Chapter 4: Systems Thinking - Seeing the Relationships Behind the Problem

~~Chapter 4: Systems Thinking - Seeing the Relationships Behind the Problem~~

Learning Objectives

By the end of this chapter, you should be able to:

define a system in the context of a business case;
identify the elements, relationships, and purpose of a system;
establish an appropriate boundary for the analysis;
recognise relationships among different parts of a business problem;
distinguish events, patterns, underlying structures, and mental models;
distinguish symptoms from root causes;
identify reinforcing and balancing feedback loops;
recognise delays between actions and outcomes;
anticipate unintended and second-order consequences;
identify common system patterns that produce recurring problems;
locate leverage points where focused intervention could create wider impact;
test whether a proposed solution improves the whole system rather than one isolated metric;
communicate a complex system clearly without overcomplicating the presentation.

Why This Matters

Business problems rarely exist in isolation. A decline in sales may appear to be a marketing problem, but the underlying causes might include:

declining product quality;
longer delivery times;
poor customer service;
outdated pricing;
weak product availability;
changing customer expectations;
a competitor's new offering;
employee turnover;
a combination of these factors.

An organisation is a system. Its:

customers;
employees;
suppliers;
competitors;
partners;
technology;
finances;
operations;
leadership;
incentives;
stakeholders

interact. A decision made in one part of the organisation can have consequences elsewhere. For example: Reduce staffing → lower labour cost → longer customer wait times → lower satisfaction → declining retention → lower revenue → greater pressure to reduce costs. The original decision improves labour cost in the short term but may weaken the organisation's financial position over time. This is why a recommendation that looks attractive in isolation can create a new, sometimes larger, problem. Systems thinking helps case teams understand:

how the problem was created;
why it may continue;
what other outcomes are connected to it;
where an intervention may have the greatest effect;
what could happen after the Recommendation is implemented.

Discover Your MAD Skills Principle

Don’t just look at the problem. Look at the system that keeps producing the problem.

A strong case solver asks: "What is causing this?" A systems thinker goes further: "What relationships are producing this outcome, and what will happen when we intervene?" This shift matters because a single bad decision, a weak employee, or an isolated event doesn't cause many recurring business problems. They are produced by:

incentives;
information flows;
delays;
capacity constraints;
feedback loops;
competing objectives;
patterns of behaviour.

A useful systems-thinking progression is: Symptom → Pattern → Relationships → System Structure → Leverage Point.

What Is a System?

A system is a group of interconnected elements that produce outcomes over time. A system contains three basic components.

Elements

Elements are the people, resources, activities, technologies, organisations, or variables within the system. In a restaurant, elements might include:

customers;
servers;
kitchen employees;
suppliers;
ingredients;
tables;
technology;
pricing;
managers;
delivery partners.

Relationships

Relationships explain how the elements influence one another. Examples include:

staffing affects service capacity;
service capacity affects wait times;
wait times affect satisfaction;
satisfaction affects repeat visits;
repeat visits affect revenue;
revenue affects the ability to hire.

Simply listing the elements doesn't explain the system. The insight lies in the connections.

Purpose

Purpose is the outcome the system is organised to produce. The formal purpose of a restaurant might be to deliver enjoyable meals profitably. However, the system's actual design may produce a different outcome. If employees are rewarded only for reducing labour hours, the system may behave as though its purpose is to minimise short-term staffing cost even when service and revenue suffer. To understand a system, ask:

What is the stated purpose?
What outcome does the system actually produce?
What behaviour do its measures and incentives encourage?
Whose objectives shape the system?
Are different parts of the system pursuing conflicting outcomes?

Systems often produce exactly what their structures, measures, and incentives encourage even when that result differs from the organisation's stated objective.

Systems Have Boundaries

A system boundary determines what is included in the analysis. If the boundary is too narrow, the team may miss important causes or consequences. If the boundary is too broad, the analysis is impossible to manage. Consider a delivery-cost problem. A narrow boundary might include only:

route distance;
fuel cost;
driver hours;
vehicle use.

A broader boundary might also include:

delivery speed;
customer satisfaction;
retention;
order density;
failed deliveries;
inventory location;
customer promises;
partner performance.

The broader boundary may reveal that delivery cost cannot be managed effectively without considering customer demand and fulfilment design.

Boundary Questions

Ask:

What outcome are we trying to explain?
What factors directly influence that outcome?
What important consequences does it create?
Which stakeholders shape or experience the system?
What time horizon matters?
Which external forces significantly affect it?
What can the organisation influence?
What must be treated as an external condition?

The boundary should be broad enough to capture the important relationships but focused enough to remain useful.

Events, Patterns, Structures, and Mental Models

Systems thinkers examine problems at several levels.

Events: What Happened?

Events are visible occurrences. Examples include:

sales declined this quarter;
a shipment arrived late;
employee turnover increased;
a customer complained;
a project exceeded its budget.

Events are important, but reacting only to them can lead to short-term fixes.

Patterns: What Has Been Happening Over Time?

Patterns reveal whether the event is isolated or recurring. Ask:

Is the problem increasing or decreasing?
When did the pattern begin?
Does it affect particular customers, products, locations, or teams?
Does it repeat at predictable times?
What other variables move with it?

For example, customer complaints rise whenever order volume exceeds a certain threshold. This suggests a capacity-related pattern.

System Structures: What Relationships Produce the Pattern?

Structure includes:

processes;
information flows;
incentives;
capacity;
policies;
decision rights;
delays;
feedback loops.

For example: Marketing promotions increase order volume, but operations receives little advance notice. Capacity becomes overloaded, delivery slows, and complaints rise. The structure, not the individual promotion, produces the recurring pattern.

Mental Models: What Beliefs Maintain the Structure?

Mental models are the assumptions and beliefs that shape decisions. Examples include:

lower staffing always improves profitability;
customers care primarily about price;
managers must approve every important decision;
failure should be avoided;
inventory is always wasteful;
more sales are always beneficial.

These beliefs may influence:

policies;
incentives;
resource allocation;
behaviour.

A deep systems analysis may therefore ask what belief causes the organisation to keep designing the system this way?

Systems Thinking and Root-Cause Analysis

Systems thinking and root-cause analysis are related, but systems thinking goes further. Root cause analysis often asks, "What caused this problem?" Systems thinking asks: What connected set of conditions repeatedly produces and reinforces this problem? A problem may not have one root cause. For example, employee turnover may be influenced by:

workload;
manager behaviour;
compensation;
scheduling;
career development;
hiring fit;
training;
labour-market conditions;
turnover itself.

High turnover may increase the workload for the remaining employees, which in turn creates further turnover. The problem is therefore produced by a reinforcing system, not by a single isolated cause.

From Symptoms to Root Structure

Consider declining profitability. Visible symptoms might include:

lower revenue;
rising cost;
declining margin;
increasing customer churn;
reduced cash flow.

These outcomes could result from deeper relationships. A useful questioning sequence is:

What is happening?
Where and when is it happening?
What pattern exists over time?
What variables influence the pattern?
How do those variables affect one another?
What policies, processes, or incentives create those relationships?
What assumptions keep those structures in place?
What happens if nothing changes?
Where could an intervention alter the pattern?

Use Why" Carefully

Asking"why" repeatedly can reveal deeper causes. For example:

· Why are customers leaving? Because delivery is unreliable.

· Why is delivery unreliable? Because orders are frequently dispatched late.

· Why are orders dispatched late? Because fulfilment cannot handle peak volume.

· Why is peak volume unexpected? Because promotions are not included in operational forecasting.

· Why are promotions not included? Because marketing and operations plan independently.

The problem is not simply delivery performance. It is weak coordination and information flow. However, repeatedly asking"why" can create an overly simple straight line. Complex problems may have:

multiple causes;
circular relationships;
delays;
factors that reinforce one another.

Systems thinking adds those connections back into the analysis.

Cause-and-Effect Relationships

A simple systems map begins with variables and arrows. For example: Order volume → Fulfilment workload → Delivery time → Customer complaints. The arrows indicate influence, but a stronger map also identifies the direction of the relationship.

Same-Direction Relationship

When one variable increases, the other tends to increase, or when one decreases, the other tends to decrease. For example: Order volume increases → Fulfilment workload increases

Opposite-Direction Relationship

When one variable increases, the other tends to decrease. For example: Delivery time increases → Customer satisfaction decreases. These relationships don't have to be perfectly linear. The purpose is to show the direction of influence.

Feedback Loops

A feedback loop occurs when a change moves through the system and eventually influences the original variable. There are two main types.

Reinforcing Loops

A reinforcing loop amplifies change. It can create growth or decline.

Positive Reinforcing Loop

Better service → greater customer satisfaction → stronger referrals → more customers → more revenue → greater ability to invest in service → better service. This creates a cycle of improvement.

Negative Reinforcing Loop

Understaffing → greater workload → employee burnout → higher turnover → fewer employees → greater understaffing. This creates a downward spiral. "Reinforcing" doesn't mean good. It means the loop strengthens the direction of change.

Balancing Loops

A balancing loop pushes the system toward a limit, target, or equilibrium. For example: Higher demand → longer wait times → lower customer satisfaction → reduced demand. The system's capacity limits growth. Another example: Inventory falls below target → the organisation places more orders → inventory rises → the need for new orders falls. Balancing loops can stabilise a system, but they can also prevent desired growth.

A Simple Loop Example

Consider a restaurant that becomes popular.

· Reinforcing Growth Loop

o Positive reviews → more customers → greater revenue → more marketing and menu investment → stronger customer interest → more positive reviews

· Balancing Capacity Loop

o More customers → longer wait times → lower satisfaction → weaker reviews → fewer customers.

Both loops operate simultaneously. The restaurant's growth strategy will fail if it focuses only on demand and ignores capacity.

Delays Matter

A delay occurs when the effect of a decision appears later. Delays can make systems difficult to understand because decision-makers may:

stop an effective intervention too early;
continue a harmful intervention too long;
overreact;
incorrectly attribute the outcome to another cause.

Examples include:

training takes time to affect employee performance;
brand investment takes time to influence customer preference;
hiring takes time to improve capacity;
cost reductions may take months to improve service;
customer dissatisfaction may take time to appear in retention data;
environmental damage may emerge years after a decision.

Example: Hiring Delay

Demand increases → workload rises → service quality declines → management approves hiring → recruitment and training take several months → capacity eventually increases. If demand continues rising during the delay, service may deteriorate before the new employees become productive. Ask:

How long will the effect take?
What happens during the delay?
What temporary support is required?
Which early indicators should be monitored?
Could leaders misinterpret the lack of immediate results?
Could the organisation overcorrect?

A timeline that ignores delays is not a credible implementation plan.

Stocks and Flows

Some business outcomes accumulate over time. A stock is something that rises or falls. Examples include:

cash;
inventory;
employees;
customer relationships;
reputation;
organisational knowledge;
trust;
technical debt.

A flow increases or decreases the stock. Examples include:

Stock	Inflow	Outflow
Employees	Hiring	Turnover
Customers	Acquisition	Churn
Cash	Receipts	Payments
Inventory	Purchasing or production	Sales, use, or waste
Knowledge	Learning and hiring	Turnover and obsolescence
Trust	Reliable performance	Broken promises

This distinction can reveal why some problems persist. For example, increasing customer acquisition may not grow the customer base if churn is equally high. Customer stock = New customers entering − Existing customers leaving. The organisation may be trying to fill a leaking bucket. A marketing recommendation that increases acquisition without addressing churn may create activity without sustainable growth.

Common System Patterns

Certain patterns appear repeatedly in business cases.

1. Fixes That Fail

A short-term solution reduces the visible problem but creates delayed consequences that make the problem return or worsen. Example: Cut employee training → reduce cost immediately → increase errors over time → increase rework and complaints → raise total cost. Ask:

What improves immediately?
What consequence appears later?
Could the solution increase the original problem?

2. Shifting the Burden

The organisation relies on a quick symptomatic solution instead of addressing the underlying cause. Example: Service problems increase → hire more complaint-handling staff → complaints are managed faster → less pressure to correct product and delivery failures → underlying problems continue. The short-term response may be necessary, but it should not replace the structural solution.

3. Limits to Growth

An activity produces growth until a constraint begins to limit performance. Example: Marketing increases demand → orders grow → production capacity becomes constrained → delays and defects increase → customer satisfaction falls → growth slows. The solution may be to remove the constraint rather than further increase marketing.

4. Success to the Successful

Resources are increasingly directed toward areas that are already performing well, causing other areas to fall further behind. Example: One regional office performs well → receives more investment → performance improves further → weaker offices receive → fewer resources → their performance declines. This pattern may be efficient, but it can also create long-term dependence or inequality.

5. Growth and Underinvestment

Demand grows, but the organisation delays investing in capacity. Service deteriorates, causing demand to fall and reinforcing the belief that investment is unnecessary. Example: Demand rises → capacity pressure increases → service declines → customers leave → leaders conclude demand was temporary → capacity remains constrained. Recognising the pattern can change the Recommendation.

Unintended Consequences

An unintended consequence is an outcome that was not part of the original objective. Consider a company introducing aggressive individual sales targets. The intended result is higher sales. Possible unintended consequences include:

employees competing rather than collaborating;
unsuitable products being sold;
customer trust declining;
employees manipulating data;
service quality being ignored;
high performers burning out.

The target changes behaviour across the system.

The And Then What" Test

For every major Recommendation, ask:

What happens immediately?
Then what?
How will customers respond?
How will employees respond?
How will competitors respond?
How will suppliers or partners respond?
What happens to cost, capacity, quality, and cash?
What new incentive does the Recommendation create?
What behaviour might people use to adapt?
What appears after a delay?

Continue until the important second- and third-order consequences are visible.

Local Optimisation Versus System Performance

A department can improve its own performance while damaging the organisation. Examples include:

procurement selects the lowest-cost supplier, but quality failures increase;
marketing generates more leads than sales can handle;
sales promise customisation that operations cannot deliver;
operations maximise production volume, but inventory accumulates;
finance delays payments, damaging supplier relationships;
customer service resolves calls quickly but doesn't solve customer problems;
HR reduces training cost, increasing operational errors.

These are examples of local optimisation. Each team improves its own measure while weakening the system as a whole. Ask:

Which metric is each department optimising?
Does that metric support the overall objective?
What cost or problem is transferred elsewhere?
Are incentives aligned across teams?
What measure would encourage whole-system performance?

A strong recommendation improves system outcomes rather than one departmental metric.

Leverage Points

A leverage point is a place in the system where a focused change can create a significantly wider effect. Not all interventions have equal leverage.

Lower-Leverage Interventions

These may include:

changing a number;
adding temporary capacity;
adjusting a budget;
treating a visible symptom.

They can still be necessary, especially during a crisis.

Higher-Leverage Interventions

These may include changing:

information flows;
decision rights;
incentives;
feedback;
rules;
goals;
relationships;
capacity constraints;
the assumptions shaping the system.

Consider the understaffing cycle: Understaffing → workload → burnout → turnover → understaffing. Possible interventions include:

hiring more employees;
reducing unnecessary work;
improving scheduling;
redesigning roles;
changing manager behaviour;
increasing employee control;
improving workload data;
adjusting growth expectations.

Hiring may relieve the symptom. Redesigning workload and retention conditions may produce a more durable change.

Finding a Leverage Point

Ask:

Which variable influences several others?
Where does the problem repeatedly begin?
Which delay prevents effective response?
What information is missing?
Which incentive drives the wrong behaviour?
What capacity constraint limits the system?
Which assumption shapes several decisions?
What intervention could interrupt a harmful loop?
What intervention could strengthen a positive loop?
Can the change be implemented realistically?

The most powerful theoretical leverage point is not always the best Recommendation. It must also be:

feasible;
affordable;
timely;
ethical;
within the organisation's influence.

Leverage Versus Impact

A large initiative is not automatically a high-leverage initiative. For example:

Hiring 100 customer-service employees is large;
Correcting the billing error that generates 40% of complaints may be high-leverage.

Similarly:

Replacing an entire technology system is large;
Changing how customer data moves between two existing systems may be high-leverage.

Ask what the smallest credible intervention is that can change the system's behaviour. This question can produce more focused recommendations.

Deciphering Case Characteristics

Systems thinking becomes particularly important when:

the problem has several interacting causes;
multiple stakeholders are involved;
the same problem keeps returning;
changing one variable affects several others;
short-term and long-term outcomes differ;
incentives shape behaviour;
significant delays exist;
capacity constrains growth;
the organisation operates in a complex environment;
sustainability is central;
stakeholder outcomes conflict;
implementation may create downstream consequences.

Turnaround Cases

Look for reinforcing decline: Lower performance → cost reductions → weaker capability → lower service or quality → further performance decline. Ask where the downward loop can be interrupted without destroying the organisation's future capability.

Growth Cases

Look for limits to growth: Demand growth → capacity pressure → declining quality or service → lower retention → slower growth. Ask which constraint must be removed to create additional demand.

Organisational Cases

Look for relationships among:

workload;
management behaviour;
incentives;
trust;
turnover;
capability;
performance.

Sustainability Cases

Consider:

resource use;
environmental effects;
regulation;
community outcomes;
customer behaviour;
investment;
long-term delays.

A decision that appears profitable over one year may create substantial cost or risk over ten years.

Platform and Network Cases

Look for reinforcing adoption: More users → more value for participants → greater adoption → more users. Also examine balancing constraints such as:

trust;
quality;
congestion;
moderation;
regulation.

A simple problem may not require a complex systems map. Use systems thinking when the relationships materially affect the recommendation.

A Worked Example

A delivery company is experiencing declining profitability. Management believes delivery cost is too high and proposes reducing the number of routes.

Step 1: Define the Visible Problem

Delivery cost per order is rising;
and profitability is declining.

Step 2: Identify the Initial Solution

Reduce the number of routes and consolidate deliveries.

Step 3: Expand the System Boundary

The team examines:

delivery cost;
route density;
delivery time;
failed deliveries;
customer satisfaction;
churn;
order volume;
service promises;
inventory location;
driver workload;
revenue.

Step 4: Map the Immediate Relationship

Fewer routes → greater route efficiency → lower delivery cost. This is the intended effect.

Step 5: Identify Downstream Effects

Fewer routes → longer delivery windows → more missed deliveries → greater customer frustration → more service contacts and refunds → higher cost and lower retention. The proposal reduces one cost while increasing others.

Step 6: Identify the Reinforcing Decline

Lower retention → lower order density → higher delivery cost per order → pressure to reduce routes → longer delivery times → lower retention. The cost-reduction strategy strengthens the underlying problem.

Step 7: Identify the Root Structure

Further analysis shows:

the company promises narrow delivery windows across a large service area;
inventory is held in one central location;
customer demand is spread unevenly;
failed deliveries are frequent;
customers receive little opportunity to adjust delivery times.

The problem is not simply too many routes. The delivery model doesn't create enough order density or first-attempt success to support the service promise efficiently.

Step 8: Identify Possible Leverage Points

The team considers:

zone-based delivery days;
customer-selected delivery windows;
route-density pricing;
pickup partnerships;
improved delivery notifications;
incentives for consolidated windows;
a review of the service area.

The team recommends:

divide the service area into delivery zones;
offer preferred delivery days by zone;
introduce customer-selected windows;
provide real-time delivery notifications;
establish pickup partners in low-density areas;
test pricing that reflects delivery complexity; and
exit locations where the model cannot become economically viable.

Step 10: Measure the Whole System

The team measures:

cost per successful delivery;
first-attempt delivery rate;
route density;
on-time delivery;
customer satisfaction;
service contacts;
refunds;
retention;
contribution margin.

The Recommendation improves the delivery system rather than one isolated cost measure.

Systems Thinking Is Not the Recommendation

A systems map doesn't determine what the organisation should do. It helps the team understand where and how to intervene. A complete process is:

define the problem or outcome;
establish the system boundary;
identify the relevant elements and stakeholders;
examine patterns over time;
map important cause-and-effect relationships;
identify reinforcing and balancing loops;
recognise delays;
distinguish symptoms from underlying structures;
identify possible leverage points;
generate interventions;
test unintended consequences;
evaluate feasibility, cost, risk, and stakeholder impact;
select the strongest Recommendation; and
monitor the system after implementation.

Systems thinking improves the diagnosis and intervention design. It doesn't replace financial analysis, strategic choice, stakeholder analysis, or implementation planning.

Testing a Recommendation as a System Intervention

Before finalising a recommendation, ask six questions.

· What Part of the System Are We Changing? Be specific about the variable, relationship, policy, process, information flow, or incentive.

· What Immediate Outcome Do We Expect? Identify the first-order effect.

· What Happens Next? Trace the second- and third-order consequences.

· Which Feedback Loop Will Change? Will the Recommendation:

o interrupt a harmful reinforcing loop;

o strengthen a beneficial reinforcing loop;

o remove a constraint;

o create a balancing response?

· What Delay Should We Expect? Determine when the result should appear and what happens in the meantime.

· What Should We Monitor? Select measures across the system, not only the target metric. For example, a cost-reduction initiative might monitor:

o cost;

o quality;

o employee workload;

o customer experience;

o revenue;

o risk;

o stakeholder outcomes.

Winning the Room: Presenting Systems Thinking

A causal map containing 30 variables and dozens of arrows will rarely help judges understand the Recommendation. The objective is clarity, not complexity.

Lead with the System Insight

For example: The company's delivery-cost problem is being reinforced by declining route density, not simply by the number of routes.

Show the Critical Loop

Present only the relationship that shapes the decision: Lower retention → lower route density → higher cost per delivery → fewer routes → slower service → lower retention.

Identify the Leverage Point

Explain: Reducing routes strengthens the cycle. Improving delivery density and first-attempt success interrupts it.

Connect the Insight to the Recommendation

Then present the intervention: We recommend zone-based delivery, customer-selected windows, pickup partnerships, and a narrower service area.

Show the Balanced Measures

Demonstrate that the team will monitor:

delivery cost;
service reliability;
customer retention;
contribution margin.

The analytical chain becomes: Visible Problem → Reinforcing Structure → Leverage Point → System Intervention.

Coach's Lens

One of the most common mistakes I see in case competitions is teams identifying several problems and then assigning one solution to each.

Customers are leaving → improve customer service.
Costs are increasing → reduce costs.
Employee morale is declining → improve engagement.

These may not be three independent problems. They may be three outcomes created by one connected system: Cost pressure
→ understaffing → workload → service decline → customer churn → revenue decline → greater cost pressure. Your job is not always to solve every symptom separately. Your job is to understand the system well enough to identify the leverage point at which a focused intervention can have the greatest positive effect. I often ask teams, if we implement your Recommendation successfully, what new problem might it create? If the team has never considered that question, it has not finished the analysis.

Common Mistakes

· Treating Problems as Independent: Connected problems may be symptoms of the same system. Map how the problems influence one another.

· Confusing Symptoms with Causes: The most visible problem may appear far downstream from its source. Trace the pattern through processes, incentives, information, and capacity.

· Looking for One Root Cause: Complex problems often have several interacting causes. Examine loops and relationships rather than forcing a single linear explanation.

· Drawing Straight Lines Only: A causal chain may overlook feedback. Ask whether the consequence ultimately affects the original cause.

· Ignoring Delays: The team expects an immediate result or misreads a delayed consequence. Identify when the effects should appear and what happens during the delay.

· Ignoring Unintended Consequences: A solution improves one metric while damaging another. Apply the"a nd then what" test.

· Focusing Only on Immediate Results: Short-term improvement may create long-term damage; analyse first-, second-, and third-order effects.

· Optimising One Department: A local improvement transfers cost or risk elsewhere. Measure whole-system performance.

· Overcomplicating the Map: A diagram with too many variables hides the insight. Include only the relationships needed to explain the problem and intervention.

· Making Every Arrow Causal: Two variables moving together doesn't prove one causes the other. Support important relationships with evidence and reasoning.

· Ignoring the System Boundary: A narrow boundary excludes important stakeholders or consequences. State what is included and excluded, and explain why.

· Forgetting the Human Element: Employees, customers, communities, and partners adapt to interventions. Consider behaviour, incentives, trust, workload, and stakeholder responses.

· Choosing an Impractical Leverage Point: A theoretically powerful intervention may be outside theorganisation'ss control. Balance potential impact against feasibility, authority, time, cost, and ethics.

· Treating the Diagram as the Recommendation: Understanding the system doesn't automatically identify the best action. Generate and evaluate alternative interventions.

MAD Skills Drill

Choose a problem from a previous case.

Step 1: Define the Visible Problem

Write the problem as a variable that can increase or decrease. For example:

customer retention;
employee turnover;
delivery time;
profitability;
service quality.

Step 2: Define the Boundary

Identify:

the time horizon;
stakeholders;
organisational functions;
external forces;
outcomes included in the analysis.

Step 3: Identify the Pattern

Describe how the central variable has changed over time. Ask:

When did the change begin?
Where is it strongest?
What else changed?
Does the pattern repeat?

Step 4: Identify Causes and Consequences

Add:

three likely causes;
three consequences;
two relationships among the causes.

Step 5: Build a Causal Map

Connect the variables using directional arrows. For every connection, explain:

why the relationship exists;
whether the variables move in the same or opposite directions;
what evidence supports it.

Step 6: Find a Feedback Loop

Identify at least one reinforcing or balancing loop. Explain how the loop affects the problem over time.

Step 7: Identify a Delay

Find one relationship where the consequence will not appear immediately. Explain how the delay could affect decision-making.

Step 8: Generate Interventions

Identify at least three possible places to intervene. Include:

one symptom-level response;
one structural response;
one higher-leverage response.

Step 9: Test Unintended Consequences

For each intervention, identify:

the immediate benefit;
one possible negative consequence;
one stakeholder response;
one measure that could reveal a problem.

Step 10: Select the Leverage Point

Choose the intervention that creates the strongest balance of:

system impact;
feasibility;
time;
cost;
stakeholder acceptance;
risk.

Step 11: Deliver the Insight

Prepare a 60-second explanation answering:

What is the visible problem?
What system structure keeps producing it?
Which feedback loop matters most?
Where is the leverage point?
What intervention should the organisation make?
What unintended consequence must be monitored?

Don't explain the entire system. Explain the relationship that changes the Recommendation.

Reflection Questions

Did your team solve a symptom rather than an underlying structure?
What pattern over time did you overlook?
Which variables reinforced one another?
Was there a delay between the cause and the visible outcome?
What unintended consequence could your Recommendation have created?
Did one department benefit while another absorbed the cost?
Which stakeholders were part of the system?
What relationship did your team initially overlook?
Which assumption or mental model maintained the problem?
Where was the highest-leverage feasible intervention?

Chapter Summary

Systems thinking helps case teams understand how relationships among people, processes, resources, incentives, information, capacity, and stakeholders create outcomes over time. A system contains:

elements;
relationships;
a purpose.

Its behaviour may be shaped by:

reinforcing loops;
balancing loops;
delays;
accumulated stocks;
information flows;
incentives;
constraints;
underlying assumptions.

Strong systems thinking follows this progression: Visible Symptom → Pattern → Relationships → System Structure → Leverage Point → Intervention. The goal is not to understand every variable. It is to understand enough of the system to explain:

why the problem persists;
why an obvious solution may fail;
where a focused intervention could create wider impact;
what consequences should be monitored.

A weak case solution treats every problem independently. A strong case solution identifies the underlying structure that produces those problems.

Key Takeaways

✓ Organisations are systems of interconnected elements, relationships, and purposes.

✓ The insight comes from understanding relationships, not simply listing system elements.

✓ Define an appropriate system boundary before beginning the analysis.

✓ Examine events, patterns, system structures, and the beliefs or assumptions maintaining those structures.

✓ A recurring problem may be produced by several interacting causes rather than one root cause.

✓ Reinforcing loops amplify growth or decline.

✓ Balancing loops resist change or push the system toward a limit or target.

✓ Delays can cause leaders to stop effective interventions too early, continue harmful interventions too long, or overreact.

✓ Stocks such as customers, employees, cash, knowledge, and trust accumulate through inflows and decline through outflows.

✓ Recognise common patterns such as fixes that fail, shifting the burden, limits to growth, and growth with underinvestment.

✓ Ask "and then what" to identify second- and third-order consequences.

✓ A department can improve its own metric while damaging whole-system performance.

✓ A leverage point is a place where a focused change can produce a wider system effect.

✓ The largest intervention is not always the highest-leverage intervention.

✓ Balance leverage with feasibility, authority, time, cost, ethics, and stakeholder impact.

✓ Use systems thinking to improve the diagnosis and intervention, not as a substitute for evaluating alternatives.

✓ In the presentation, show the critical loop and leverage point rather than a map of the entire system.

Looking Ahead

Systems Thinking helps case solvers understand how the parts of a problem connect and how interventions create consequences across the organisation. The next chapter introduces Analytical Thinking, the ability to break complex problems into manageable components, structure the analysis, identify patterns in evidence, and determine what the information is actually telling you.