Thinking in Systems (abridged version)

11 min readOct 8, 2019

How to apply system's thinking to managing a company

Disclaimer

The purpose of this post is to give a very brief summary of the wonderful book Thinking in Systems: a Primer by Donella Meadows and the accompanying blog post Leverage Points: Places to Intervene in a System, by the same author, and how I see it applies to organizational structure. My intention here is to bring the general intuition, rather than be precise with system's theory (I'm most definitely not an expert in this) concepts. I also intentionally left out specific points made in the book that I felt were too abstract when applied to a company, or just generally hard to act upon in this context. All the quotes are directly taken from Donella Meadows' works.

A Simple System: A Bathtub

A basic system has the following attributes (as shown in the image below):

Inflows: The resources, capital, labor, and/or information that increase the state of the system.
Outflows: The resources, capital, labor, and/or information that consumes the state of the system.
State of the system: The standing stock of importance for the system.
Perception of the state: A measurement of the standing stock of importance.
Goal of the System: The desired level of the standing stock of importance
Discrepancy between perception and goal: The difference between desired state and the actual state.

Let's use bathtub as an allegory for a system. In this case the attributes would be as following:

Inflows: water getting into the bathtub from the faucet
Outflows: water leaving the bathtub down the drain
State of the system: the water level, and temperature of the bathtub
Perception of the state: the observation of water level, and temperature made by someone monitoring the bathtub.
Goal of the System: the desired water level, and temperature
Discrepancy between perception and goal: the difference between the desired temperature, and water levels, and the actuals.

I will use the bathtub analogy whenever possible in order to simplify some points.

A company as a system

"A system is a set of things—people, cells, molecules, or whatever—interconnected in such a way that they produce their own pattern of behavior over time. The system may be buffeted, constricted, triggered, or driven by outside forces. The system’s response to these forces is characteristic of itself, and that response is seldom simple in the real world."

As such, one can consider any company is a system, with complex, intertwined dependencies between people, teams, investors, customers, media, regulators, competition, utilities providers, and even elevator companies. It's also a bit more complex than a bathtub.

Because everything is interconnected, any action, or decision may lead to trickling second-order effects that were never initially anticipated. As a consequence, changes in any element of the system will lead to reactions that have massive impacts. Ensuring these impacts are virtuous, and improve us is paramount. On the other hand, if we are not capable of considering these effects in our decision making, they very easily may lead to destructive cycles in the long run.

Inflows: ideas, user requests, competitor information, investor demands, people, etc.
Outflows: features, reports, statements, interviews, presentations, bugs, etc.
State of the system: backlog, dependencies, blockers, engagement, stress, etc.
Perception of the state: how the company is doing in the eyes of: users, investors, employees, management, regulators, media, etc.
Goal of the System: Vision, OKRs, team mission, individual aspirations, etc.
Discrepancy between perception and goal: Performance evaluation, company results updates

Places to intervene in a system (in increasing order of effectiveness)

These are the ways one can intervene in a system, in increasing order of effectiveness. In this context, effectiveness means the leverage a change has in changing the whole system. Small changes in higher leverage points may lead to transformational changes in the system, both for good, and for bad. Lower leverage points have less of a transformational quality to it, but can also significantly impact the system, also either for good or bad. Generally speaking, we should be more careful and deliberate when changing higher leverage points, and we shouldn't expect massive changes when only dealing with lower leverage points.

8. Constants, parameters, numbers

This is where most managers engage in a system: the number of employees, the number of teams, the number of features, the number of hours worked, the number of ideas had, or decisions made. While definitely a leverage point, and the simplest one to engage with, it's the least effective in impacting the system by itself.

"If the system is chronically stagnant, parameter changes rarely kick-start it. If it’s wildly variable, they don’t usually stabilize it. If it’s growing out of control, they don’t stop it."

Turning a faucet harder won't do anything if the pipes are blocked, adding more faucets won't help reach the goal if the temperatures of each are not set, bringing in a firehose might just lead to more chaos if it's not possible to turn it off in time, etc. Just adding more people, or resources to a team won't necessarily make it successful. Just adding more features to our product won't necessarily make it successful.

Because of this, I'd encourage us to go beyond (not ignore it) talking about parameters, and consider the other leverage points that exist in a system.

7. The lengths of feedback delays (relative to the rate of system changes)

Feedback delay can be best exemplified going back to the bathtub analogy. We've all turned a faucet that took some time to heat up or cool down the water (and naturally burned ourselves or froze ourselves in the process). The time from the faucet turn, and the change in the water temperature is the length of feedback delay.

Some of the delay is structural (measuring a temperature will always take some time, and water flowing from point A/B also takes time), but some might be the result of inefficiencies that could be improved with investment (maybe the heater is in the basement and could be moved closer to the bathroom in a renovation).

"Delays that are too short cause overreaction, “chasing your tail,” oscillations amplified by the jumpiness of the response." If we react to every user feedback, or to every new competitor article we will never be able to deliver on a specific goal. If we don't allow time for credit tests to perform, we might realize a few months from now that we've turned the heat way too high and find ourselves in scalding water.

"Delays that are too long cause damped, sustained, or exploding oscillations." If we can't deliver a bug fix in a timely fashion, we might go bankrupt. If we can't deliver customer value before another competitor we might lose market share. A system just can’t respond to short-term changes when it has long-term delays. That’s why a massive central-planning system, such as the Soviet Union and any organization with centralized/top-down decision making, necessarily functions poorly at scale.

On this point, also consider that any change in prioritization is filled with feedback delays. A decision takes time to be aligned, a new strategy takes time to be validated, a product takes time to be developed, and adopted.

6. Regulating negative feedback loops

"A complex system usually has numerous negative feedback loops it can bring into play, so it can self-correct under different conditions and impacts. Some of those loops may be inactive much of the time — like the emergency cooling system in a nuclear power plant, or your ability to sweat or shiver to maintain your body temperature — but their presence is critical to the long-term welfare of the system."

In the bathtub example: to prevent water from overheating, the heater should shut itself down after a certain temperature is reached; to prevent water from overflowing one can use floating devices that stop the water supply once it reaches a certain level (similar to how toilet tanks' valves work).

Ensuring we have regulating actions that are triggered by negative feedback loops is fundamental to ensure we're on the right track.

What's important with this leverage point is to capture the main feedback loops that have the chance to derail the business, which are the reinforcing feedback loops on core parts of the company. By reinforcing I mean, when there's a negative output, that negative output increases the likelihood of more negative outputs (essentially a vicious cycle).

In addition to that, it's important that when these negative feedback loop triggers are reached, that the actions they triggered are effective in normalizing the system, taking it back to a stable state.

5. Driving positive feedback loops

"A negative feedback loop is self-correcting; a positive feedback loop is self-reinforcing. The more it works, the more it gains power to work some more. The more people catch the flu, the more they infect other people. The more babies are born, the more people grow up to have babies. The more money you have in the bank, the more interest you earn, the more money you have in the bank. The more the soil erodes, the less vegetation it can support, the fewer roots and leaves to soften rain and runoff, the more soil erodes."

Identifying opportunities for positive feedback loops that can be fostered safely is a huge leverage point. The clearer example would be creating viral functionalities that encourages a member get member dynamic. However, another positive feedback loop would be investing in better tooling for development, as with better tools you can create even better tools, and build on top of. Developing open source projects have similar dynamics: making more people contribute to something will make it more widely used, and will lead to more people wanting to contribute to it.

The thing with this leverage point is that it's hard to find instances where these positive feedback loops can actually be triggered (very often there are self-correcting feedback loops that stop them at some point), and when they are triggered, it's possible for them to get out of control and lead to chaos.

4. Information flows

"There was this subdivision of identical houses, except that for electric meter placement. In some of the houses it was installed in the basement, and in others it was installed in the front hall, where the residents could see it constantly, going round faster or slower as they used more or less electricity. With no other change, with identical prices, electricity consumption was 30 percent lower in the houses where the meter was in the front hall."

This is an example of the inclusion of an informational feedback loop into a system. The addition of more information modifies the system to consider more variables in its decisions. Turning on the hot water to the bathtub now has the added information of its cost, and this directly, or at times indirectly, modifies how the system behaves.

On the other hand, more information, and more feedback loops are not necessarily always a good thing. If in addition to measuring the cost of electricity, the meter included water pressure, room temperature, external temperature, precipitation likelihood, and levels of water salinity, it would be a lot harder to filter signal from noise, and the system would either react inappropriately to irrelevant information, or waste time filtering it out.

At a company, creating the appropriate level of information flows is fundamental for better decisions to be made at all levels: what does the customer care about, what are the costs associated with this decision, how was this PR statement received, how did someone feel when you gave them feedback, what are the complexities of pursuing an idea, etc. At the same time, we need to be aware that no individual team, and especially no individual person, will be able to carry all the information that impacts the company as a whole.

Additionally, very often the act of fetching a new piece of information will impact the system. Placing a thermometer into a bathtub to check its temperature will change its water level, asking a team to create a presentation on their progress will divert their efforts from delivering on their goals to generate that information. This doesn't mean at all that we shouldn't demand some pieces of information, it just means we need to be aware of the costs.

3. The rules of the system (incentives, punishments, constraints)

"The rules of the system define its scope, its boundaries, its degrees of freedom. Thou shalt not kill. Everyone has the right of free speech. Contracts are to be honored. The president serves four-year terms and cannot serve more than two of them. Nine people on a team, you have to touch every base, three strikes and you’re out. If you get caught robbing a bank, you go to jail."

The importance of the rules of the system are so fundamental we often take them as a given. Imagine inverting some baseline rules from a company and how much of a change it would lead to: what would happen if each employee could define their own salary, if social media posts needed to be approved by the government, if we always decided promotions based on a democratic vote, if we fired anyone who disagreed with the CEO, if discrimination was allowed, etc.

The rules of a company as a system are one of the key drivers of culture. Who and what do we reward, reprimand, or punish what direct consequences on the overall behavior of the system. Decisions around promotions, raises, hiring, recognition don't exist in a vacuum, and should be considered as such. It's easy to incentivize behavior we don't actually want (praising hero culture antics, while not praising stability; praising effort/frantic motion, instead of results, etc), and to punish behavior we ultimately want (punishing honest feedback because it's uncomfortable, censoring dissenting ideas, etc).

2. The power to add, change, evolve, or self-organize system structure.

"The most stunning thing living systems, and some social systems can do is to change themselves utterly by creating whole new structures and behaviors. In biological systems that power is called evolution. In human economies it’s called technical advance or social revolution. In systems lingo it’s called self-organization.
Self-organization means changing any aspect of a system lower on this list — adding completely new physical structures, such as brains or wings or computers — adding new negative or positive loops, or new rules. The ability to self-organize is the strongest form of system resilience. A system that can evolve can survive almost any change, by changing itself.
Insistence on centralized decision making, rigid organizational structures, one size fits-all decisions shuts down learning. Cuts back resilience. Any system, biological, economic, or social, that gets so encrusted that it cannot self-evolve, a system that systematically scorns experimentation and wipes out the raw material of innovation, is doomed over the long term on this highly variable planet.
The intervention point here is obvious, but unpopular. Encouraging variability and experimentation and diversity means “losing control.” Let a thousand flowers bloom and ANYTHING could happen! Who wants that? Let’s play it safe and push this leverage point in the wrong direction by wiping out biological, cultural, social, and market diversity!"

At scale there's a constant desire for more standardization and processes, in the desire to both ensure control (which is generally a futile exercise), and leverage learnings (which is extremely important with parsimony in order to reap company wide benefits from these learnings).

To battle this, organizational structures should encourage diversity in org structure, decision making process, goals, etc. The ability for a team to self-organize, within some constraints, is key for keeping the system adaptable.

1. The goals of the system

The ultimate leverage point is reestablishing the goals of the system. Every element of the system, and every subsystem in them will adapt themselves in order to reach the goal of the system. This is is the reason selling a dream, or establishing a vision can be so powerful: it gets everyone working towards the same direction leading to amazing results. This is also the same reason why a change in priority can feel so painful: it fundamentally destroys or repurposes every element of the system, and getting everything to connect back towards a new direction takes a lot of effort and time.

Operating in this leverage point has extremely high leverage, and therefore clarity is of the utmost importance. Any element of the system working towards disparate goals will lead to confusion, misalignment, and destructive actions as the whole system tries to correct it. Establishing clear goals, clear definitions of success, clear strategy, and clear expectations are the highest leverage thing we can do at any company.