In the maze of complexity
Two weeks ago, the landline telecom-network in Switzerland collapsed for about eight hours. Emergency numbers were also affected. The company Swisscom thus made it back into the headlines. Similar system failures, in which the national telecom service provider did not cut the best figure, were not so long ago. Since the emergency numbers were down on a night with severe storms, the media coverage was correspondingly high. Thus, in the aftermath of the events, I make an interesting observation in an interview with the CEO of Swisscom. A journalist asks him whether, in view of the seriousness of the incident and the accumulation of similar incidents, it might not be time to resign.
Resigning would be equivalent to a severe punishment for the CEO. He would take responsibility for what happened. But what would be achieved? The public would have a scapegoat, and it would go back to the 'courant normale' with the illusion that the problem had been solved. Case closed. Interestingly, it does so even though, apart from confirming a questionable sense of justice, it knows full well that the problem has not been solved.
The journalist's question with a prompting character is understandable. It comes from the widespread and well-entrenched error-blame paradigm. Those who make mistakes must take responsibility for them, bear the consequences, accept the blame. This is done in the misguided belief that justice will be done and that the obviously incompetent person in charge will be replaced by someone who can do it better.
Let's get to the bottom of the matter and examine the error-blame paradigm for its suitability.
What happened?
Swisscom was performing maintenance work on a network element of a telephone platform for business customers. A software update caused a malfunction that triggered a domino effect. As a result, large parts of the network were affected by the malfunction. The reason why it took so long to fix the glitch was that the supplier of the network component had to be called in for this purpose. For a better understanding, it is worth mentioning that there is no system leadership for emergency calls in Switzerland. The CEO explains in the interview that numerous partners are involved in the emergency call system. It is like a machine with 1,000 cogs: Swisscom controls perhaps 700 of them, and the company has only limited or no influence on the remaining 300. But all these cogs have to mesh. Otherwise, the emergency call would no longer work.
Without knowing the details of what actually happened, we can imagine the chain of circumstances. It is extremely rare that maintenance staff go to work with the intention of producing a breakdown. It is equally unlikely that update software was deliberately programmed incorrectly. We can safely assume that the programmers were not aware of the undesired domino effect that their software would trigger. The fact that in this case the quality check of the software was not up to the requirements may be of systemic nature or may be due to the circumstances in the wider context in which this check was performed. The domino effect triggered by the software update could most likely not have been anticipated by the maintenance crew on site. It was probably more of an unknown phenomenon. To what extent the aforementioned 300 Non-Swisscom cogwheels in the emergency call system had an influence on what happened is beyond our knowledge. The only certainty is that they played a systemic contributing role.
That's how complexity comes along.
What may we conclude from this?
In this case, we are dealing with a system that is not under the control of a single organization. The system boundaries run far outside the Swisscom company. They extend into politics and the federal structure of Switzerland. Replacing the CEO would bring a person to the top of the company who only potentially knows or can do more, but who would be confronted with the same system. A system that cannot be changed with the push of a button within the company. Whether such a person could be found at all, of course, remains to be seen.
The characteristic of complex systems, apart from the thousands of existing mutual relationships, is also the fact that they can no longer be controlled, but only influenced. And that they are able to produce undesired events, although all have worked correctly according to instructions. To start here with the medieval punishment club is like trying to start an ice hockey match with jogging shoes. In a complex environment, the sanction has degenerated into a blunt tool. As we know, it even hinders the required learning process.
It makes much more sense for us to realize that, not least thanks to digitalization, we have built systems that are way out of step with our understanding of the law and our idea of control and accountability. They relativize the importance of individual people - including bosses. We can change them all at any level of the hierarchy and will find that this will not solve the problems. The complex systems challenge us with our desire for control and ask us to find a different way of dealing with them.
On closer inspection, they ask only one thing of us: that we learn to understand them. We have developed them in many iterations, often over a long period of time, and there has never been anyone around to take care of keeping up with the user manual. So we have no choice but to take the painstaking reverse route and learn by observation how they really work. To do this, we need information and data, tips and reports from employees and supervisors that reveal systemic weaknesses. We need a culture in which reporting one's own mistakes is valued. Where people who report mistakes do not have to fear disadvantages or even sanctions. Where punishment is only in the management's toolbox for the very gross and deliberate.
Letting go
We must not resent the journalist's question about resignation. It comes from a long-gone time of simple systems. A time that has left a lasting mark on us and our understanding of the law and our legal system. So it is not surprising that in the whole interview the reporter never asked the CEO of Swisscom what should be done to prevent the company from such undesirable events. Let's hope the next generation of journalists can do it.
The widespread application of artificial intelligence is just around the corner. The accompanying massive increase in complexity demands a social and cultural development push from us in dealing with complex systems. This starts with the first step: Let's throw the error-blame paradigm overboard. It has done its job - may it rest in peace.
