January 17, 2020

The Story of Allthings

My friend Stefan Zanetti has always been an entrepreneur, if not by title, then at least in spirit. I first got to know him as my boss, running a small team that was dedicated to projects within the HR department of an insurance company. He used to take up topics that he found important, and drive them until they became reality – often with the support of the organization, sometimes with their tolerance, sometimes against their resistance. So it came as no surprise to me when in 2005 he decided to leave the world of big corporate in order to found his own startup. At the beginning there was an idea, or rather, a strong belief. It can best be summarised by a statement published by the industrial insurance company FM Global in an outlook. It says that technically speaking, 50% of damage volume to property could be predicted and therefore prevented or reduced. For example, much of the damage created by factory fires is actually due to a lack of maintenance of the fire alert and sprinkler systems. The missing link to making these potential preventions and damage reductions true is access to sensors and data, in order to read weak signals, monitor maintenance and the occurrence of damage, and the size of impact. The technology that would change all this: the internet of things (IoT). 

For the next seven years, Stefan built up and ran two companies which concentrated on building that link between selling insurance and making people’s life at home safer. After some initial ups and downs, both companies were doing o.k. But Stefan realised that over time they had ended up providing webshops and digital services in the insurance and later on also in the utility sector, but they were not engaged on the initial idea of using the huge potential of IoT to make lives safer. The main reason was, of course, that the IoT underwent the same cycle as many revolutionary technologies. As Roy Amara of the Institute of the Future put it, people have the tendency to overestimate the short term impact of a new technology, and underestimate its long term effect. The IoT revolution would take much longer in the making than the hype talk suggested.

The result of this refocus on the original idea of their startup was the realisation that nonetheless, something could be done. There are problems of a simpler nature which can be addressed by connecting objects to the digital world. Not necessarily by connecting all of them in a big network, but simply by connecting them 1:1 to a digital identity, which allowed for additional things to be done. Your e-bike comes with its digital user manual that is always accessible, your spectacles remind you when it is time to have your eyesight tested, your camera informs you of a new set of lenses that has come out. A platform that does this could be open to third-party services such as insurance. It was the beginning of Qipp, later re-christened Allthings. 

Qipp-Allthings was to be the platform that provided objects with a digital identity – a kind of Facebook for things. At that time, the business model was Business to Business to Consumer (B2B2C): If you get your new e-bike, and are invited to register on the platform, it is not you who are going to pay for it, but the producer of the bikes. What they learned very quickly at Allthings was that many producers of consumer goods were indeed ready to pay for this service. What they paid for was access to the end consumer. Think of a company that produces DIY tools, or SLR cameras, or small household electronic devices. It sells hundreds of thousands of products each year, and at the same time has no direct channel to any single end consumer. It was the perspective of this channel that made businesses want to work with Allthings, the prospect of learning about customer behaviour, and of binding customers to the brand. But which industries, which companies would be most interested, and which features did they and their end consumers want to see on the platform? There were many assumptions, and zero evidence. 

Therefore it was clear that the initial phase of the company would be one of experimentation. They needed a platform that was capable of running many different features, they needed to try out many different packages of features with various partners, and they needed to modularise the offer as much as possible, in order to be able to re-use and scale. From the very beginning, they did plan for a large variety of such features to be accommodated for the various customers. But the variety they encountered was even larger. 

I remember one of the very first features they tried out with friends and family was a lost-and-found service. You get stickers with QR codes to put on computers, keysets and the like, and connect the QR code to your online profile. Whenever a third party reads the QR code, they can get in contact with you without knowing who you are, or where you live. 

The commercial applications were richer: 

  • A large wine merchant had QR codes put on their bottles, especially to invite guests in restaurants to their webshop. 
  • A bike manufacturer delivered the whole documentation and repair history with their bikes, and upselling opportunities for additional features. 
  • An insurance company delivered baggage tags with included travel insurance. 

This phase of exploration came with its own array of leadership requirements. In retrospect, one of the key tasks Stefan sees for himself as a leader is the one to provide endurance, or momentum, in the sight of uncertainty and failure in detail. At the moment of founding Allthings, Stefan did not know which industry, which product, which technical solution or business model would bring success. But he believed in the potential – in the validity of the vector – and that was how he drove the business. He believed so strongly that he invested almost all his own money in the venture, at a moment when he did not know which product or service they were to deliver to which client. The strength of this bet that the overall direction made sense, enabled him to face the many little issues that needed to be solved, but more importantly, to tolerate some ongoing problems for the sake of focus, and to endure regular failures in small areas. In hindsight Stefan realises that many people, both within and without the company, cannot really deal with that mindset. He would cause irritation by adopting an Oscar Wilde quote I once gave him: “How can I know what I mean before I see it?” While other people may try to create areas of relative calm and stability, he was constantly on the search for the new, the different, of areas where he was not happy with the status quo. In some way this is another version of the explore-exploit tension: the friction between the forces pushing for stability, and for change. The key is to make this friction productive, and not destructive. Stefan’s job was to constantly go around and pitch his ideas to anyone, both within the company and with potential partners or clients, see what reactions he provoked, and try to connect it to intentions of the person in front of him. It was less a question of methodological planning, and more of the right psychological attitude. 

Of course this emphasis on belief and attitude over planning doesn’t mean that Stefan was frivolous in how he spent the company’s assets. He knew that their own capital had to last for a relatively long time. But he did feel a bit like Christopher Columbus: how do you plan your supplies when you are going to India around the other side of the planet? The answer they came up with took the unknown issue of each customer project very seriously. Every engagement with a business client was an experiment they believed was worth making, but they wanted to make sure that the other party, who knew their own part of the business better than they did, shared this belief. Therefore they never engaged in a project without being paid by the customer. It didn’t mean that the bill necessarily covered all their cost, but it served as an indicator that the customer had thought carefully about the experiment, and found it worthwhile. 

At the beginning, Allthings accepted to do any project on this criteria alone: It had to be roughly convincing, and the customer had to pay something. Once the first experiments started producing results, they could start saying no to that part of the spectrum which evidently was about repeating things that they already knew didn’t work. 

A few months later, Stefan was able to come to some more general conclusions, and they were not necessarily promising. First, the reality of scalability was flawed. Each business, no matter how similar their product was to others, wanted specific, customised features that went beyond Allthings’ architecture of re-usable building blocks. This also meant that the variety of possibilities still outran even Allthings’ imagination. It was not enough to scan the world and estimate the potential of each product you could make out, in order to actively target the most promising customers. Whenever you sat down at the customer’s table, the reality would again be different from what you imagined. In addition, for many products the platform was simply never going to be economically feasible, because it would demand a too high share of cost. If the customer produces something at the cost of a few dozen Swiss Francs, any kind of monthly fee per registration weighs too much in their calculation. Conclusions such as these represented valuable learnings, but for the time they did not show a feasible way forward. 

What happened next was pure serendipity. A very weak signal hitting a very sensitive ear. At one of their children’s party, Stefan got to chat with another father about their jobs. As it turned out, the other father worked as a project leader for a large housing development company. And as they explained their respective challenges and realities to each other, in Stefan’s head a whole handful of pennies dropped at the same time. What is particularly noteworthy about this encounter: years before, Stefan’s team had already looked into housing. Things like an automated noticeboard or an online assistant caretaker had looked interesting for a short moment, but did not pass the scalability test. And exactly there was the difference to the person he was talking to: scale. In Switzerland, the typical apartment building has six to twenty apartments, with less than 1% of buildings over that number. But the project in question was about 200 apartments, with an overall objective for the whole development of over 700. Beyond that, other housing developments would have very similar needs for additional services. And the share-of-cost issue, which was a business case-killer for relatively cheap objects, is a no-brainer once you are talking about a monthly fee within the per mill range of an average Swiss rent of over 1500.- CHF. 

The stack of virtual layers Allthings could add to the apartment is impressing. Think about the pile of paper tenants get with the key to a new apartment. There are the house rules, the manuals for all the electrical appliances in the apartment. Usually, about six months after people move in, the real estate manager has to spend hours on end answering to requests to replace lost copies of user manuals. Now people get a login key to an online platform. All the necessary documents are always there, always up to date. You can add substantial services such as a pooling of household insurances to get better rates. You can add a webshop, so that when the first toothbrush holder or soap dispenser that came with the bathroom design is broken, people no longer have to go looking for a replacement in the DYI shop, which as a matter of course never fits. You get an online ticketing form to alert the caretaker, and optimise it in such a way that in certain cases, they can directly send the plumber, carpenter or other, avoiding the caretaker to take a trip just to confirm the obvious. Next, you can encourage responsible behaviour in the smart energy building, by letting tenants monitor the energy consumption of their apartment, and maybe compare it against the average in the estate. You can add platforms to build community life: announcements of parties to take place on the shared playground, online flea markets if you are looking for someone to take over the safety seat your kid has grown out of, or your dog’s puppies. You can even encourage the sharing economy, if you think of how many people own an electric drill and other tools, and use them for five minutes a year. The benefits were innumerable, not just from a tenant’s perspective, but also from a real estate management’s point of view: In new buildings, up to 100% of tenants registered within 3 weeks. 

What they soon learned during this phase of focused exploration (as a kind of bridge between exploration and exploitation), was that they had to re-think the fundamentals of what business they were in. It was not a B2B2C business, but they were a multi-sided platform provider. The property developers were the entry gate, but at some point in the project they were gone. The tenants were the main users, but they were not the ones to make the buying decision. The property and facility management companies were another type of user, but they also did not decide. The ones who decided were the real estate owners, investors and asset managers. And once you looked at the situation from their perspective, you discovered another type of benefit: analytics. Real estate owners are interested in understanding what motivates tenants, how they tick. 

With this experience, the new business model was clear and focused in a way it never had been before: Become the standard for digital tenant relations and service management in real estate, target large settlements and large owners. With this choice of focus, the rules of the game changed fundamentally in comparison to how Allthings had behaved before. The shift from exploration to exploitation had become decisive. The owners’ interests still called for a certain level of diversification, but it was limited: some features changed once you moved from housing to office buildings, or to retail or other commercial tenants. But by and large, the requirements for tenant relation management were the same in the whole real estate industry. 

The shift to exploitation started with management. While in the early years, Stefan had probably been the key driver of variety in the company, now he became the one to tone down the still existing tension between the variety of customer needs and the efficiency of the internal landscape at Allthings. Part of this change of role came with size. From this moment, growth was much steeper, and the strategy was focused on accelerating it. This implied changes in management activities. Sensemaking, making sure that staff had a common idea of the nature of the business and how it was changing, had happened pretty much on its own when the company was small, despite the fact that they were running in two separate locations from the very start. Now, it had become an important management task, not the least due to the fact that the larger company started attracting other kinds of people. At the beginning, it was an almost unspoken prerequisite for staff to have a startup mindset, which included open attitudes towards risk, self dependency, improvisation and the lack of structure. Now, both the reality in the company and people’s mindset changed. For the management team it was clear how important it was to over-communicate across locations, to spend much time on the common understanding of the “Why” for so many of the things going on in the company. For example, it is clear that tenants and the asset management companies owning their houses have fundamentally different mentalities, conflicting interests, and a critical view on each other. These points of view tended to replicate within the company, depending on which end of the process people were working on. It was essential to relentlessly underline the common interests, and the win-win potential in what Allthings did, for example if they provided a ticketing system so that damages were identified early and requests were actually dealt with in time. 

Size also meant that the business started generating more data, and it became possible to take data-driven decisions. But the interpretation of data had become less self-evident, because more people were further away from the context that generated the data, or had a partial view on this context. Like the common picture mentioned above, making sure that people had enough informal, qualitative customer knowledge to be able to interpret data, had shifted from being self-evident to becoming a management task.

Looking back on the early history of Allthings, it is important to note that the coherence of the story as it is told now is something that happens only in retrospect. Stefan is able to tell which events make sense as a crucial step in the journey, and to leave others out. But not only could he not have known that before. At any time, the present, and maybe the short term past, too, looked to him much more disorienting than the coherent story we are getting now. For example, the understanding that the better way to think of their business was not in terms of a B2B2C model, but as a multi-sided platform provider, was something that occurred to them only when in reality they were doing that very business for quite some time – only with somewhat less orientation, more incoherent reflections, and probably some less educated decisions on the way. Stories like this can make you understand what can happen. They don’t tell you whether, when or how they actually will occur.

November 7, 2019

The Sorcerer’s Apprentice in a Knight’s Armour

At the peak of the Middle Ages, the Mallorcan monk Ramon Lull wrote a small treatise called The book of the order of chivalry, which became a cornerstone of the genre. In Medieval society, the deal was that peasants farmed the land and left part of their income to those of noble birth – the knights – in exchange for protection. Why then, Llull wrote, was it so important that the knights were instructed about moral virtues? It was not that they were better people. But with their horse and their armour, they were people of bigger potential. A bad knight had a worse impact than a bad peasant. A good knight had a better impact than a good peasant. For this reason, it was more important for everyone to make sure knights became good knights, and demanded higher standards of them than of ordinary men. 

I was recently reminded of that rationale when thinking about the rising potential of modern mankind through the use of technology. Like the knight with horse and armour, in today’s world we command a rising arsenal of objects that increase our impact – Sigmund Freud called us “a prostetic God”. We have immense strength with the use of robots and motors. We can fly with the use of a plane. We can be omnipresent with the use of telecommunication.

But we have more in common with the metaphoric knight than that. The stronger and more protective a knight’s armour became over the centuries, the more limited and partial was their perception. Their field of vision was a small slot in the helmet, all the while the helmet was bumping up and down with the galloping horse. The armour was clanking so loud it was impossible to hear what was going on behind you in the melee. So while the reinforced armour made the knight’s thrust more powerful, it certainly rendered it less precise. Only that did not cool down most knight’s impression of being all-powerful, of being able to act, to compensate any lack of understanding with a healthy dose of testosterone.

I think this pretty well sums up how many of us feel about using modern technology to face complex challenges. 

Today’s technology being more subtle, more digital, more immaterial, the metaphor shifts from the armoured knight to the Sorcerer’s apprentice – maybe taking the healthy dose of testosterone along he way. We tend to overlook consequences of our own actions. Partly because at the beginning, they are practically invisible. We are confronted with a huge amount of information, and there are virtually no relevance filters. Partly, because we tend to have a too narrow view of our playing field. We believe we can disregard what is happening in the wider context. One typical area where this is happening is the interface between technology, business, and politics. 

The Sorcerer’s Apprentice from Disney’s Fantasia

One of the big changes brought about by digitalisation is that it fundamentally increases our range: easy access to more information, or easy access to a large audience. Businesses provide this access, and build models about how to monetize on it. But what also happens is that changes in access may affect changes in selection, or changes in visibility, and as a consequence affect what information does in our lives. 

In 2015, in the race to attend its goal of one billion hours a day of viewer engagement, Youtube changed the algorithm that selected videos for recommendation (the ones that appear on the list on your screen, and those that appear on the main screen once you have finished watching a video). The recommendations have been Youtube’s most successful instrument to keep user’s eyes on the screen, maybe longer than they originally intended – the main commercial objective. But with the old algorithm, those videos that were seen most often, were recommended most often. Over time, recommendations would converge on the mainstream, and be considered a bit boring by many viewers. The new algorithm changed that. It still showed many of the most viewed videos in the context of the one you were watching, but it also showed some of the more marginal ones, maybe less closely related to your topic, but most importantly watched by fewer people. Commercially, the change was a success. By October 2016, Youtube had reached its goal of a billion hours a day. But what happened to the content that people were watching? Well, let’s ask ourselves the simple question: what kind of content, in a given area, is viewed less than the mainstream? Probably the more boring things, and the more extreme. If, amongst the typical mainstream recommendations, you get to see thumbnails of some boring videos, and of some extreme, which of the two are more likely to be clicked? The extreme – for two reasons. First, you see the title and a picture, you don’t see the video yet. And extreme videos tend to have catchy titles. Second, the psychological law that describes what catches our interest, and makes us react and engage: emotions work better than rational, cognitive stimuli. And amongst the emotions, negative ones such as insolence and hatred work better than positive ones such as curiosity and joy. Therefore, in the wake of Youtube’s change of algorithm, obscure youtubers with extreme, polarising standpoints, hate speech (as long as it was just within the rules), conspiracy theories and wild accusations, made the career of their lives. Many, who before were on Youtube as a side job, could now make a living off it. Youtube, as an enabler, contributed to the division and polarisation of our society. Was it their intention? Not at all. They were just pursueing the innocuous task of increasing viewer engagement.   

July 24, 2019

So many Fish in the Ocean – somewhere

I am sitting on a beach and watching the ocean. On rare occasions, I can see a fish jump out of the water, not too far away. I imagine all the fish that are under water, just out of sight to me. Some are close together in big swarms, some are alone. And now I imagine being a fisherman, on my tiny little boat, crossing that huge expanse of water. There are definitely lots of fish down there, just under the surface. But in the absence of tech gizmos such as radar, sonar and the like, I don’t know where to find the fish. How can I know where the swarms are? How can I know what to do? If I am not catching any fish, is it because there are no fish in this area, or because there is something wrong with my bait?

Now I get lucky, and I catch some fish – quite a few, actually. Now I not only have fish, I have information. The information that what I just did was successful. But what does this mean, how can I make this information useful for my next decision? I have to find a pattern. What if I stay in the area, will I still catch many fish? What if I come back here tomorrow? Should I switch off the motor of my boat, because it shies away the fish, or should I keep it running, because the sound attracts them?

When we face complex situations, we are often just a little fisherman on the ocean. Objectively, there is a lot of information in the system: where the fish are, how they behave. But we don’t have access to the relevant information. All we see is that there is a lot of water, everywhere. Objectively, we have many many options to choose from: go close to that spot three miles out where the cold deepwater current comes up, or go slightly to the left of it, just over the underwater cliff. But since we only see the surface, it actually does not matter which nuance we choose, we don’t know what the implications of these nuances are anyway. And that leaves us in fact with very few options indeed: stay where we are, go somewhere else, keep moving, somehow. 

To the bureaucratic mind, this feels very uncomfortable.

Only in retrospect do we tend to see everything as a cause to the outcome. When we caught a lot of fish, we conclude that we evidently did make the right decisions. But in truth we have no clue which of these decisions were important, which were of no consequence whatsoever, and which were actually keeping us from an even better result. We still don’t understand the intricate interdependencies that existed that day in the ocean, and even if we did understand what happened on that one day, we have no indication whether this course of events will repeat the next day, or not. We are victims to the teleological fallacy, the tendency to retrospectively see everything as a relevant means to the outcome. 

The conclusion, for the moment, is humility and hope. If I am not catching any fish here, I move somewhere else, acknowledging my ignorance, accepting the possibility that maybe a big swarm was just coming my way, and I am now steering away from it – and I will never know. This acceptance of my ignorance is the basis of that little informed action I can take. I do get rid of the bait that has never worked so far. I do come back to a spot that was relatively successful so far, although I have no idea that it has to do with underwater currents and temperature. When I accept that everything I do is a more or less educated guess, I am much more open to adapting my plan. The options at my disposal actually become few, not many, and I consider them as constantly changing. As the economist John Kay noted in his wonderful book Obliquity, in complex situations even the most powerful men such as Lincoln or Roosevelt must proceed by choosing opportunistically from a narrow range of options.

June 10, 2019

Nudge the System, Wag the Dog

When we think about perception, our first question is whether it is true or not. Our focus is on the relation of perception to the perceived reality. But often we forget about perception’s other side. On the other side, our perception is connected to our action, and this, in turn, influences reality. Let me show you what I mean. 

Imagine there is a glass full of water in front of me. But I assume that the glass is empty. My perception of the glass is objectively wrong. But as Epictetus says: People are moved not by things, but by the view which they take of them. Acting on the basis of my perception that the glass is empty, I take it and hold it upside down. In so doing, I actually create the reality that just a moment ago was my wrong perception. Now my perception is correct: the glass is empty. And yes, I may discover there is water on the floor, but as long as I only consider the glass important, that doesn’t matter.

One step up, we can find the famous Vase scene in the movie “The Matrix”. Neo has just entered the kitchen of the woman called the Oracle, and while she casually talks about cookies, she says: “…and don’t worry about the vase.” – “Which vase?” asks Neo, and turns around to look. In so doing, he knocks over the vase that was just next to him.  – “That vase.” And there, in a nutshell, is the key to the whole movie: The prophecy the Oracle is going to make, about Neo not being “the One”, will eventually be a crucial influence on his behaviour which will make him become “the One”.  

Now let’s take this from fiction to reality: A few months after his election to the presidency, Donald Trump said that Sweden had problems with violence connected to immigration. This perception at the time was wrong, and based on some fake or exaggerated news from one of his favorite sources. But now that the President of the United States had said it, it became something people reacted to. The Swedish government, indignant, denied it. Fox News set up an interview with a “Swedish government expert” who happened to be someone who once did a project for a Swedish government institution. He confirmed Trumps view, and as a consequence depicted the Swedish government’s denial as a tentative hush-over. Since by now everybody was on edge about the topic, one week later a small incident about how Swedish police treated a foreigner led to violent protests. Now Sweden did in fact have a problem with violence related to immigration… News produces its own reality, like in the movie “Wag the Dog”. 

While in this case things happen very fast, they appear more regularly, and less visibly, in systems that develop over a longer period of time. Think about how parents educate their kids’ attitude towards risk, for example in traffic. At the start, everyone’s assumption about how their two-year-old toddler is capable to move in traffic is the same: at that age, kids are mostly oblivious to oncoming cars and bikes. But the conclusion from perception to action is different. Some parents think that for a long time, the only way a kid can move on or near a road is with constant attention and intervention of an adult, and the best way is to take them by their hand. Others think different, expose their kid to a certain risk, and gradually increase their independence. If you compare the kids’ behaviour at the age of 6, both parents see their views confirmed: the child that was always guarded and guided, is objectively much less capable of moving in traffic than the other. Similar behaviour can be observed on the playground, when it comes to climbing objects and jumping down off them. Kids whose reality consists of having an adult constantly two meters behind them, crying out “Don’t go there! Don’t do that!” whenever it might get dangerous, will learn to rely on that warning signal, and not develop one of their own. The ones who are more independent, may sometimes hurt themselves, but can be seen to ask for help if they assess the jump as too difficult. 

Once we understand this way that agent-system coevolution plays out, we can use it for our purposes. In my case, it was precisely the Matrix that opened my eyes – by having arguably the most surreal effect a film ever had on me. It was in Paris, and after the closing scene where Neo stopped the bullets, I walked out of the cinema with this sense that reality can be influenced. In the crowd in front of the Odéon, I walked straight, shoulders wide, looking right through people – and they all stepped out of my way. Then I became bolder. I walked onto the street, in between traffic lights, gave clear signs with my hand to the cars, never looking at the driver for more than a split second – and they slowed down, and drove around me. After five minutes, slowly the thought came creeping in: and what if one of the drivers doesn’t see me, or doesn’t believe that I will not stop? And all the puffed-up sense of self-assertion ebbed off again…

While this example may make the point, here are some more practical ways in which this can play out. My friend Diane has a habit of being very trusting, even in situations where others would have backed off long ago. But here is how she explains it: “If you trust people – from the start, and sometimes relentlessly – then some may call this risky, and sometimes even naïve. But it is amazing how many people want to be worthy of your trust.”

There is an episode in the life of Hugh Bishop of Lincoln in the Twelfth Century, when he was in conflict with Henry II., King of England. The conflict had escalated to the point where the Bishop had threatened to excommunicate the King, and the King had threatened to remove the Bishop from office. As a result of a difficult mediation, Hugh had the chance to come to court to make his excuses to Henry, with all the powerful nobles and magistrates present. But given the proud nature of both King and Bishop, it was quite uncertain whether, or how sincerely Hugh would apologize, whether Henry would accept, and which way the opinion of the court would lean. The King, and most of the nobles, were of the firm opinion that Hugh needed a beating, and needed to work hard to earn back Henry’s trust and friendship that he had once enjoyed. So when the Bishop entered the throne room, the King sat on his throne paying no attention to him, mending a button on his coat. Two notes on the context for the modern reader: In the Twelfth Century, it possible for a King to mend his own clothes, but it was very uncommon to do so in public, on the throne. Second, William the Conqueror, Henry’s grandfather, was an illegitimate son, born by Arlette de Falaise, a woman of humble origin. Now Hugh enters the throne room, sees everybody in silence, and the King is mending his clothes. The tension rises, everybody awaits Hugh’s reaction. Then Hugh says: “How you remind me of your cousins of Falaise.” In the context of a humble return of an outcast, this remark was insolent to the extreme. But in another context, that of old friends, of a Bishop who was allowed certain personal jokes that others were not, it was ok. The surprise was so big that Henry couldn’t help laughing out loud. And that created the re-framing. The court joined in the change of mood, and Hugh was forgiven without further rites of humiliation. 

I remember when I was in a situation with a client. The team working with me introduced me to their top manager. I was to join him in a workshop with many people, and the success of the workshop depended partly on how much the people in the room trusted me to have a close relation with their top manager – on eye level, so that I could, if need be, stand up to him and push back. In the very short briefing I had with him, I did several things very consciously: contradict him, make a casual remark to a weakness both he and I had, make a remark that showed I saw behind his public façade, while at the same time considering it perfectly normal to speak about this. Each of these actions were just a bit unacceptable for a first meeting, but would have been normal under the assumption that the two of us had a developed, trustful and buddy-like relation. He could, of course, decide to confront me there, or at least question me internally. Or he could go with the flow, which meant participating in a kind of behaviour that could only be thought of as normal if he assumed that we had a good relation. And that was what he did. Our internal model said that this behaviour was the outcome of a good relation. But since it was the behaviour that started it, it actually created the only condition under which it could be conceived as acceptable. 

Agent-system coevolution, the fact that one reality is open to many perspectives, and the role of internal models in directing our behaviour, combine into a toolset which allows us to consider many situations as an occasion to nudge the system. It is up to us to use it. 

February 17, 2019

How to be a Servant Leader, Center Stage

I don’t like the concept of servant leader too much. I think it is often over-stretched, and unduly loaded with moral baggage.

But there is one thing I am convinced a leader should do, and should be capable of doing: promote greatness in others. The first step to do this is often difficult enough for many: step back, and leave space for others. But then the next question arises: once you have left space, what can you do more?

Here’s an inspiration as to what you can do. In 2007, 21-year old bass player Tal Wilkenfeld joined Rock legend Jeff Beck for a European tour, this being the second major engagement in her career. At the end of the tour, they appeared at the Crossroads festival, organised by Eric Clapton, Jeff’s predecessor as guitarist of the Yardbirds, back in the ’60ies. In their first tune, Jeff opens with his signature intro to “Cause We’ve Ended As Lovers”, and after just one chorus, hands over to Tal. Notice what Jeff does during her solo. Think what effect it has on the audience’s attention, on how they think about it. Basically he is being a role model to the audience. Notice also how it encourages Tal to increase the intensity of her playing for the second chorus, and how the drummer Vinnie Colaiuta undermines it with his play. At one moment, you can see keyboarder Jason Rebello join Jeff in his mixture of admiring and encouraging observation.

This is how it’s done.

February 3, 2019

How to Achieve Aligned Autonomy

Imagine you are a shop assistant in a telco shop. A customer walks in with a broken mobile phone. It is clearly a warranty case. Unfortunately, his two-year warranty has expired a month ago. He claims he has been abroad on another continent for two months, he couldn’t come earlier. What do you do? Do you insist on the warranty terms, or do you replace the phone for free? 

If we put ourselves in the shop assistant’s place, the first thing we want to know is: are we allowed to take a decision? What is the level of delegation, and conseqently of autonomy and flexibility we can take in this case? As we have explored in a previous article on delegation, if the company wants to react to exceptions and unforeseen events, it better delegate autonomy to the front. And since in our case the company has taken this lesson to heart, we do in fact have the autonomy to extend the warranty ex gratia in appropriate, justified cases. 

Is this enough for us to take our decision? We may follow our personal idea of customer satisfaction vs. cost savings, and decide. But from a company’s point of view, this is not enough. Think about the brand’s identity on the market. Whatever the company stands for, it is essential that the identity is seen clearly and strongly by the customers. How does this happen? It is not so much through advertising campaigns, or the visibility of the logo. Rather, it is the consistency of customer experiences such as the warranty case, that creates a strong brand. You see an advert, you go to the shop, you call the company, you get a service person to your home – and in all these cases you have a consistent sense of what the company stands for. You get the company’s “groove”. 

Therefore, as a company we don’t want employees to decide how they see fit for themselves. We want them to decide how they see fit in the light of the company’s identity. In our case, the question is how the company stands in the tension between customer satisfaction and cost. If we are a shop assistant working for a discounter, we may want to be more restrictive – and we could explain to the customer that in order to achieve the low prices which they so much appreciate, we unfortunately have to be strict with wavers of replacement costs. On the other hand, if our company is in the premium segment, we will replace the phone without hesitation, because customers expect a premium service for a premium price. The company’s identity serves as an anchor to guide behaviour. 

So much for the company’s intention. 

I happen to have had a similar customer experience with a network provider which was positioned in the premium segment. I moved house, and in the process bought a longer router cable in the shop. By sheer luck, I tried to connect the router to the cable a few days before the movers came. To my surprise the cable didn’t fit the socket. I found out that as an business customer, I needed another cable than the residential customers, though the router looked the same. Could I expect the shop assistant to know about that? No, because usually business customers do not go to a shop. But what next? I phoned my support, and the right cable arrived just in time before the movers put a huge wardrobe in front of my network socket. So far so good. 

But by now my flat was full of cardboard boxes, and as you can imagine it took two months until I found the cable again, together with the receipt. Although the usual period for returning unused goods is 30 days, I had the above story of a warranty waver in mind and went to the shop to have my money back. Here is my experience: 

First, the shop assistant checked my story in the computer system: Yes, he could confirm that I had in fact moved house two months ago. Then he excused himself to go discuss the case with his boss. When he came back, he gave me the money, saying: “We accept your claim this time, Sir. But note that we are not obliged to do so. Next time you will want to be more careful and bring back the cable in time. We have to be very scrutinizing, because we have too many customers who try to trick us.” I was speechless. 

What had just happened? 

Let’s look at this story from a manager’s point of view. In the structural sense of delegation, everything had worked fine. The shop assistant did have the authority to take the cable back after the usual 30 day return period. He did apply this authority to my case, and took back the item. If the manager were to consult the log of the shop’s transactions, he would get the impression that everythig was all right. 

But of course it wasn’t. 

With his behaviour and comment, the shop assistant nearly lost himself a customer. He conveyed the exact opposite of the “premium service for premium price” identity of his company. In a way, he destroyed the very reason why the authority to be flexible had been delegated to him in the first place. He did not have the correct anchor to guide his behaviour. 

Now if, as a manager, we are aware of such cases – how do we improve the situation? What can we do in order to achieve more alignment, more appropriate behaviour in the light of the company’s identity? If we apply any kind of rule or regulation, in the sense of “These kind of cases must always be dealt with like that!”, we in fact withdraw delegation, and move our case towards instructions – the capability to deal with the specific nature of each context is gone. The challenge is to improve alignment, without taking away autonomy. Here’s a method by which we can do so. I call it anchoring. It consists in systematically anchoring the preferred course of action in a specific situation to a general, purpose-giving statement. Many good leaders practice anchoring off the cuff, without thinking.

Anchoring, step one: Timeout Discussions

The key lies in the discussion of individual trade-off cases. Just like the broken phone, also my cable story could be a case worth discussing from the point of view of a shop assistant. Take the team together, and offer the situation for decision, exactly like I did at the beginning: “A customer walks in with a broken mobile phone… – what do you do?” Three things are noteworthy about the choice of case, and the way to present it. First of all, it is a case. Why is this noteworthy? For whatever reason, many managers seem to shy away from the individual example. There is a tendency to drift towards the generic, towards the one rule that covers them all, towards making one statement which is valid for many cases. And, of course, a large part of management practice is about that. But not here. We have to choose an example, knowing that there are many other examples which are slightly different.

Second, the case represents a relevant trade-off decision. Both the broken phone and my cable are examples of the trade-off customer satisfaction vs. cost. Other trade-offs may be speed vs. security. Long-term vs. short term advantages. Standard vs. exception. Our team’s problem vs. a problem for someone else in the value chain. Etc. etc. The reason for choosing our case is exactly that we have noticed, or suspect, a need of alignment, therefore we expect the team members to propose different approaches to the case, and not all will be equally apt.  

Third, although we most probably base our case on a past observation, it is important to accentuate the trade-off, stop at the moment of decision, and not give any indication as to how we, or the person we observed, have reacted in the real story. From now on, the real story is of little importance – it’s about the way the situation challenges our values. 

The next step is to collect the diversity of answers around the table. I have seen teams simply tell their answer one after the other. While this approach is simple and fast, the risk is that as soon as the first person begins to speak, the others quietly alter their proposed approach based on what they hear. Since at this point we want the diversity of uninfluenced approaches, we need some maturity, and psychological safety in the team to let everyone understand that they should tell their original idea, no matter what the others say, because modifying their idea at this point would defy the purpose of the exercise. Other teams I have observed have team members write down their answer on a paper, and then read them aloud. In any case, it is important that we don’t start a discussion before everyone has stated their approach. 

Now that we have the diversity of the team’s approaches on the table, we can open the debate: What is particularly good, what is ok, what is not such a good idea, and what is a no-go? Here, obviously, it is important to let the team speak their mind first, and accept anything that is in line with the company’s intent. Only in case the team comes to a conclusion which goes against this intent, is it time for the team leader to exchange their facilitator role against the role of a judge, and contradict, or adjust, the team’s dominant opinion. In such a case, we take care to legitimize our statement with the anchor, which is described below. At all cost, we need to avoid that the team sees our intervention as arbitrary. 

Another thing we don’t do, is end the session with a verdict – and again this is in fundamental opposition to many management practices, where it’s all about decisions. It is not the point of these discussios to come up with a regulation, because no matter what, the next case will again be slightly different. It is exactly the intent of this method to prepare for the next slightly different case. If somebody has participated in timeout discussions of ten cases around a certain topic, when they meet the eleventh case in practice, and stand alone to decide, their decision is in fact substantially more aligned to the intent of the company than before. 

To sum up, we may note that the practice of timeout discussions follows four principles which go against what I observe as widespread habits of management. I call them the four EX: 

Experience: Focus on collecting individual experiences and augmenting them in exchange with other experiences instead of devising instructions top-down or out of theory.

Example: Focus on alignment across a series of examples instead of creating the one generic rule that is valid for all cases. 

Exchange: Focus on exchange and influence in teams instead of individuals learning by themselves. 

Experiment: Focus on experimenting and adapting approaches instead of one upfront solution that guarantees success. 

By applying these principles to our timeout discussions, step by step we increase the alignment across the team, without inhibiting the autonomy of decision that is part of our delegation, and thus we uphold the organisation’s capacity to react to the exceptional and the unpredicted. 

There is, however, another part of the method that remains to be established – the anchor itself: the strategic storyline. 

Anchoring, step two: Strategic Storylines

We are in the middle of a timeout discussion. We have chosen our case well, which means there are many different, contradicting courses of action on the table. People argue with conviction, revealing the different values that drive their behaviour. At one point we may invite the team to search for alignment, to come up with some authoritative statement, that helps prefer certain proposed behaviours over others. What is the appropriate form of this authoritative statement? Take for example the above case of the broken phone. WHY is it better to replace the phone? If the answer is just “Because that’s what the company wants” we’ll never get people to use this information for anything else but phone replacements. Team members may even come back and ask for precision. Until how many days after the warranty has expired? What is the maximum price of the phone? How convincing must the customer’s reasons be? – And that is the wrong approach. By comparison, observe the following explanation: “The market for mobile phone subscriptions is dominated by an obsession with price. We are a premium supplier, that means we manage to defend our market position, although we are never the cheapest. This only works under two conditions: First, we have to offer premium quality both of product and service to legitimise our premium price. Second, on every occasion must we lead our customer’s focus away from the price, and towards quality, where we outshine the competition. Our existing customers are very quality aware. By emphasizing the quality argument, we respond to their expectation.”

What is the difference between these two explanations? The second version connects the phone replacement to the company’s strategic positioning – the replacement is not just the thing to do. It makes sense. By comparison, in the first version the connection is just to the company’s intention, no matter how arbitrary this intention may be. This has fundamental pragmatic implications: a team member can use and apply the second explanation for so many more situations, while the first explanation is only valid for replacing phones. Carsten Schloter, CEO of Swisscom at the time and one of the top managers that I’ve observed using strategic storylines, guessed that every company has probably no more than half a dozen storylines on corporate level, that provide orientation and meaning for almost every trade-off decision in the organisation. Next to the corporate storylines, there may be some more storylines on local level, such as the need to replace IT systems, or the fact that you’re the small foreign subsidiary of a large national company, and therefore often overlooked when it comes to applying company decisions abroad. In any case, for an employee there is a manageable number of such storylines that help them make sense of decisions and priorities. A storyline is like a reservoir out of which we can draw the justification of behaviour in an individual case. While discussing the case, the relevant explanation can take as little as one sentence. But the reservoir holds more. As we will see, the full potential of the storyline as an anchor lies in its repeated application to many cases, and that this repetition is recognised by the audience – again: if we have heard the same storyline applied ten times, the eleventh time we can do it ourselves. It is therefore worth thinking them through at a level which may rarely be applied in everyday conversations, but which does justice to its versatility as a reservoir.

Sometimes people make the mistake of creating one big storyline for everything in the company. The consequence is that they mix so many aspects of the company’s reality that it is difficult to understand which parts of it give meaning to which other parts. An essential starting point for a storyline is therefore the choice of focus. What aspect of our company’s reality do we want to emphasise? In the above case we want to explain what it means to be premium. Another company might have a storyline which explains the consequences of growing more internationally than in the domestic market. A car manufacturer may have one to explain the consequences of having a mind-blowing number of product options to choose from. An IT company may have one which explains dealing with the insane speed of technical evolution.  

Once the focus is clear, a good storyline has between two and four building blocks, that each respond to their own question. Let us look at them one by one.

  1. What is going on out there? Anyone’s actions only make sense in their context. Therefore one should not start with what the company does, but what in the company’s context makes its actions look like a good idea. Some may be about the product, the customer, rules of the game of the market, regulations, or supplies. They may have to do with the strategic positioning of the company, or simply with how the business works, e.g. what the key customer interests are. For example, an insurance may start a storyline with the fact that customers are not really interested in insurances, they just want to have one so they don’t have to think about it anymore. And their most regular form of customer interaction is a bill. A hotel having recently become part of a chain may have a storyline about the leverage of customer loyalty, pointing out that if a customer is loyal to a chain, they may prefer a hotel of this chain wherever they go on the globe, whereas for an independent hotel, customer loyalty only makes a difference on the occasions that customers happen to come back to visit the same town. A producer of machine equipment may have a storyline which starts with the fact that due to digitalisation, batch sizes along the value chain become smaller, and therefore the cost of retooling between two batches becomes a more important factor of the cost of each piece. It is important, in this part of the storyline, not to describe the environment in general, but only that part which explains why the next step is a good idea.
  2. How do we respond? The second step is inward looking. It describes what the company does in order to respond to the context shown in step one. The insurance company may point out that in all interactions apart from sales, the most important effect to create in the customer is ease. Ease of payment, ease of getting information if needed, ease of handing over the paperwork in a case of damage. Therefore, it is important to optimise end-to-end processes across silos, and to avoid creating difficulties, or questions, in other parts of the company. The hotel may point out to what length they expect employees to go in order to produce customer loyalty, what means they are given to do so, and how they should increase interest in the chain, not just in the hotel where they are staying. The machine equipment producer may emphasise the focus on lead time, and that it’s worth the effort to group similar batches in order to reduce retooling time per batch. In this part, it usually happens in one or two steps that the storyline moves from a general statement about the company, to the context for which we want to make this relevant. In the insurance company, the same storyline becomes to the front-end employees a tale about understanding the customer’s needs and reducing the number of interactions. For the back-office, it may be more about process optimisation. The anchor provides relevant orientation for the behaviour we focus on.
  3. How do we tick in order to make this happen? The third step is optional, and gives a nice twist on how culture can be influenced, not by talking about culture, but by talking about the business. If there is a conclusion to be drawn from the first two points towards mindset and behaviour, then it is helpful to make this obvious. In the first example, making it easy for the customer implies elements of empathy – understanding the customer’s needs, and gaining trust: we are the experts who want to be trusted by the customer – in order to do that we need even more trust across our departments. With the hotel chain it may be about learning about the individual customer and understanding what is important to them. Engaging into conversations with customers has a lot to do with the self-confidence of employees. And in the factory, it may be about a culture of responsibility and collaboration in order to take initiatives that go beyond the blind execution of individual tasks.
  4. How do we lead in order to make this happen? A second optional element actually depends on the people to whom the storyline is addressed. If they lead others, then it is often possible to draw conclusions to particular aspects or tasks of leadership which become essential – usually related to step three. If you want to increase cross-departmental trust in an insurance company, you should stop the blame game whenever you see it, let people work more together, let your team discover the effects of their work across the value stream etc. If it is about increasing the self-confidence of hotel staff, it has much to do with treating them on eye-level, with how you delegate, praise and blame. And the initiatives you desire in the factory happen more if they are encouraged, and if people have direct access to other departments.

With this structure in mind, we now have he anchor around which to encourage alignment. The key here is repetition. People should recognise the same few storylines in as many occasions as possible. Of course, a systematic use of timeout discussions is a good way to get there. But it can go beyond that. A person is promoted, another one leaves for retirement. In both cases we are asked to give a speech. There are a thousand aspects which we could choose to praise – why not connect the ones we choose to a storyline? Why are we stopping this project, and increasing the resources in that other one – explain it with a storyline. The more we move from the team to higher, more visible levels of management, the more important is the coherence of messages sent into the organisation. Carsten Schloter pointed this out many times:

“Imagine an employee who has the autonomy to act, and they want to take a decision – and they can see: If I choose this path, it will please that top manager, but it will absolutely not please that other top manager. In such a case, the employee will not want to decide. They will try to move the delegation back up the chain of command – the organisation becomes stiff and ineffective. Therefore, it it is absolutely essential that the management team does not send contradicting messages into the organisation concerning the key trade-offs of the company.”

As a consequence, Schloter organised several top management off-sites per year where they conducted something similar to timeout-discussions. They discussed two questions: What is the impact of employee behaviour on the business? And what is the impact of top management behaviour on employee behaviour?

If we take this back to timeout discussions, we can see that for top management, it is not only helpful to discuss cases of leadership, but to discuss cases of the employees in various parts of the company.

In other cases, the anchoring and the alignment effect happens simply because top managers tell stories with the same strategic storyline as anchor. For example, Urs Berger, when he was Swiss CEO of Baloise, had a storyline under the title of “High tech – high touch”. What he meant was that in times of the first e-commerce bubble, many customers wanted to move to digital access, and in the process also called more into the hotline in comparison to the old-fashioned sales agent, or visits to the general agency (step one in the storyline). High tech – high touch meant that the company wanted to be present on both channels, and let the customer choose. This implied for example that there was no difference in prices, although the cost of the various channels was very diverse (step two in the storyline). Berger used this anchor, and the motto, on every occasion he could. I remember once he unveiled a “damage mobile”, a trailer that the company wanted to use in cases of hailstorms – when hundreds of cars had to be checked for hail damage in a region. The trailer could be placed somewhere in the area, customers were informed and brought their cars for inspection. On one side, the trailer had all the equipment necessary for the inspection – notably the special lamps used to discover small bumps in the paintwork. On the other side, there was a coffee machine and a little counter for the waiting customers, and maybe a sales agent who was ready to talk to them about household insurance: high tech – high touch.

Another famous example of the use of storylines were the Dormann Letters. Jürgen Dormann was CEO of ABB in the early 2000s, and in this period sent an e-mail to all employees every Friday. The letters were so famous they were even printed as a book after the end of this period. The key storyline of the period was that ABB did not need a new strategy, but that they had to become much more disciplined in executing on the existing strategy. But very few of the letters are about strategy and the like from a corporate perspective. Most of them are along the lines of: “Last week, I visited a site in the U.S., and here’s what happened…”. Again, it is not the generic statement that covers it all. It is examples, and all of them are different, all of them are typical for something, and not at all typical for many other things. And slowly, the storyline became apparent as the anchor of a series of individual examples.

Today, this idea can be further developed into various forms of management blogs. I have worked with several senior and top management teams who have decided to write a common blog – taking turns in recording examples around a small number of strategic storylines. The effect of this exercise is twofold: First, the staff perceive very clearly the unity of voice amongst the team: different team members may talk about different examples, but in essence, the message is the same. Second, this unity of voice was established in the preparation work. Not the least since many managers do not feel comfortable writing longer stories, they prepared a bunch of blog posts together. And the conversations – guess what! – had an effect similar to timeout discussions: across the examples, team members developed their alignment of mindset and values. And on this basis, their autonomous activities out in the field became more aligned, too.

February 3, 2019

Are You Sure You Really Know How To Delegate?

One of the oldest methods to organise the division of labour is delegation. I have discussed delegation with thousands of managers. My preferred approach to inviting people to use delegation more consciously, is to let them clarify its purpose, and compare it to some of its alternatives. Delegation is a tool to commit useful contributions. A very telling alternative are instructions – including procedures, checklists, workflows etc. When asked about the advantages of delegation over instructions, I invariably get two points. First, people are more motivated when they are delegated a task. They feel ownership, a sense of responsibility, trust being given to them, and therefore self-worth. And second, they can develop their skills over time, whereas with instructions, the development stops once you know how to follow the checklist. After these two answers, the flow of suggestions usually slows down a bit. What is disconcerting about this, is one essential point: the importance of delegation in the face of today’s increasing complexity is only rarely on people’s minds. 

What is the connection between delegation and the VUCA world? Let us have a look at a simple case. I have come to work with my pet monkey. Unfortunately, I am called to an important meeting, and I ask you to look after the monkey for two hours – a delegation. Let’s assume you know how to deal with monkeys, and it’s fine with you. I leave the office, and the animal is looking expetantly at you. What do you do? Do you take her outside, or stay in? If you stay in, she might get difficult to control for her need of exercise. So you take her outside. Here, many things can happen. A lorry arrives on the parking lot to deliver some goods. She gets afraid and nervous. You take her to the park instead. A group of excited small children are running around – if your monkey participates in that play, there might be some bad reactions. Therefore, you take her on the leash. Some people are walking their dogs nearby. You make sure to keep a distance. And so on…

In times of increased complexity, the advantage of delegation has an effect for the organisation as a whole. With delegation, the organisation is capable of dealing with exceptional and unpredictable cases. Contextual decisions can take place at the front.

Why is that different to instructions? Let’s have a look at instructions first. Many routine tasks, especially if they are sensitive to safety or security, are organised by instructions, and wisely so. The checklists pilots use before starting an airplane: They have just checked all these things before the last take-off two hours ago! Yes, you still follow the checklist. Or how to prepare an operating theatre for a certain type of surgery. They know that in this particular case, it is practically certain that this one instrument will not be needed! Yes, and still you are going to put it on the table, and the whole team can rely on the fact that with this type of surgery, that instrument is there. In an instruction, the executor’s obligation is to follow the instruction. If the instruction was wrong, the executor is not responsible for the bad result. This means that the capacity to deal with situations, both in choice of response, and in timeliness, lies to 100% with the instruction’s author. If a surgeon thinks that in this particular intervention, he would like to have an instrument ready which is not on the checklist, she has to tell the nurse. No nurse can be blamed for forgetting to put something on the table which was not on the checklist. What happens if we do blame them, and they need to “disobey” the checklist in order to do the job, can be seen in the case of the Cernobyl catastrophe. The checklists and procedures were so outdated that in order to operate the nuclear power plant, some of them needed to be ignored, and replaced by an experience-based local approach. The authorities back in Moscow designed a test to be run on the power plant, but the assumption of the test designers was that the reality in the power plant corresponded to the procedure they knew. A fatal misunderstanding. Instructions rely on rigidity and obedience. 

Once we move from instructions to delegation, this becomes different: Since the executor has both the responsibility for the result, and the authority to take the necessary decisions to succeed, they will put their power, knowledge, skills and common sense at the service of delivering the desired result – they will do whatever it takes to keep the monkey safe and happy. 

What you always half knew about delegation but never had the time to think through properly

At this point it is helpful to dig a bit deeper into the mechanics of delegation. There is nothing new about delegation as such, but it so happens that many people disregard these mechanics, and therefore delegate badly. As a consequence, the people on the receiving end do not like delegation, taking its bad execution for the thing itself. I would even go so far that much of today’s dislike of hierarchies actually roots in the fact that delegation is done in a faulty way. 

Most delegations happen within a chain of connected responsibilities. The person who delegates a task to us, is herself commited to a task from someone else. I have asked you to look after my monkey, because I am obliged to appear at a meeting. I have given you the task, and I no longer have it. But what about responsibility? Imagine that whilst looking after my monkey, you are called to a meeting, too, and therefore you delegate the monkey to John. You are allowed to do so, because I have not made a delegation ad personam. But now, for whatever reason, John decides to stay inside. After two hours, the monkey is very nervous, and the room is a mess. I come back, and I am not happy. Who do I complain to? Not to John, but to you, because I still hold you accountable for the task. That you have delegated its execution to someone else does not in the slightest change your obligation towards me. 

So when you delegate the task, you hand it over, and you no longer have it. But the responsibility for the result has has not been handed over, but cloned: John is responsible to you for the execution, but you are still accountable to me for the result. 

Given this situation, it is only natural that you care about the execution of your delegation. What does that mean? Think about what is happening to the monkey when John takes her outside. You know what is good for the monkey, it is only natural that you put your experience at the service of the result. The lorry comes, you open the window and remind John that the monkey is afraid, and behind the house there is a park that she usually likes. The children arrive, and you point out to him that it is better to take the monkey on the leash. The dogs arrive, and you warn him to keep a distance. You do this because you care, since you are still accountable for the result. But what happens to John? Step by step, with every intervention in the delegation, in his head the delegation transforms into an instruction. You have told him to put her on the leash? If he won’t let her off the leash for the remainder of the time, you cannot possibly blame him for that. 

Such kinds of interventions can take very subtle forms. John may ask you about how to deal with a particular problem. And since you know it well, you reply to his question. John goes off and puts your suggestion into practice, and all is well. But if the situation evolves and it is no longer sure whether your suggestion is still a good idea, he will hesitate just a bit more to deviate from what you have said, than if it had been his own idea. 

Therefore, together with the task that you delegate, you also hand over the authority to take the necessary decisions to achieve the result. At the same time, you are still fully accountable for the result – and you should keep your hands off from intervening? Truly this is a very uncomfortable situation. 

And the answer to this is: yes, it is uncomfortable. This is the price you pay for the advantage of delegation over instructions, to preserve the investment of John’s competence, creativity and wit, to deal with unexpected situations. 

The damage created by interventions is taken to the next level when we look at interventions that bypass one or several levels of hierarchy. If John is out on the parking lot, and it is me, not you, who remind, point out and warn him about the various risks, I not only destroy the delegation in John’s head and transform it into an instruction. I also damage your relationship to John. If later in the day John has an issue with the monkey, and sees both of us taking a break from our meetings, who will he talk to? Me, not you. Because he has learnt that it is me who calls the shots, me who will intervene and hold him responsible in case he does something wrong. Your authority over John is substantially broken. 

To complicate things even further, there are a few instances where it is perfectly legitimate to intervene into a delegation. The emphasis is on “legitimate”. Feeling on edge is not a legitimate reason. But legal obligations are. In additions, there are three contexts worth keeping in mind, where interventions are appropriate: 

1.     It is ok to intervene if the person is in the process of making a mistake. This is not about vague risks, or that we think our idea works better and therefore everything else is a mistake, but about an evident situation where, on stopping the execution, we can explain the consequences so that the person sees that they would have committed an error. 

2.     It is ok to intervene if their task surpasses either the competence or the authority of the person. In fact in such a case we take back the delegation. If the person deals with a client, and at one point the volume of the contract crosses the line up to which they are allowed to sign, of course we intervene. If the task looked simple, but now has become too complicated for the level of experience of the person, we take it back, and we can legitimise why. 

3.     The third reason is not always welcome, but it becomes very evident in the light of a complex environment: If a change of objectives promises a better result, it is ok to change the objectives. Take the example of a salesperson. By September a main competitor goes bankrupt, and their customers all try to find a new supplier. The salesperson may lean back and think: From now on, I will achieve my annual targets just by picking up the phone. However, if we look at it from the company’s perspective, this event has opened up a window of opportunity: There are customers out there who are very easy to convince, and for the remaining suppliers, the race is on. Once the customers are settled, we will be back to the tedious work of unsticking them from their existing contract, in order to get them to move to us. In this phase of customers on the search, the most active competitor will make a huge gain. We want people to invest as much as they can, to cancel their holidays if necessary, just this once to get as many new contracts as possible. Therefore, if we are sharp, we fundamentally increase their sales target, and we increase their bonus on the sales target exponentially. We can see with this example that there are two cornerstones of this kind of intervention. One, it must be made evident why the change is legitimate. If people think our intervention is arbitrary, we’ve destroyed the delegation, and created cynicism. Two, many systems that govern targets and bonuses, would not allow for such a change – we are back to the incompatibility of bureaucracy with a complex environment. 

Delegation and co-ordination in a self-organised environment

While delegation hits its limitations as an instrument to deal with complex environments wherever bureaucracy cuts its wings and renders it overly rigid, it can boost its ability on the other end of the spectrum. In the age of unknowable futures and emerging strategies, we can take the concept of delegation out of the hierarchical context, and use it more openly: It does not matter whether the definition of the desired result has happened top-down or bottom-up, nor whether it is of small or large granularity. A delegation is a commitment between two people to deliver a contribution, leaving adequate margin to the executor to take the necessary decisions in response to the context.

If you look at delegation this way, you get a very versatile tool. The key is that authority and responsibility is with the executing party, that’s the difference to instructions. Is the delegating party the direct line manager of the executor? Maybe, but not necessarily. Is the executor more de-central, and the delegating party more central in the organisation? Probably, but not necessarily. The only condition is that in the given form of organisation, the delegating party is actually allowed to delegate to the executor. If we are in a hierarchical organisation, the line manager needs to rely on the availability of the resources attributed to them to fulfill their task. Therefore, other people can only delegate to a person if that person’s line manager has agreed to it. If the organisation has a centre, logic demands that most delegation flows from the centre to the margin, where the execution happens. 

In this sense, delegation can happen between colleagues. In hierarchical organisations, superiors can define the space in which colleagues delegate to each other along a process. The obligation towards the superior is to act according to the defined space, while the obligation towards the colleague concerns the individual task. The most obvious examples are projects. Someone is assigned by their hierarchy to work for a project up to 30% of their time. What happens in these 30% is decided within the project management structure. A brief, or a specification passed from one colleague to the other, may be the same. The famous “internal customer” is in the better case the source of a delegation. In the worse case, it is the source of an instruction. As Dave Snowden pointed out regarding spec sheets, the lists of precise user requirements for software developments: “You make the user sign a document they don’t understand, so you can hold them responsible later if they don’t like the result.” If, instead of the spec sheet, we put delegation at the bottom of the relationship, the user enters in a dialogue about their needs in order to get a software that actually fulfills the task – the core of Agile methods. 

The concept of delegation between colleagues is the cornerstone of highly de-centralised, and dynamic organisations. A famous example is The Morning Star Company, the largest processor of tomato in the world. The company is totally self-organised, without hierarchical relations. Their network organisation is formalised in one-to-one relationships of delegation, called the Colleague Letter of Understanding, or CLOU. In essence, the CLOU includes five elements: The personal commercial mission, which defines the person’s contribution. The key activities necessary to fulfill the mission. The key measures by which the performance can be gauged. The time made available for these tasks. And the colleagues to whom the contribution is delivered, and who in consequence sign off on the CLOU from the very start.

What you get is a networked organisation. As such, the CLOU increase the capability to adapt to changing conditions, to reach out and use relevant knowledge to solve local problems. In 1990, Morning Star was able to oversee the construction of a new, 27 Million Dollar factory with a completely self-organised team of 24 people in a few months only.

In 2007, Morning Star made a big improvement when they introduced the digital CLOU, which made short term adaptions much easier. Putting the threshold as low as possible, you make sure that co-ordination actually happens with the help of the instrument, and not outside it. 

Autonomy and alignment

To summarise: Delegation is a commitment between two agents to deliver a contribution, handing responsibility for the result, and the authority to take the necessary decisions, to the executing party. As such it is a key instrument to create de-centralised autonomy. It enables an organisation to deal with exeptions and unpredictable situations. 

It may surprise to see that one of the institutions who have taken this lesson to heart is the military. A widely regarded stepping stone was the publication of the Study Power to the Edge within the Information Age Transformation Series of the Command and Control Research Program of the US Department of Defense. One of the most striking appliers of this approach was General Stanley McChrystal as head of the US Joint Forces against Al-Kaida in Iraq. He led a task force which included soldiers from the Army, Navy and Air Force, and agents of CIA, FBI, NSA and other agencies, against an extremely versatile, connected and – in terms of methods – unorthodox enemy. In such a highly complex condition, hierarchical decisions and procedures were way too slow. One of the key principles was to radically de-centralise – i.e. delegate – decisions. As soon as they did that, they realised that many of these decisions were coupled to other units. In the old system, the central command structure allowed for a co-ordination of orders and therefore actions across several units. This could happen faster and better if the de-centralised units could co-ordinate directly with the relevant counterpart. For example, someone is observing an enemy meeting point that has been raided the night before. The observer sees a car arrive, people get out, stop cold, get back in the car and speed away. The U.S. observer, 23-year old with little mission experience, contacts other operators to mobilise helicopters to intercept the car and arrest the suspects. When later the parties in the car separate, and take two different cars, the leader of the assault team orders another helicopter to follow the second car. They stop the first car, find out the people are of little importance, keep them arrested, redeploy the helicopters to arrest the person in the second car – who happens to be a senior Al Qaida member. All this within 46 minutes, and no senior officers involved. 

It is fascinating to follow McChrystal’s conclusions for senior leadership from this sort of experience. First, of course, it is a proof that the work “on the system”, to de-centralise decision-making, is adequate for the highly complex conditions, and pays off. Second, however, and in contradiction to many manager’s ignorance or fears with de-centralised organisations, there are some key tasks “in the system” which remain with central leadership. One is to facilitate strong lateral connectivity: the direct connection between units, and the creation of trust. It would not be realistic to simply expect people on the ground to find out who could be the relevant counterparts, reach out to them and form relationships which work under pressure. Therefore, it is the job of senior leadership to create occasions for this kind of connections to take place easily, and improve from there. The second task is to increase the systemic consciousness of the whole group, “the understanding of the nature of the war we are fighting”. One of the means to achieve this task was a daily, one-hour phone conference with up to 7000 participants across the globe – breaking the culture of secrecy, of passing on information on a “need to know” basis, deeply rooted in the military and secret services. What was going on during this call were not generic, top-level commands or definitions of strategy, but mostly examples and points of view from the ground. McChrystal’s job was to comment, facilitate, but most of all choose the right kind of stories to be shared. 

While this glimpse into military command draws a clear profile of the leadership tasks involved, the nature of the mission may make it a bit difficult for many of us to draw actionable conclusions for our business, or administrative contexts. This is why in my next contribution, How to Achieve Aligned Autonomy, I want to show practical methods, and business examples, of how to increase aligned autonomy in any organisation. 

February 3, 2019

Instruction or Education? Complexity, contextuality, and the nature of help

As practitioners we are here to produce results in a specific context. How do we get help?

One way to look at it is that we cluster similar contexts into groups where the same approach works. Dave Snowden’s Cynefin Framework itself is such a kind of clustering on the highest level of abstraction: contexts can be obvious, complicated, complex, chaotic, or of unknown nature. Like in a fractal system, within each of these domains we may form sub-clusters to further specify our understanding and approach. But the way in which we form these sub-clusters is different for each of the main domains. In an obvious context the expectation would be that it is relatively easy to form these groups of similars, and that the shared approach for this group includes every step of the solution, or almost all of it. We can make the recipe so detailed that it is impossible not to succeed. Let us make an example “for dummies”, and assume we need sugar:

1.     We put it on the shopping list. 

2.     We go to the supermarket in due time, for example weekly. 

3.     We take the list with us when we go to the supermarket.

4.     We check the list in the supermarket for each item. 

5.     For sugar, we go to row 23 at the bottom, and take the same label that we’ve had at home (if we cannot remember the label, we have to be more precise on the shopping list). 

6.     We take it through check-out, and take it home.

As long as the supermarket has sugar, and we have the money and just enough brains to follow these instructions, nothing can go wrong. 

Now if we take this from sugar to salt, the only thing that changes is the row in the supermarket. And so with toothpaste, clingfoil, mineral water etc. You have another procedure for re-fuelling your car, but it is the same as the one for buying windshield washer fluid. The procedure for buying newspapers applies to candy and cigarettes. The one for post-its, to glue, paper and envelopes. We can have categories with a relatively large number of individual cases each, write a detailed procedure that covers all steps of the solution process, and this one approach covers all the cases in the category.  

In the complicated domain, this changes a bit. On the one hand, the groups of problems that share a solution become smaller. On the other hand, the help we can get from these commonalities only covers part of the individual solution. Think of the purchasing process for IT. It may include a form to apply for the budget. A minimal number of offers to be considered. Who sits at the table for the decision. But it does not describe your decision criteria in detail, how to research for suitable suppliers, or all the requirements that you put into your request for a proposal. The same applies to a software development framework such as Scrum. It defines some of the roles, artifacts and rules. But in order to develop our software, we have to take many more decisions about the way we collaborate, which depend on our context. Some of them are improvised on the spot: How we make sure Andrew doesn’t interrupt Mike all the time. In other cases, we build an individual, contextual solution out of modules. If we use story points for estimation, there is a recipe how to do it. If we use a burndown chart, there is a way to do that. 

When we move to the complex domain, there is a fundamental shift towards the context-specific. Since each situation is path-dependent, so are the conditions that make some solutions work better than others. In addition, most of these conditions are invisible, and therefore a good solution emerges from the interaction of our initial approach with the environment. Still, we are not totally blind when we approach a new situation. Not every case is novel, and not every chance of success is unknown. But the way these generalisations work is fundamentally different. Instead of categories, clusters or modules of procedures, we are talking about principles. In my line of work, such a principle may be that consulting is mostly a people’s business. It does not tell me what to do, but it sounds like a good idea if I pay attention to building a relation with the customer, and not just to selling a solution to their problem. This conclusion is still a principle. Step by step, my reasoning becomes more concise, more actionable in the specific context, until in the end it makes me decide what sentence to say next in our conversation. Most of this way from principle to action has been trodden by me alone, and could not be replaced by anything generic.

This has an impact on how to get better, and how to get help in the complex domain. There is a fundamental difference between science and the humanities, as it has been laid out by the Neo-kantian school of thought in the theory of science from Max Weber to the present day. Science aims at explaining the type, what you can repeat in an experiment ceteris paribus, under the same conditions. In comparison, the humanities are interested in explaining the individual. If behavioural biologists observe a sealion, they are only interested in what this particular sealion can tell us about sealions in general. If an historian looks at the role Nelson Mandela played in the end of Apartheid, or a literary scholar looks at the influence of Marcel Proust on the modern novel, they are interested in the individual case. The way practicioners learn from these scholars is therefore different, too. Employees in a zoo may read the scientific study of aggression amongst sealions, because they have a particularly nasty dynamic going on in their pool. By replicating what dampens aggression amongst sealions in general, they may expect to dampen the behaviour of their particular bulls. If you apply this principle to Nelson Mandela, the only people reading his biography would be those planning to peacefully overthrow a racist government in their own post-colonial state. Since this would not explain the sales figures, there must be other reasons to buy such a book. Many readers may be looking to be impressed by great deeds, to make sense of what happened, or understand more about humanity. The process that usually takes place here is what Hans Robert Jauss called an oscillation between alterity and modernity: you look at Apartheid, and say. “How very different to my world!” You look at the power relations of leading Apartheid politicians, and conclude: “No, how very similar!” You dig some deeper into the values and beliefs, or their readiness for violence, and go: “No, how very different!” And so forth. And in the process of ongoing comparison, you see your perspectives on your own context challenged, and you learn – not in the form of progress towards the one universal truth, but as an enrichment of overlaying possible perspectives on your situation, and of the assumptions and argumentations that distinguish each one of them. In this process, you may also come to more actionable conclusions. Maybe you get inspired to stand up to a bully at work, or to seek reconciliation with somebody. The difference to learning from science lies in the verbs – replicate versus be inspired. In the second case there is much more building of the individual, contextual solution involved.

And this brings us back to complexity. In the complex domain, the conclusions of scientific research can provide us with understanding of some of the domain’s most generic principles, especially if they are counter-intuitive. For example, our understanding about tipping-points, positive and negative feedback loops, attractor basins, or trophic cascades, can make us better comprehend how an apparently stable and slow-moving system such as public opinion can suddenly shift surprisingly fast. For the large remaining part where we are forced to build our context-specific solutions, the way we can learn from somewhere else is not by replication, but by inspiration. Yes, the most effective way is to probe the system and learn from how it reacts now. But experience, the famous educated guess, brings us to better assumptions on which to choose our probes. And this education, rather than instruction, is what we aim for. 

February 3, 2019

Managing for the Exceptional: the Ritz-Carlton Story

The problem with the exceptional is it doesn’t happen the same way twice. If, as a manager, you see your organisation like a huge prosthesis, which allows you to directly steer a thousand limbs, you won’t get far with the extraordinary. In the luxury hospitality business, the base-line that every competitor meets is to deliver excellence through expensive and thought-through infrastructure, or procedures and standards devised to provide perfection with a majority of relatively low-skilled staff. The way to outperform, to shine in comparison to your competitors is not with your standard, but with the exception, the extraordinary act in a specific context. Again, if you look at the entirety of these exceptions, you face a complex phenomenon. As we will see, the main trait with which the Ritz-Carlton achieves its results is culture, and the way to build a culture of excellence is to nudge for its emergence. Let us go through the elements one by one. 


Many companies try to influence culture by formulating some kind of document – values, a vision, a motto. The reason why many of these efforts are essentially naïve is not that these documents exist. It is that not much more exists around them. In the Ritz-Carlton, the respective document is called the Gold Standards. It includes several elements, most notably a motto, and service values. In the form of a small leaflet, employees are required to carry it around at all times. The company’s motto, as stated in the document, claims that “We are Ladies and Gentlemen serving Ladies and Gentlemen.” What is the nudging intent of this sentence? Employees are referred to as ladies and gentlemen. They are seen as people who can meet a guest with a certain level of self-esteem which is the condition on which service and respect can be expressed beyond the ordinary. 

Management actions to show the anchor is taken seriously

If you tell employees that there is a sentence that describes how much the company values them, and ask them to carry this sentence with them on a leaflet, you would get cynical reactions in most organisations. Not so at the Ritz-Carlton. It is the sum of elements that create a context in which this potentially empty claim is filled with life. The most obvious is to expect a discipline of wording. On all occasions, managers at all levels speak of their staff as ladies and gentlemen. 

Taken for itself, this approach would still be shallow. It is obvious that managers use the words because there is a rule that tells them to do so – so it is up to each manager to fill this wording with authenticity. In another instance, the Ritz-Carlton have actually learnt the lesson that mandatory wording, if exaggerated, can ruin the experience. For some time, employees were asked to say “It is my pleasure to…” on every possible occasion. The effect was a slight too monotonous, and sometimes out of place, so that guests and travel journalists started making fun of “my pleasure”. Since then, the Ritz turned to working more on its staff’s posture, instead of the wording. But the “ladies and gentlemen” remained.

To reinforce the authenticity of the “ladies and gentlemen”, you can give people a practical experience how seriously the company takes its claim. In a so far unconfirmed story, I’ve heard that new hires are asked to stay several days as guests in the hotel. They work their shift in the kitchen, and then they have dinner in the hotel restaurant, pass some time in the spa, and sleep in a room that usually sells at $ 400.- a night. In another case, when the chain acquired an existing hotel in Shanghai, and renovated the building, the first thing they renovated was the staff entrance. The wording of “ladies and gentlemen” becomes tangible through experiences like these. 

Vectors, and authorisation to act

At this point it is still only a signal – not yet embedded in the daily experience. But it explains something about the functioning of the next building block. The Ritz-Carlton is most famous for their $ 2000.- rule. Every employee – or lady or gentleman – is empowered to spend up to $ 2000.- to solve a customer issue, or create a Wow-moment. Many managers who hear this, are shocked at the thought of it in their own company. And they are incredulous when they hear that yes, no member of staff is ever told off for the way they spend money under this rule. And no, there is no inapropriate, or excessive spending of this allowance. To be clear, the sum needs to be put in perspective to the average $ 250 000.- lifetime customer value of a Ritz-Carlton guest. But again, the relevance and effectiveness of this rule only appears in the context of other nudges. The anchor for this behaviour, in the Service Values of the Gold Standards read: “I build strong relationships and create Ritz-Carlton guests for life.” The long term purpose of the employee’s behaviour. “I am empowered to create unique, memorable and personal experiences for our guests.” The focus of the behaviour is where standards and procedures cannot reach. “I own and immediately resolve guest problems.” This last one is particularly telling. Someone phrased it this way: “Who sees a problem, owns the problem.” No matter what your job responsibility, if you lead a guest to their room, and on the way see that someone has put their breakfast tray out their door, it is your obligation to make sure it disappears. If you finish your job in a minute, you come back and take the tray down to the kitchen. If your job takes longer, you call someone from room service. As one Manager of Guest Services phrased it: “There is nothing of which I can say: this is not my job. Everything is my job.” 

The Service Values act as vectors. As part of the anchor, they are just claims on paper. But tangible actions, and the authorisation to act, make them impactful. What is the outcome? Here are some examples. 

A facility manager was changing a light bulb when a guest asked him how long it took to get to the airport, because they had to go there fast. “No problem, Sir”, he replied, “I’ll take you to the airport.”

A family with two boys stayed in the Ritz Carlton Toronto. The boys played floor hockey in the hallway. Another guest complained. The member of staff went to the room and explained to the parents that their kids could not play floor hockey on the hallway. The parents were very understanding, and asked the boys to stop. And this is how far any hotel would have gone. But the employee said that one of their ballrooms was not in use at the moment. He invited the boys to come play floor hockey in the empty ballroom. He organised some other members to join, and they challenged the family to a match. 

In Berlin, a family was on their trip home, when it turned out the little girl had forgotten her teddy bear in the room. They called the hotel, and of course the hotel had found the teddy and offered to send it home. Again, this is how far any hotel would go. But they took the teddy to the various departments, took pictures of the teddy with members of staff, and wrote a little story about all the adventures the teddy had had during his time at the hotel, and that it hadn’t been lonely at all.  

Regular practice to focus attention towards the desired way to act

How do we learn about such Wow-stories? They are shared. Of course, when they appear on videos made by hospitality trainers, or on the Ritz’s public learning institution’s blog, or are mentioned by the CEO in an interview, they work as multipliers of a marketing effect. But the more important role of sharing is internal. And again, there is a structural element which ensures that. At the Ritz-Carlton, the practice is called the Lineup. It takes place on department-level at the beginning of each shift. For fifteen minutes, the team members discuss one of the Service Value statements from the Gold Standards, share Wow-stories, and use some time for other relevant information, birthdays or anniversaries. The value statement of the day is chosen centrally for the entire Corporation, and the central communication department prepares a worldwide collection of Wow stories for the teams to choose from. Both the statement discussion and the Wow-story part are facilitated by a team member. 

Motto, Service Values, $ 2000.- rule, and Lineup are all rigid, centrally imposed elements. The Motto puts the person in a certain place – I am a lady or gentleman. The Service Values offer orientation – what would be a good thing to do. The $ 2000.- rule opens opportunities – what am I allowed to do. And the Lineup guides attention to how desired behaviour in specific contexts can look like – how could I do it. Together, they create the structural conditions for nudges which happen in situations that nobody has foreseen, and which turn them into a unique, but entirely appropriate and desired experience. 

p.s. And if any of my Agilist friends is wondering what this article is doing in their feed – try finding the values of the Agile Manifesto in the story.

February 3, 2019

Manage for the Next Possible Step

Many have pointed out that the word to manage comes from the Italian maneggiare, and originally refers to the training of horses and riders in the arena, or manege. The classical metaphorical use is this: the rider is the king, or governor, the horse is the people, and both learn. On a slightly different note, the image reminds me of a situation when we brought our daughter to a public horse training session at the Circus. The trainer had a microphone, and commented on his moves. The horses were in a queue, and he invited one after the other to do a trick, and then join the end of the queue again. Not every horse did the same trick, though. For each horse, the trainer chose the one level which was the next possible step, the mental barrier the horse had to work at, and overcome. Whether the horse succeeded or not, it inevitably joined the back of the queue again. For each horse, the trainer chose a rhythm between attempts and little corrections. If a horse failed, it got an easier trick next turn round, to regain self-esteem. Keeping an eye on that balance of self-esteem as a condition for competence, was key to the trainer’s success. What was important – and here’s the nudging in the story – is that the horses were able so see each other. On the one hand, they were stimulated by watching the others. There was a certain competitiveness, although it was very important to accept that the degree of competitiveness was different with each horse, and there was no point in trying to coerce an uninterested horse into becoming more competitive, e.g. by some kind of carrot-and-stick method. On the other hand, by witnessing their fellow’s exploits, the horses learnt from each other.

What I like about this analogy is the notion of the next possible step, and the intricate conditioning of the individual succession of tricks, and the group, that created a progress of performance. These lessons remain useful even if you take away the superiority of the trainer in terms of knowledge and purpose orientation, and replace the figure by an agent in the system, who can sometimes modulate the architecture, and otherwise influence with intent. 

This insight has become one of the most important principles of my activity as a coach and consultant. Think of organisational culture, team dynamics, but also individual things like how much push back a person is ready to accept, or how much parental guidance a teenager takes in. It is one thing to identify what would be good for the system in focus. But the more important question is: What is the next possible step the system is actually capable of going? And in many cases, the second point covers less than 5% of the first. But what is the point of going beyond and failing? 

The key is to be as close to the limit as possible. And since we cannot know where this limit is unless we hit it, we have to experiment. Any experiment that falls within the limit is not an experiment, but a measure that works. But from time to time, an experiment will overstep the limit, and fail. In the mind of the horse, that creates frustration, and the trainer has to go two levels lower next time. In the mind of people, the memory can stay on longer: there is a stronger path dependency in the system. If our experiments fail too often, whatever we do next, people will say: “Oh no, not another experiment!” Think of a politician in an executive role. She may try many important, but unpopular decisions such as reforming the health or pensions system. There will always be resistance. Sometimes, she has to step back, and withdraw her proposal – a sign of where the limit is. The more withdrawals, the more the opposition understands that resistance pays off, and increase resistance each time. If the politician withdraws too many times, or oversteps the line too far, it will be the reason she won’t get re-elected, and the end of her career. This is the essence of wicked problems: we only have one shot – if we miss, our problem changes, and usually becomes more difficult.

Older Posts