Complex dynamic situations where things interact with each other are messy and require systems-level thinking. In her book Thinking in Systems, Donella Meadows gives us the tools necessary to understand and improve the systems around us. The applications of these tools are vast, ranging from personal life to business to global issues. This is going to be a two-part post. In this post, I summarize the basics of systems thinking, including its loops and characteristics, and in the next post, we will look at common system traps and how to go about changing a system.
What is Systems Thinking?
Systems thinking is understanding how its constituent elements interconnect and work together to achieve the overall goal. The elements of a system do not follow a linear relationship; there are loops in systems that reinforce or balance specific metrics or actions. Non-linear feedback loops are the essence of systems thinking.
See the example illustrated below. It shows how (follow it from left to right) as the demand increases, production is increased, leading to filling up the inventory, which helps marketing or sales to increase, thereby increasing the customers and demand, which increases the profits, and so on. This is a reinforcing loop, and we’ll look at feedback loops in more detail in the next section.
Systems have their own inherent behaviors. Outside actions can influence or unleash these behaviors, but the same actions will affect a different system differently. To understand systems, we need some common terms like stocks (which represent quantities) and flows (which represent the increase or decrease in certain values of stocks). We will use these terms to understand the 4 laws of systems thinking.
There are also some unique idiosyncrasies of systems that we need to understand and be wary of:
- Today’s problems come from yesterday’s solutions - Coal-generated electricity led to pollution; fertilizers to boost production led to cancers and low yields, etc.
- The harder you push, the harder the system pushes back - Prohibition, war on drugs, marijuana ban, etc.
- The easy way out usually leads back in - When we act without thinking, we end up with solutions that need more and more solutions.
- Faster is slower - Systems slow down when we try to put in quick fixes. As the old saying goes, “we never have time to do it right, but we always have time to do it over.” The classic example was how in Covid times we fast-tracked the vaccine, then we had trouble distributing the vaccines, and they were not as effective while the vaccinated people stopped using masks, and so on.
Feedback Loops
There are two types of feedback loops:
- Self-balancing or stabilizing - These cause a stock to decrease after an increase.
- Amplifying - These usually cause exponential growth.
Looking for feedback loops, instead of simple causal linear relationships, is the first thing we need to do when trying to think in terms of systems. It must be said that all feedback-driven actions affect future results, but not immediately, as there are always delays in reactions. Take the example of a thermostat that reacts to the loss of heat in a room due to someone opening the door. The heat loss begins immediately but becomes noticeable by the thermostat after a delay, which then activates the heater, which then takes some time to reheat the room. Systems thinking eventually comes down to understanding and planning for these delays.
Most systems have both reinforcing loops and balancing loops. Feedback loops help systems gain three essential qualities:
- Resilience - Feedback loops that restore the core behavior of a system, like the human body that self-repairs organs, make the system more resilient.
- Self-organizing - This is the system’s ability to take basic constituents and turn them into something bigger and complex. The best example is how evolution took a bunch of chemicals (primordial soup) and turned them into a set of organisms.
- Hierarchical - Hierarchy, here, implies a system is composed of stable subsystems that deal with a small part of the overall system. These subsystems may be further divided into subsystems which are less complex and therefore, more durable.
Evaluating Systems
You can’t evaluate a large system by looking at each small part or feedback loop and its actions. You must look at long-term behaviors and structures to understand it as a whole. The following characteristics, when looked at in detail, can shed light on the system’s holistic operations and help one understand it better:
- Boundaries - Start by understanding the extent of the problem. For example, when thinking about sewage disposal, you can’t just dump it all in the river, as the town downstream will face the brunt of the pollution. They will then want to extend the boundary of the pollution to include the river itself. Drawing the correct boundary is the first step in understanding the system.
- Bounded Rationalities - Simply put, if the invisible hand of the market really worked always, we would not have situations where people do things in their own interest but end up making things worse for everyone in the long run. Overfishing, farming surpluses, and tourists complaining about other tourists crowding out famous destinations are all examples of people acting rationally for themselves and ending up making things bad for the whole. This happens because of our lack of overall information about the distant parts of the system. Herbert Simon called this bounded rationality. Fishermen fish as much as they can as they don’t know how many fish there are and how much or how little will be left for them if they didn’t.
- Limiting Factors - Every system has elements that limit the progress of that system. Bread will not rise if not for yeast, no matter the flour; plants won’t grow without sunlight, no matter the fertilizers, etc. It is common for species to be interdependent in nature. For example, trees of various kinds and worms can form self-balancing delicate systems. Move any one of them slightly, and it may send the entire ecosystem into chaos.
- Non-Linearities - As discussed above, a peculiarity of systems is that its behaviors are not linearly causal. Taking an example from nature again, predator-prey relationships are mostly non-linear. When one species grows, the other grows unbounded with them, and then one falls, the other can also dissipate entirely. This is why it is important to understand the non-linear relationships in a system.
- Delays - Elements comprising a system have in-built delays, like a car production system has delays in the design and manufacturing cycle or a pig breeding setup has a delay of 6 months, etc. These delays play the role of oscillators that affect whether to focus on short or long-term policy making. It also drives the frequency of actions and is thus important to understand and alleviate problems before they become a limiting factor.
In the next post, we will look at the common system traps and how to go about altering a system’s characteristics to make it more efficient.
I run a startup called Harmonize. We are hiring and if you’re looking for an exciting startup journey, please write to jobs@harmonizehq.com. Apart from this blog, I tweet about startup life and practical wisdom in books.