Changing Recurring Outcomes
Has it ever seemed to you that you’ve been here before? Have you ever gotten that sense of deja vu that the situation around you is completely familiar, even though some of the details are different?
I have a motto:
“When some behavior persists, a feedback cycle exists.”
What does that mean?
When W. Edwards Deming said
“A bad system will beat a good person every time”
he meant that it doesn’t matter if you’ve got good intentions and work hard if the system is designed to produce bad outcomes. What I mean is that when a system repeatedly produces certain outcomes, it’s a signal that the system is designed to do so.
That doesn’t mean that someone intentionally designed the system for bad outcomes. I’m talking about the de facto design—the one we happen to have. In a complex system, there are lots of effects going on that we can’t predict, or even easily observe. And any system made up of people in inherently complex.
Perhaps someone did try to design the system to produce desired outcomes. The mere fact that they called it “designing the system” rather than “growing the system” signals that they were assuming the problem was merely complicated rather than complex. Complex systems have all sorts of non-linearities and unintended side-effects.
Complex systems also have extraordinary staying power if they’ve survived at all. They have feedback loops where some the increase of some observable behavior either inhibit or promote some other behaviors. Each of those behaviors, in turn, inhibits or promotes others, and so on, until we have a behavior that either inhibits or promotes the one we started with. If the number of inhibiting influences in the circle is an even number, then they “cancel out” and the result is that our starting behavior promotes itself. That’s what makes it impossible to change the behavior by “trying harder.”
Instead, we have to change the system.
Changing the system is not an easy thing. Again, you can’t go to the drawing board, draw the system you think you’d like to have, and “make it so.” It’s a complex system, not a complicated mechanistic one.
Nor is it likely feasible to break an existing influence of one behavior on another. The system is maintaining that status quo by its design. Complex systems adapt themselves to produce the results that they are producing. (There are also many other influences going on than the ones we’ve been able to identify.)
The best approach I know is to try to introduce new paths of behavioral influence that counteract some of the existing ones. Looking at the patterns of influence that we can observe and intuit will give us some hints on where we might introduce these. The proof is in the pudding, however. We can’t be sure that we’ll get the outcomes we want. We have to try things and see what results we get.
If the results are going in the direction you want, then perhaps doing more of that will move it further in that direction. If not, then perhaps we should try something else.
In Cynefin terms, we probe the system by trying something, sense how the system responds to that probe, and respond by amplifying this probe and/or adjusting our next probe. It’s not easy, but anything less can only succeed by chance.