One Simple Rule To Control A Business: Emergent Behavior
Even a small business is pretty complicated. I have heard some people say that business is so complicated that no one can really understand or control one. While the former might be true, you don’t need to understand everything about a business to manage it. And it can be understood by watching flocks of birds.
Let’s say you are made the CEO of a large company today. Congratulations! But how do you manage such a monstrosity?
Usually there are metrics that the person at the top of the organization is watching. These could be things imposed from the outside, like market capitalization, or self-selected, like profit.
Remarkably, many companies don’t have much more than that — they are missing organization-wide metrics. It should terrify you if you are a manager at such a company. Imagine running your daily processes without having any information on how those processes were running and not finding out until weeks later if what you made today was any good. Most companies couldn’t survive in such an environment.
Our approach is to use a company’s words for where it wants to go and how it creates value for its customers to identify high-level metrics. We then use the concept of hoshin kanri to translate those business success metrics into metrics at every level.
Sometimes I get resistance to this idea. The reasoning goes that there are too many things going on every day for any set of metrics to measure the chaos, much less control it. And in a strict sense, maybe this is true. But thankfully, the universe conspires to make real processes different than theory.
There are three principles that allow us to create metrics that make managing a business possible: emergence, regression to the mean, and the Pareto principle.
Emergence in this context is the idea that simple rules can lead to complex behavior. A good example we have all seen is a flock of birds. They all fly together, and then suddenly change direction together. It almost looks like they are one big organism. Similar behavior can be seen in schools of fish and locusts. One bird CEO doesn’t issue commands to make it happen, each autonomous individual has a simple rule — do what the bird next to you does. This one simple rule results in the flock wheeling around together, and then landing together.
Similarly, when we start to translate metrics from the top of the organization down, we are following a simple rule set at each level, which might be stated as, “If your boss wants this, what do you do that relates to it?” From this simple (not easy) question, we can build metrics that make sense for an entire business. There doesn’t need to be one person who decides what everyone should measure, each autonomous individual gives their best guess as to how what they do relates to the business objectives, all the way down to the process itself.
Emergence also shows up throughout the organization as changes happen. Whether it is a dynamic change in the machines, work environment, or strategic direction, each individual responds in ways that, optimally, tune their area a bit better to adapt to the change. Overall, it looks like the business is adapting dynamically, but it is really the action of a lot of individuals to what they see day-to-day. This can be a bad thing, though, in the absence of the integrated metrics I mentioned before. Dynamic change without metrics to keep individual adaptation on the rails means dynamic change like cancer — doing things that locally look like the right thing to do, but hurt the organism as a whole.
Another thing that keeps businesses manageable is regression to the mean. This just means that, while there is variation every day, it doesn’t suddenly take off to the stratosphere. If things were a little extreme today, they will be likely less so tomorrow. As a manager, it is important to know about regression to the mean. If you don’t, you might ascribe that change to something that you or someone else did, not realizing it was likely to happen all by itself. Statistical Process Control (SPC) used in the process and in management will help you keep a handle on when it is time to react and when it is time to leave the process alone.
Finally, and most hopeful for managers, is the idea of the Pareto principle. There are about a bajillion things going on right now in your business. Mary is having a hard time with her boyfriend, Mac just got that date with his crush, the docks just received a batch from a supplier that is a little bit different. Heck, even the phase of the Moon is different today! But not all of these things have an equal effect on your business. The Pareto principle posits that 80% of the consequences comes from 20% of the causes. In practice, that means that, while there are a lot of things in motion right now, the number of metrics that really control the business are few. You don’t need to measure everything to measure enough to know what to do as a manager.
Of course, the trick is in finding out which metrics those are.
In the process I just described, we deploy metrics by asking what it is you do that affects what your boss needs. My experience is that most people get about 80% of them right. They do miss some important metrics and there are other metrics they thought were important that turn out not to be, but mostly they do a pretty good job of identifying their own metrics. Of course, we don’t just rely on their guesses — once you start collecting data you can test to see if they are missing something or if they have metrics that are not valuable enough to continue to collect.
What I have just described might sound like it is prone to spiraling out of control, as in a game of “Telephone,” where you whisper a sentence to the first person in a line of people, then they whisper to their neighbor and so on, until the phrase, “Elephants are scared of mice” becomes “Elegance is sharing rice” and the original intention is lost.
How it actually works is each level down tends to become more robust to errors above them. Frankly, I don’t know if it is a consequence of emergence, self-correction, or the Pareto principle, or some combination. Unless someone makes a really big error, or the organization adds or changes its success measures in some fundamental way, you don’t see big changes required down closer to the process.
That said, most everyone gets some metrics wrong, so each level of management must use their collected data to test the usefulness of their own metrics, and if they change anything, it might have an effect down into their area.
The big take-away here is that running a business is a knowable endeavor. That, while there is a lot happening, you don’t need to manage all of it. You only need to manage those few metrics for each person that controls the outputs you need for your business.
And that is good news for you when you get that job as CEO!