2 June 2021, 08:14

on systems thinking


The behavior of a thing is latent in its structure.

e.g: A Slinky toy rests in a person’s palm and another hand grasps the top of the slinky several coils down. Removing the hand results in the Slinky dropping and then bouncing up and down. The hand did not make the Slinky bounce. The Slinky itself possessed the necessary properties for that specific behavior. Understanding the relationship between behavior helps us understand how poor outcomes happen and how to generate better ones.

“It is a way of thinking that gives us the freedom to identify root causes of problems and see new opportunities.”

A system is an interconnected set of things (people, cells, molecules, components, etc.) that produces a specific outcome over time.

  • A system can be triggered or influenced by external forces.
  • The system’s response to external forces is a characteristic of the system itself (see slinky example).
  • An external force that triggers a system response will likely result in a different outcome when applied to a different system

“The behavior of a system cannot be known just by knowing the elements of which the system is made.”

System Structure

A system consists of three things:

  • Elements
  • Interconnections
  • A function or purpose

e.g: The digestive system

Elements: Teeth, enzymes, stomach, intestines. Interconnections: The physical flow of food, chemical signals to regulate the process. Function: To separate nutrients from food (to maintain and provide energy to the body) and to collect and discard unusable waste.

Systems don’t exist in isolation.

  • Systems can be interconnected with other systems.
  • Systems can be embedded within other systems

Conglomeration: A collection of things that LACK interconnections or function. Example: Sand scattered on a road is not a system

“When a living creature dies, it loses its ‘systemness.’ The multiple interrelations that held it together no longer function, and it dissipates, although its material remains part of a larger food-web system.”

“A system is more than the sum of its parts. It may exhibit adaptive, dynamic, goal-seeking, self-preserving, and sometimes evolutionary behavior.”

The flow of information, from one part to another, is a common way that system interconnections are manifest.

“The best way to deduce the system’s purpose is to watch for a while to see how the system behaves.”

  • Watch what the system does not what is says or advertises itself to be doing.
  • e.g: “If a government proclaims its interest in protecting the environment but allocates little money or effort toward that goal, environmental protection is not, in fact, the government’s purpose.”

Self-perpetuation is an important function for almost every system.

System purposes need not be human purposes nor need they be the intended purpose. It is common for a well-meaning system to result in unintended outcomes and behaviors.

Successful systems keep “sub-purposes and overall system purposes in harmony.”

Changing elements in a system often has the least impact on the overall system behavior. Modifying the interconnections and purpose often results in dramatic or fundamental changes.

“The least obvious part of the system, its function or purpose, is often the most crucial determinant of the system’s behavior.”

System Basics

A stock: This is an element in the system that is accumulated, depleted or stored over time.

e.g: water in a bath, the population of a city, books in a bookstore, money in a bank.

“A stock is the memory of the history of changing flows within the system.”

A stock changes over time based on flow. A flow represents elements entering the system stock as well as elements leaving or being depleted from a system stock.

e.g: For a population, an inflow might be represented by births (these ADD to the population stock). Outflows might be represented by deaths (these result in a reduction in the population stock).

Stock-and-flow diagram: A visual representation of a system that shows inflows, outflows, reinforcing and balancing loops, stocks and interconnections with other systems.

graph LR
    A[Water] -->|inflow| B(Water in bath)
    B --> |outflow| C[Water]

System diagrams are simplified representations of the real world “The map is not the territory.”

"If you understand the dynamics of stocks and flows—their behavior over time—you understand a good deal about the behavior of complex systems."

Dynamic equilibrium: A state in which inflows and outflows are equal resulting in an unchanged stock.

“A stock takes time to change because flows take time to flow.”

This is important because people underestimate the time needed to change large or complex systems.

"Stocks allow inflows and outflows to be decoupled and to be independent and temporarily out of balance with each other."

e.g: A water reservoir allows us to maintain stability in the availability of water. People can receive steady flow of water whether there is a temporary drought or if it’s the rainy season.

“Systems thinkers see the world as a collection of stocks along with the mechanisms for regulating the levels in the stocks by manipulating flows

Feedback loops

Feedback loop: A mechanism for regulating a system’s behavior. Feedback loops occur when a change in stock affects the inflow or outflow of that same stock.

Feedback loops can affect stocks in several ways:

  • Maintain the level of a stock within a narrow range.
  • Cause a stock to grow.
  • Cause a stock to decline.

Balancing feedback loops are loops that aim to stabilize a stock within a desired target or range.

The discrepancy or gap between actual and desired levels in a stock causes a decision to increase or decrease additional inflows into said stock.

Feedback loops can work in both directions (inflows/outflows) and can regulate increased or decreased flow rates.

e.g: A coffee drinker might drink a cup when their energy is running low (in this example, the stock is energy). The drinker notices their energy level is ebbing and opts to drink a cup of coffee. The inflow of caffeine results in an increased stock of energy.

“Balancing feedback loops are goal-seeking or stability-seeking....a balancing feedback loop opposes whatever direction of change is imposed on the system.”

Remember: the mere presence of a feedback loop isn’t sufficient to ensure that a system is working well (the loop may be insufficient or too weak to maintain the desired result).

Reinforcing feedback loops are loops that amplify or reinforce growth or destruction within a system.

“A reinforcing feedback loop enhances whatever direction of change is imposed on it.”

e.g: Inflationary pressures. As prices for goods increase, wages increase. As wages increase, prices also must increase to maintain profits. e.g: Compound interest. Money is saved and as interest in earned, the interest is added to the stock of savings which increases the overall stock that is earning interest.

“Reinforcing loops are found wherever a system element has the ability to reproduce itself or to grow as a constant fraction of itself. Those elements include populations and economies.”

Reinforcing loops can lead to exponential growth but that can also result in runaway collapse.

Consider: If A causes B, is it possible that B also causes A?

e.g: If someone says that population growth causes poverty, ask yourself if poverty causes population growth.

“The concept of feedback opens up the idea that a system can cause it's own behaviour.”

A Brief Visit to the Systems Zoo

One-Stock Systems:

A Stock with Two Competing Balancing Loops

e.g: A thermostat

  • Stock is the temperature of the room which is influenced by inflows and outflows.

  • Inflow comes from heat from the furnace.

  • Outflow results from temperature outside (which is presumably colder).

  • One balancing feedback loop regulates the inflow: A desired temperature compared against the real temperature of the system stock. The delta between desired and actual influences the subsequent flow of more heat (or reduced inflow).

  • One balancing feedback loop regulates the outflow: The outside temperature influences the inside temperature based on the delta between outside temp and inside temp. A higher delta (e.g. between high room temp and low outdoor temp) results in increased outflow of indoor heat.

  • The two balancing loops compete.

Note that owing to the flow of information and time needed to impact the stock, there are inherent delays to rebalancing the stock. Any actions taken (e.g. to increase inflow of heat) can only affect future behavior and stock.Because of the competing outflow, the thermostat inflow needs to be set HIGHER than the target or desired temperature.

  • Because of the competing outflow, the thermostat inflow needs to be set HIGHER than the target or desired temperature.
A Stock with One Reinforcing Loop and One Balancing Loop

e.g Population and industrial economies

  • Stock is the population (city, nation, world).

  • Systems have key driving variables. For a population these are fertility and mortality.

  • Inflow is the result of births which add to the population stock.

  • Inflow is governed by a reinforcing loop. As births increase (fertility levels), the population stock increases which drives more births and adds more rapidly to the population stock.

  • Outflow is the result of deaths which reduces the population stock.

  • Outflow is regulated by a balancing loop—mortality rate.

  • Population grows when the birth rate outpaces the mortality rate. Population declines when the reverse is true.

  • Changes in the flows change the (over time) the behavior of the stock.

  • Shifting dominance refers to situations where one feedback loop dominates the system. The loop that dominates the system determines its behavior.

“A stock governed by linked reinforcing and balancing loops will grow exponentially if the reinforcing loop dominates the balancing one. It will die off if the balancing loop dominates the reinforcing one. It will level off if the two loops are of equal strength.”

  • In reality, loop dominance will shift back and forth in sequence over time.

fertility and mortality are governed by their own feedback loops (they can be modeled as discrete systems that interconnect with the population system).

  • An economy bears similar behavior to the population loop.

    • Stock = capital
    • Inflow = investment
    • Reinforcing feedback loop: increased capital stock leads to reinvestment and increasing capital over time.
    • Outflow = depreciation
    • Balancing loop: lifetime of the capital affects depreciation. The longer the lifetime, the smaller fraction of capital needs to be retired/replaced annually.

“Systems with similar feedback structures produce similar dynamic behaviors.”

A System with Delays — Business Inventory
  • Stock is the product inventory itself — vehicles for sale.

  • Inflows are the deliveries from factories.

  • Balancing feedback loop regulates the inflow to ensure that there are 10 days of vehicle inventory in stock.

  • Outflows are the sale of new cars to consumers.

  • Balancing feedback loop regulates the outflow (customer demand). Dealer can monitor sales and sales trends. If forecast is higher, the dealer can modify the inflow of new vehicles accordingly.

  • Delays are inherent to the system. The response to each balancing loop is not immediate or always accurate.

  • Perception delay: The dealer bases their ordering decisions on a 5-day average to smooth the temporary spikes and dips in demand.

  • Response delay: The dealer doesn’t adjust inflows in a single order. They make up a fraction of any shortfall with each subsequent order. Changes in inflows occur over several days (rather than as a one-time, immediate response).

  • Delays in balancing feedback loops result in system oscillations (fluctuations over time of inventory levels).

Other delays: production delays, delivery delays, construction delays.

  • Other delays: production delays, delivery delays, construction delays.

“Delays are pervasive in systems, and they are strong determinants of behavior. Changing the length of a delay may make a large change in the behavior of a system.”

Two-Stock Systems:

A Renewable Stock Constrained by a Nonrenewable Stock

e.g: An oil company

  • This model considers environmental constraints (which is lacking in the simpler single stock models).

  • Entities that exchange things with the environment need both a supply of the resource and a way to dispose of waste and byproducts of the process.

“Any physical, growing system is going to run into some kind of constraint sooner or later. That constraint will take the form of a balancing loop that in some way shifts the dominance of the reinforcing loop driving the growth behavior, either by strengthening the outflow or by weakening the inflow.”

  • Systems that exhibit growth: look for a reinforcing loop driving inflows and a balancing loop that constrains it.

Renewable vs. nonrenewable pollution constraint (outflow):

Nonrenewable: The environment cannot absorb the pollutant. Renewable: The environment can absorb the pollutant and make it harmless.

Two constraints:

Resource-constrained systems: Inflow supply constraints. Pollution-constrained systems: Outflow constraints (production is not feasible because the damage to the environment is too great or hazardous).

  • Constraints on a system can be either temporary or permanent.

“A quantity growing exponentially toward a constraint or limit reaches that limit in a surprisingly short time.” The higher and faster the growth, the faster and more pronounced the fall (when building capital stock against a nonrenewable resource).

  • The size of the constraining resource is the key variable in this system.

A Renewable Stock Constrained by a Renewable Stock

e.g: A fishing economy.

  • Living vs. nonliving renewable resources:

  • Nonliving renewable resources: sunlight, wind, water in a river are regenerated through steady input regardless of the state of that stock.

  • Living renewable resources: fish, trees, cattle. These resources can regenerate and regrow. Note that the regeneration rate of these resources are not constant and are constrained by their stock and environment.

The fishing economy is governed by three variables:

  • Price (scarcity determines the price).
  • Regeneration rate (scarcer fish stocks are replenished more slowly).
  • Yield per unit of capital (efficiency of the fishing technology and methods).

“Renewable resources are flow limited. They can support extraction or harvest indefinitely but only at a finite flow rate equal to their regeneration rate.”

  • Some over-extracted renewable resources can be driven below a critical threshold and effectively become nonrenewable (e.g. fish stocks).

Why Systems Work So Well

  • Many systems possess one of the following characteristics: resilience, self-organization, or hierarchy.

Resilience: A measure of a system’s ability to adapt and survive in a changing environment. Brittleness and rigidity are the opposite of resilience.

  • Feedback loops that regulate the balance in a system are essential to system resilience.
  • e.g: The human body can tolerate different temperatures, variations in food supply, and repair itself and compensate for missing parts.
  • Resilience is not the same as being static or constant.

“Because resilience may not be obvious without a whole-system view, people often sacrifice resilience for stability, or for productivity, or for some other more immediately recognizable system property.”

Self-Organization: The characteristic in which some systems can learn, evolve, self-regulate, and develop complexity on its own.

“Self-organization produces heterogeneity and unpredictability.”

  • Freedom, experimentation, and disorder are necessary conditions for self-organization (many people are uncomfortable with these properties).

Hierarchy: Self-organizing systems generate hierarchies of nested systems. Systems with subsystems as you move down and systems aggregated/comprising larger systems as you move up.

e.g: A cell in your heart is a subsystem of the organ. The organ is a subsystem of a larger circulatory system. The circulatory system, along with other physiological systems, comprises a person. In turn, a person is a subsystem of a family or group of people. Groups of people are subsystems of a larger societal organization (neighborhood, city, state, nation, etc.).

  • Hierarchies give structure to systems (yielding stability and resilience) and also allow for efficiency and specialization. Certain subsystems are tasked with certain goals and subsystems only need information essential to those goals and behaviors.

  • Balance in the hierarchy is necessary for optimal behaviors:

  • Sub-optimization occurs when a subsystem’s goals dominate at the expense of the overall system goals.

  • Central control occurs when the system at the top of the hierarchy prevents subsystems for operating efficiently or freely.

“Hierarchical systems evolve from the bottom up. The purpose of the upper layers of the hierarchy is to serve the purposes of the lower layers.”

Why Systems Surprise Us

  • Much of our thinking makes use of models to explain and make sense of the world. These models may or may not accurately reflect reality.

“System structure is the source of system behavior. System behavior reveals itself as a series of events over time.”

Remember: many system relationship and interconnections are NOT linear. Moreover, the relative strength or influence of these relationships are dynamic.

Remember: the language we use to describe a system biases our understanding of the system.

E.g: Labeling an undesirable outcome of a system as a “side effect” suggests that the outcome was unanticipated, poorly considered, and unwanted.

Garrett Hardin (ecologist):

“Side effects no more deserve the adjective ‘side’ than does the ‘principal’ effect. It is hard to think in terms of systems, and we eagerly warp our language to protect ourselves from the necessity of doing so.”

  • Simplified systems models can lead to errors. For instance, with simple, single stock-and-flow diagrams, we make inherent assumptions about the nature and availability of the source for a flow or the sink for an outflow. This may be necessary to examine a system closely for a given task, but it’s worth noting that this activity opens us up to blindspots and assumptions.

Remember: Systems form a continuum. Systems are connected with other systems (on both ends of the model). The boundaries we draw around a system are arbitrary and serve the purposes of our analysis and the questions we want answered.

Limiting factor:

“A necessary system input that is the one limiting the activity of the system at a particular moment.”

“Any physical entity with multiple inputs and outputs is surrounded by layers of limits.”

“There always will be limits to growth. They can be self-imposed. If they aren’t, they will be system-imposed.”

Remember: Do not underestimate the importance of system delays.

Bounded rationality: People make reasonable decisions based on the information available, but they don’t always have complete or correct information available.

“The bounded rationality of each actor in a system may not lead to decisions that further the welfare of the system as a whole.”

System Traps and Opportunities

Archetypes: System structures that exhibit standard or characteristic behaviors. These include problematic patterns of behavior.

These systemic problems are often blamed on specific actors or events but the problems are more likely to be inherent to the system.

“Blaming, disciplining, firing, twisting policy levers harder, hoping for a more favorable sequence of driving events, tinkering at the margins—these standard responses will not fix structural problems.”

Archetype: Policy Resistance (aka “fixes that fail”)

“Policy resistance comes from the bounded rationalities of the actors in a system, each with his or her own goals.”

E.g: Wars on drugs that fail to reduce the use or prevalence of drugs.

Consider a single-system stock of the drug supply in a city.

Consider the divergent goals of different subsystems or actors within the system:

  • Addicts want to keep the drug stock high.
  • Law enforcement wants to keep the drug stock low.
  • Pushers want to keep it in the middle so prices are moderate (not too high and not too low).
  • Average citizens want to be safe from crime.
  • When one actor gains an advantage that moves the system stock (drug supply) to their advantage, the other actors will increase their efforts to move it in the opposite direction. The result is a standoff in which the stock remains constant (which nobody wants).

Two possible resolutions:

Overpower the system with one of the goals: this requires tremendous resources (say from law enforcement) and carries high costs whether it succeeds (resentment from the population, mass incarcerations, etc.).

Capitulate (counterintuitive): stop spending resources and energy on enforcing or resisting. e.g: Ending Prohibition in 1933 in the USA.


← on being anti-fragile
on service design →