← Portfolio · ← Collaboration 2.0
1.2 · Clues — where the regress terminates
Why the graph in 1.1 has no single root, and what the two nodes outside every cluster turn out to be.
1.2.1 · Clues
At first glance, the graph we arrived at in the previous section does not point us to a single root that started the process. The question we set out to answer, at the first step, was to find the underlying element that causes so many surface-level problems.
The first question that comes to mind is how to get to a tree — one that points at the original blocks, the priors that give birth to what we saw: a bunch of entangled, closed-loop symptoms. Looking carefully at the graph, I came to the idea that a few of the nodes are candidate answers. Those nodes differ from the others, as explained below.
This is also where the clustering in 1.1.5 is repaid. The historical question is hard to ask about eighteen symptoms — you get eighteen answers, most of them local accidents. It is tractable to ask about three clusters and four routes, because there are only seven objects to explain and each is large enough to have a history worth telling. A careful taxonomy is what makes the roots findable at all; a sloppy one leaves you with the original graph and no way in. If the crossings had been thirty instead of four, there would have been nothing to ask this question about.
These are nodes we cannot attribute a causal relation to from any higher-level symptom. They seem to be this is what it is. If you want to reason further and find what is behind them, you get a continuing chain — not a loop any more — which eventually arrives at a node that is not a symptom at all, but an explanation about the company itself. Let us look at a number of such nodes, which are also distinguished in the second graph by not being placed inside any cluster.
In what follows I list a set of facts that chain one after another. The nodes that were not placed in any cluster can be found in this list.
Herein, the Fact 1 mentioned below, is nothing more than a simple sentence describing the situation of big proud company that so far had been very successful in its previous missions and made tremendous revenue. yet, in the sections coming after Fact 1, you will see how this situation with the company makes the corner stones for upcoming issues.
Fact 1 — The division we talked about so far, was a branch of a bigger company that decided to invest in digital adaptation. The previous missions of the bigger mother company had been to work on engineering problems based on transferred technology. You can think of it as receiving the engineering handbooks and the manuals, together with a clear, well-defined engineering problem.
Culture/norms resulted from Fact 1
- Fact 2 — Fact 1 made the company’s original processes, and its codes and norms of behavior and collaboration, demand a great deal of regulation.
- Fact 3 — This one is the natural result of Fact 2. These norms evolved as best practices over more than a decade of in-depth work on the previous missions. So for people who have worked here for decades, the norms sit in their minds like natural beliefs. It is hard for them to imagine that the disciplines coded so deeply into them are now a disruption in the new place.
- Fact 4 — The hierarchical structure, which you find in so many places in the graph that I got tired of repeating it, is itself a product of Fact 3.
Market and strategy challenges resulted from Fact 1
- Fact 5 — The company has extremely weak experience of thinking about strategy and market analysis or adaptation. Because of Fact 1, it has the least possible experience of entering a new and unknown market. Until now, the problem to solve and the criterion for judging whether the mission had been accomplished were always clear.
Hierarchy and norms being emerged from Fact 1
- Fact 6 — This one is very hard to go into. as mentioned in the Fact 1, The truth is that the division was a branch of a bigger company that decided to invest in digital adaptation. So whoever was chosen to lead the division was probably already in a relatively strong position in the parent company. As a result, the relationship between that leader and the whole business of the division is not like that of the CEO of a small startup, or a founder who has risked their own money on a new business.
- Fact 7 — Fact 6 causes the interests of the highest authorities in the company not to be necessarily aligned with what makes the division itself grow.
- Fact 8 — Based on Fact6, the highest authorities are mostly motivated to satisfy the level above them, which produces disruptive phenomena like echo chambers and political environments.
Now, reviewing the two lists above, we realize that two facts are causing the rest: Fact 6 and Fact 1. Let us name them: - Imbalance of the responsibility and authority of the highest authorities - Incompatibility of the digital business and industry with previous missions of the mother company
Now our “Aristotle regress of ends” looks like this (Figure 4):
Figure 4 — the two enabling conditions. The same three clusters as Figure 3, now shown against what let a circuit of this kind form. The figure is acyclic by construction: these are enabling conditions, not sustaining causes, and 1.2.2 explains why that distinction decides the whole design.
First: authority and responsibility were decoupled at senior levels. Authority sat above the information. Responsibility sat below the authority. Nothing connected the two.
This is the condition that lets a circuit like this form. Where authority and responsibility have come apart, people cannot tell disagreement with a plan apart from disloyalty to a person. And once they cannot tell those apart, they stop moving information upward on their own.
Every route running out of the structure cluster passes through this.
The survey that came later put a number on it. Seven in ten named the complexity of the hierarchical requirements as their main obstacle to getting information out of another team, which is the highest count on any option in the instrument (4.3).
Second: the group’s previous mission had a different nature from digital industries. The parent organisation was built for work in which the customer was known in advance, the requirement arrived from outside, and the product was defined before the engineering started. A division doing digital product development needs a function that discovers what to build. The parent had never needed one, so it had never built one, and it had no way to notice the absence — an organisation cannot miss a capability it has never had a reason to develop. This is why the strategy and market cluster looks less like a set of failures than like a set of vacancies.
Where Collaboration 2.0 sits inside the wider business direction, and what had to be done about this vacancy, is 4.2 · the phased strategy.
1.2.2 · What these two conditions are, and what they are not
They are not root causes you can pull out, and this document should not be read as offering them that way.
Fix the authority imbalance tomorrow and nothing stops. The secrecy it produced now keeps itself going. The people who would have got information moving again have already left. The norms that formed under it are enforced by colleagues now, not by anyone in charge. The second condition cannot be fixed at all, because you cannot give a company a different past.
They are enabling conditions: what had to be true for a circuit like this to start, and to survive without being challenged. Once such a pattern is running, it no longer depends on how it started.
That is not a technicality. It decides everything in the design that follows.
If the two conditions were removable causes, the right move would be to remove them — rebalance the authority, build the market function, and wait for the symptoms to clear. That is the obvious response, and it is what most reform efforts try. Here it fails, for a reason you can see in the figures. The loop no longer needs its origin. Taking away what let a fire start does not put the fire out.
1.2.3 · What would be next (intervention)
So far we have seen that clustering transformed the raw graph into something easier to work with, and easier to explain different phenomena with. What it did not do is satisfy the natural human tendency to find the single source of the problems — in other words, who is to blame? So we carried on trying to terminate the regress, until we ended up with a model that explains how some factors coded deep inside the company itself made trouble on its way to adapting to the digital business.
Of course, nobody is going to be blamed in practice. I believe this is just a natural tendency of the logic in our minds to think about the problem in this way. Later, in the sections on my proposed intervention, I will explain why I believe that this natural inertia — going after a node at the head of a chain of causes — might be misleading when we are thinking about a solution. Because it is very likely that one concludes: if we remove the very root from which the problems started, then we have uprooted every problem it caused.
There are several mistakes in that conclusion. The one that matters most here is that the root of the symptoms is the patient itself, and no intervention worth proposing kills the patient. 2.1.3 takes that argument up in full and turns it into a different question. 2.2 is the solution itself — the one I name Collaboration 2.0, to keep the patient alive.
⟵ Up: Collaboration 2.0 · Previous: 1.1 · Symptoms and diagnosis · Next: 1.3 · Disclaimers →