop and they immediately get a bad reputation. What we don’t see is that for half the month the IT system was down, or the supplier wasn’t paid on time and refused to deliver the parts needed. 

Even worse Kahneman found that the less information you have, the more coherent the whole story gets in your mind. Simpler pictures are easier to understand, easier to fool you. It’s like when I come into the kitchen. The cookie jar is empty and Max – the dog – is licking his lips. I think, the dog ate the cookies. I inhale to scream at the little devil when my son comes in and says: “Can you buy more cookies? I ate them all.” Subjective confidence cannot tell you if you are right. This is the problem with fast thinking or System 1.

If there is System 1 there of course there has to be a System 2. Well that would be the slow, analytical thinking. The one you should use, if System 1 does not produce a well-built option. Sigh, but that is so slow and strenuous! Isn’t there another way?

Herbert Simon – ha, another economics Nobel Prize winner – forms the bridge with his model of bounded rationality: we are capable of rational decision making, but we only can do so within our limits. These are mainly time constraints, the amount of information we have and our cognitive capacity. So we satisfice — meaning we find a solution that is just good enough within our limits — with what we have. That is of course dangerous. The less time or information I have the less rationally sound my solution will be. Now ask any manager out there what they have least of and it will be exactly that: enough time and sound information. Plus cognitive capacity is reduced by what? Right, stress, at least that managers have enough of!

The firefighter, the nurse, the poop dance watcher, they live in a world where the conditions remain the same. A lot of repetition, immediate feedback, stable patterns. Move up the hierarchy and things change. The smoking corner manager is on a whole other level. They’re making strategic decisions in complex environments, but they are still using the tool that belongs to the operational level, the level of frontliners.

So it has to be System 2. But what would that exactly look like in organizations? When at a loss for ideas, look to Deming, he will have it. Of course not always, but W. Edwards Deming is an exceptional figure in management theory. Hardcore statistician – intellectual son of the aforementioned Shewhart, Japan post WW2 economic guru and one of the first system thinkers. He has that whole thing about what he calls Profound Knowledge. In this context it means that your experience needs framing. So Max’ pre-pooping dance… no, let’s not go there again. Let’s say a customer service teams receives bad customer satisfaction ratings. Don’t just say agents need to do better. First you have to understand the processes, tools, suppliers and customers they work with. Always develop an appreciation for the system. 

Then you need to analyze the data. Data without a context is just blah, numbers. Of course Deming is a bit more concrete – and sometimes a bit narrow too. Variations are his main thing for working with information in systems. Common causes, special causes, how you have to react to them – spoiler: never change a whole system because of something special happening – investigate, have a plan, should it ever recur. Yet they are rare – most organizations have to put most effort into getting control of the big pile of the common stuff. These require system-level changes. Maybe in the customer service one agent was replaced with a two headed alien from the vicinity of Betelgeuse who has no clue how to operate a human telephone. Don’t buy a betelgeuzian telephone system with four-eared headsets and a depressed AI assistant built in. You still have 99 human agents. 

So it would be wrong to redesign the whole system around such special events. You have to always understand the whole system first, by learning, then the data, then the people. 

Which brings us back to the PDCA cycle. Deming was always associated with PDCA, though he later preferred PDSA — ‘study’ captured the point better than ‘check’, he said. And he was right. Because it is not a management tool, it is really a learning cycle. A structured method for organizations to learn continuously about themselves. Plan – Do – Study – Adjust. 

In my view the whole variation thing is not the only aspect. You’ll need to understand all the different components of your system and how they interact. You need to understand the elements, the interconnections, the purpose, the stocks, the flows, the feedback loops and all that jazz. Like the modern system thinkers do – shout out to Donella Meadows. A real marksman doesn’t shoot from the hip. They account for their equipment, distance, wind, elevation, their own breathing. Engineering, Physics, Metrology, Biology (and then you realize that shooting is something stupid anyway). Every variable in the system matters. Same idea. Sound like a lot, yes, it is, maybe another tool might help.

Because PDCA is not the only help managers have. In my trainings I always say that you need to use at least two of these funny acronym tools. PDCA and RCA – Root Cause Analysis. The first is your overall working mode, the other is the reinforcement – don’t just scratch the surface, always dig deeper.

Yet after the gazillionth PDCA or RCA training I did in my career, I see that it doesn’t change a thing. I ask myself, am I a bad teacher. But I tested it. I asked people months later if they can remember them. They usually do. They just don’t use the stuff. They keep going back to that smoking corner and shoot from the hip.

I see a cultural problem here. We know the tools but we operate in a culture which keeps rewarding the wrong behavior. We have 2500 years of experience in sound decision making, at least 60 years in management thinking. The gunslinger has had so much time to be wrong, but eventually they still get promoted. I guess there is more to explore.

And yes I know that I conflated decision making and problem solving in this piece, but while there is a fine intellectual difference, for the sake of simplicity, just shut up.

If you were wondering: The gunslinger misses in the end and dies, because that’s the way management of that sort will eventually turn out. 

Recognize any gunslingers? Or are you one yourself? Share and discuss — LinkedIn-Post

Don’t just take it from me, here is some good stuff to read:

Peter F. Drucker, The Effective Executive, 1967

Gary Klein, Sources of Power: How People Make Decisions, 1998

Daniel Kahneman, Thinking, Fast and Slow, 2011

Herbert A. Simon, Administrative Behavior, 1947

W. Edwards Deming, The Essential Deming, edited by Joyce Orsini, 2012

Donella H. Meadows, Thinking in Systems: A Primer, 2008

Douglas Adams, The Hitchhiker’s Guide to the Galaxy, 1979