Why Big companies fail to flex

Decisions are complex and multi-staged. In the simplest model, there is a stage of information gathering and analysis, a stage where the decision gets made and an execution stage. Put even simpler, there is the stuff that happens before the decision and the stuff that happens after the decision.

The Ballmer peak (and many similar ideas) already explores the idea that there is a relationship between effectiveness and the number of people on a project (or BAC if you're learning about it through xkcd). The relationship looks like this:

Basically, the ability to execute on a particular course of action depends on how many people are involved in the execution. When it comes to a decision, you need to socialize the decision and the same relationship exists. Too few people and you won't make the decision stick. Too many people and you get bogged down in administrivia and communication problems. There is some narrow range of people within which you have reached a required thresh hold of effectiveness:

I think the shape of this curve is a function of the size of the organization that is impacted by the decision (the size of the organization that is expected to execute the course of action being communicated). Larger organizations need more people. In general, the larger the organization, the more people you need involved and the lower the overall effectiveness of everybody involved will be.

When it comes to everything up to the point of decision, there is a similar relationship. I think it looks like this:

With the same sort of threshold. Too many people and you cannot effectively make a decision. This may seems counter-intuitive, but it really is true. Larger and larger teams for information gathering, analysis and recommendation creation get bogged down under their own weight and become debate societies. (Maybe this curve should be another normal shaped curve since too few people is pretty ineffective, but that doesn't matter for what we're doing here).

Different decisions require different decision structures and I'm not going to get into the difference between decisions that require consensus, those that are best with authoritative decision structures, etc. I will say that I think the more complex the decision is, the higher the threshhold of effectiveness and --paradoxically-- the fewer the number of people you want involved in information gathering and analysis stage. I think it looks like this:

Ok, so what?

Well, here's the sticky part. As long as you can overlap these curves and come up with a feasible number of people for your decision team, you can make and socialize a decision:

In this case, the space between the green dot and the left end of the purple range represents the size of the decision group that can actually accomplish something. For this decision, you can actually make a recommendation and see it executed.

But, for some decisions that are sufficiently complex and that need to land in organizations that are sufficiently large, you end up in a paradox. There is no team size that is sufficiently large to socialize the decision but sufficiently small to make it in the first place:

I think this is a big contributing factor in the repeated and predictable failure of big companies. They simply get too big to make decisions and execute against them. This problem may not become emergent right away -- lots of big companies can just keep going the way they have been without much trouble. For decisions that are simple enough, the threshhold is low enough that you can assemble teams that can work (should we just keep on doing what we're already doing?).

The issue comes up as soon as a really big company (or organization or division or whatever) needs to make a radical change.

What are your thoughts?