A Beautiful Mind
Bureaucracy ignores human nature because it's not tractable.
Bureaucracy fails because it abstracts away the messiness of human nature. To make its model tractable, bureaucracy has to make itself practically useless for its stated objective of efficiency.
In the end, bureaucracy just makes everything weird.
I’m reading The Geek Way by Andrew McAfee. It’s great. I didn’t expect to see the excoriation of bureaucracy. That was an added bonus.
It’s a book about culture and management in tech companies, how they differ and why they may be better than traditional corporate environments for making interesting things happen. I agree they’re different. They’re certainly efficient. I disagree with some of the decisions they make, but that’s just me. At least by reputation, they’re very good at executing, regardless of what we may think of their choices.
He starts out by looking at research done by David Ford and John Sterman into the question, why are so many large projects delivered behind schedule? You can see their paper, “Overcoming the 90% Syndrome: Iteration Management in Concurrent Development Projects,” in Concurrent Engineering: Research and Applications, Volume 11, Number 3, September 2003, p. 177-186.
“The syndrome describes a project that reaches about 90% completion on schedule but then stalls, finally finishing after about twice the originally projected duration.”
They built a “dynamic project simulation model” with several parallel constraints:
1. Some activities require a minimum amount of time no matter how many resources you devote to them.
2. Typically, development consists of activities executed in sequence.
3. Phases of development depend on information within and across different phases.
4. Work is often released in discrete packets, not as it is completed, affecting the flow of information across phases.
5. Sometimes released work needs to be redone and this requires coordination between those who did the work in question and those who discovered the need to redo the work.
All of these can gum up the works. Testing reveals problems late in the game, requiring rework. As they retrace their steps, progress stalls. This is called an iteration cycle. Some of it is to be expected. Nobody’s perfect. It’s the unplanned iteration that gets you. For example, if you start a subsequent phase before you’ve completed a prior one and obtained all the relevant information can lead to problems that you must fix subsequently.
The need to coordinate the efforts of different teams and to ensure the flow of the right information to the people who require it in a timely manner can lead to all kind of process design. It’s a hop, skip, and a jump from process engineering to bureaucracy. This is a systems dynamics problem.
The solution (according to Perplexity) is: “surface errors early, improve the quality and stability of information passed between phases, reduce unnecessary coupling among activities, and build fast mechanisms for detection, coordination, and engineering changes. The key managerial mistake is treating apparent completion as real completion when unknown rework remains.”
But here’s the really interesting part.
In a subsequent paper, “The Liar’s Club: Concealing Rework in Concurrent Development,” Concurrent Engineering: Research and Applications, Volume 11, Number 3, September 2003, p. 211-219, they introduce the actual behavior underlying all this process engineering work.
“Concealment is often standard practice. At a major defense contractor, weekly meetings of project team leaders were known as ‘‘the liars’ club’’ because everyone withheld knowledge that their subsystem was behind schedule. Members of the liar’s club hoped someone else would be forced to admit problems first, forcing the schedule to slip and letting them escape responsibility for their own tardiness. Everyone in the liar’s club knew that everyone was concealing rework requirements and everyone knew that those best able to hide their problems could escape responsibility for the project failing to meet its targets.”
This is a glorious example of the assumptions bureaucracy makes that are inconsistent with human nature. I suspect this behavior persists across cultural lines. Indian managers, British managers, American managers, European managers. They’ll all behave the same way.
McAfee introduces the game theory here. (He calls the game the “Are You On Time?” game.)
“Let’s focus exclusively on these late players in the AYOT? Game. Each week they have a choice between two strategies: be deceptive (and say “Yes, we’re on time”) or be honest (and say “No, we’re behind”). What are the payoffs associated with each of these strategies? Well, if they’re honest, the payoff is that their reputation suffers and they get chewed out by the boss, blamed for throwing the whole effort off schedule, subjected to increased scrutiny and work as well as decreased career prospects. If that sounds too harsh or cynical to you, my guess is that you haven’t worked on a lot of big projects at industrial-era companies.
“What’s the payoff to being deceptive in the AYOT? Game? It depends on whether your overall project manager knows your team’s running behind. If so, the payoff is all the bad things listed above plus gaining the reputation of being a liar. That’s bad. But remember, observability is low in this game. It’s unlikely that the manager knows who’s late (if they knew, why would they bother to ask?). So the most common payoff for players who are lat and deceptive is the same as the payoff for players who are on time and honest: escaping further scrutiny and going back to work with an intact reputation.”
This reminds of the old joke, “That’s all very well in practice, but how does it line up in theory?”
That joke explains everything you need to know about bureaucracy.