
A day at the beach
In between babbling about memetics to friends, every now and then one will get a word in edgewise, and ask “OK, Mr. Smartypants. How do we change outcomes?” My answer is always the same. “You have to consciously and deliberately apply Conway’s Law and change the social structure. If you don’t change the social structure, the emergent outcomes that you’re complaining about won’t change.”
They look at me, somewhat knowingly, like they MIGHT have gotten it. And then they walk away, just hoping their organization is going to suddenly morph into something that will produce different outcomes than the current version. Of course, this is insane. What they’re really doing is retreating back into The Matrix. And what people don’t get is that The Matrix has physical laws. You don’t get to break them. Like Grandma always used to say, “wish in one hand, shit in another, and tell me which one fills up first.”
That Grandma.
If you’ve read this far, I’m going to assume that you might be looking for an example (or something) that really conveys to management how this is true, and how they might need to readjust their thinking, as well as their org. chart, if they want a different product design.
Look at a nuclear reactor power plant. It’s a big thing, and at the center is a nuclear reactor. The reactor has support services that connect to it, like cooling fluid, and maintenance departments in charge of things like pumps, environmental controls and whatnot. You better believe that the org. chart looks like that hierarchy.
Last week, I had a visitor from Germany, an organic farming advocate. I was explaining to him that if he wanted small, robotic crawlers and grabbers, as opposed to enormous wheat combines (he got to ride in one) that had to start with a change in social structure of the organization producing them. Not easy to do with no obvious market, and folks making fine money on giant machines, now even driven by satellite.
But let’s take an even simpler design example. An airplane.
Virtually all but a handful of modern aircraft have a tube, called a fuselage, in the middle, upon which are hung wings. Engines then are hung off the wings, and then of course, there are subsystems inside the wings that drive things like flaps, ailerons and control wires.
But getting back to the social configuration, you’d better believe there is a wing group, concerned with the aforementioned components, and there is a primary structures/fuselage group. (Remember this is a simplification.) Within the wing structures group is a sub-group who talks to the fuselage folks regarding attaching the wings to the fuselage. Those folks might even hang out together, because pinning the wings to the fuselage correctly has all sorts of issues, and is obviously critical to entire system integrity. This probably seems obvious (it is). There’s also going to be a vertical stabilizer group (that’s the tail) and a horizontal stabilizer group as well. That’s aircraft 101. The people that interface all that in the design will talk to each other. And then at a large aircraft manufacturer like Boeing, they even have people called “Liaison Engineers” (which they pronounce “Lie-uh-zon” accent on the first syllable) that basically cruise around and make sure the interface people, as well as other folks having integration issues, are taken care of.
I’ve trained my share of Liaison Engineers.
All this seems obvious. But how would such a group ever change an aircraft design to something like a blended wing/body design? See below.

They couldn’t, of course. They are locked inside their social structure, refining the parts of the airplane they are responsible for. This is not entirely a bad thing, in that in the process of refinement increases reliability of the current configuration. Which is a primary reason that airplanes don’t drop out of the sky, and the biggest thing you have to worry about on a flight is whether you’re going to get a package of mixed nuts. Or not.
Any development of such a blended wing-body aircraft would require major redesign, with a major reshuffling and pulling of experts out of all the current groups into an entirely different social structure. Because you’d take some of the old v-Meme NA of the design (e.g. probably attaching engines works much the same way, and trust me, Boeing has specialists for nailing engines to wings) but other things would be entirely different. You’d definitely have to have an entire group for ground crew liaison, and on and on. Those people would then have to talk to the outside vendors providing those sky bridges we’re all accustomed to.
Short version — you’d have to create an entirely new social structure, which then would have to create an entirely new CAD model (think knowledge structure) which would then be instantiated in the final blended-wing-body design.
On top of that, I’m willing to bet that the people in the current organization aren’t used to jumping out of their hierarchy to talk to other technical specialists in the other hierarchies producing the other large parts. Remember — they have those Liaison Engineers for reasons. But you’d need that, and then you’d also have to have some evolved leadership so that if people aren’t “staying in their lane” they know that’s OK, and even encouraged to bring issues that the other parts of the new configuration might need to know about. People that are used to phoning it in because they’re master of their small square of real estate in the old design are going to have to be encouraged to seek out places where the new design requires design synergy.
And without that, you couldn’t make that cool new design at all. Now take that down to manufacturing and you’ll have a markedly different assembly line for such an aircraft. Boeing 737s are built on a continually moving line. If you wanted to build that new plane, you’d have to rethink all of that. And that would require a very different org. chart as well. As well as integrating the manufacturing people far earlier in the process, so you could actually build the thing. It would be revelatory to put the 737 assembly process org. chart up next to a B-2 assembly org chart. While there would be commonalities, I guarantee they’d be significantly different in topology.
If you’ve been reading much of this blog, you now can start seeing how you have different conflicts, as well as synergies in reliability and validity. And that’s going to require different brains, with different abilities to talk to other people with different brains. If you haven’t had an organization that has evolved empathy, that’s not going to exist. And trust me on this one — the thought of jumping out of the org. chart won’t even occur to most people, except in the context of whistle-blowing wrongdoing.
So that’s Conway’s Law in a nutshell. Your org. chart, and how you develop your people, which is largely due to how you set up your communication culture, is destiny.
Because Conway’s Law is The Law.