At any managerial level, the point is to make correct assessments in adequate time with partial information. The ID-DFS mode I mentioned earlier yes is an effective protocol but it imposes a very high cognitive load. And added to the inherent computational complexity is the noise, i.e., the non-determinism of the system under analysis, the unnecessary statistical significance of the detections we use as a basis for decision-making, the lack of heuristics in the proximity of the exploration frontier, the impossibility for this reason to do pattern recognition and reduce to a better known problem. And added to these two levels of difficulty is a third, which is the non-objectivity and impartiality of the decision maker.
Often roadblocks before being corporate are personal. Companies often plateau for reasons other than organizational-management. Sometimes simply decision makers work with what one U.S. author has called ghosts in the machine, residual functioning systems with respect to past situational needs, which no one has ever told that their task is finished (ref Jarry Colonna). Imagine having someone constantly suggesting tactics to you that are objectively dysfunctional: I doubt you would want to keep them close. Ghosts must be hunted for, and they are detected precisely when recurrent or otherwise significant behavior cannot be explained in tactical terms.
Modern IDSs function in the same way. They observe a network of endpoints and discriminate between physiological and pathological processes by detecting the “meaningfulness” of behavioral patterns. And from a statistically significant detection begins the tracking down of the dysfunctional process, backdoors, C&C mount of APT. I step out of the metaphor. If a thought process causes me to enact dysfunctional behaviors, that process needs to be investigated and reevaluated. I am not saying that all blocks to a company’s growth are of this type, I am saying that many are, and to work on companies you work on people. And increasing one’s toolbox is a competitive advantage.
I start with an example and then generalize. Usually one of the elements that hinders process optimization is the Dunning-Kruger effect. There are two facets of this effect, two distinct local manifestations. It is that mechanism whereby 1, those who know too little about a topic do not know enough to realize that they know too little about it, and 2, those who know a lot have the information so well structured and flowing that it is easy for them to interact with it and so they devalue their abilities about it. Evading the DK takes time, and communication-type bypasses to prevent the situation from stiffening. This I have described is a typical example of cognitive bias, that is, an automatic but dysfunctional thought process.
More generally, an imperfect decision-making system has internal biases and resistances, and elicitative methods that systematically attempt to circumvent them are needed. I give an example. It has happened to me as an administrator to unconsciously ignore possible ways to solve a problem because I as a person would not want to pursue them. Sometimes solutions exist but are subject to resistance. The elicitative technique in this case was to stop thinking “what can I do – personally – to solve my problem” and instead start thinking “what would I have X do to solve my problem.” Thus I also make detectable those potentially effective hypotheses that I personally for whatever reasons was avoiding. This is just an example, but it gives the idea. Working with companies and people means working with powerful but imperfect systems.
Yes, these things need to be thought about to talk about optimization. I do not launch a shuttle without fine-tuning the engine because I would not be successful. Similarly, I will not be able to launch a company on a market trajectory if I do not take care of people before processes. As usual without an outside point of view, and unaccustomed by and to the situation, you don’t even start. I don’t fix an engine from the inside, I periodically take it apart I observe it I test it and improve it.