Chapter 8: Decision-Making Under Uncertainty - Making Defensible Choices When the Future Is Unclear

~~Chapter 8: Decision-Making Under Uncertainty - Making Defensible Choices When the Future Is Unclear~~

Video: Decision Trees Visually Show Your Strategic Choices and Why You Chose Them

Learning Objectives

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

distinguish risk, Uncertainty, ambiguity, and incomplete information;
identify the assumptions on which a recommendation depends;
distinguish assumptions from risks;
prioritise uncertainties according to importance and confidence;
evaluate multiple possible outcomes;
use sensitivity analysis to identify the variables that matter most;
construct coherent scenarios rather than relying on one forecast;
use decision trees when decisions and outcomes occur in stages;
calculate and interpret expected Value when reasonable probabilities are available;
understand the limitations of expected-value analysis;
evaluate the Value of obtaining additional information;
use pilots, phases, partnerships, and other real options to manage uncertainty;
identify no-regret actions, hedges, signposts, triggers, and contingencies;
distinguish an optimal strategy for one forecast from a resilient strategy across several futures;
make recommendations that remain credible when important assumptions change.

Why This Matters

Case competitions rarely provide perfect information. You may not know:

exactly how customers will respond;
how competitors will react;
whether the market will grow;
whether the organisation can build the required capability;
how quickly a partner will perform;
whether regulation will change;
whether implementation will proceed as planned;
whether the financial assumptions will hold.

Yet you must still make a recommendation. That is not a flaw in the case; it reflects the business reality. Leaders make decisions before they possess complete information. They commit resources while customer demand, technology, competitors, costs, regulation, and stakeholder behaviour remain uncertain. The objective is not to eliminate Uncertainty. It is to:

understand it;
prioritise it;
measure it where possible;
reduce it where valuable;
protect the organisation from unacceptable downside;
design a strategy that adapts as new information emerges.

A recommendation is not credible because it predicts the future perfectly. It is credible because it remains sensible or can change intelligently when the future differs from the forecast.

Discover Your MAD Skills Principle

You don't need certainty to make a decision. You need a defensible decision and a plan for managing uncertainty.

Students often believe confidence means presenting one answer without hesitation. Experienced decision-makers understand that confidence comes from knowing:

what is supported;
what is assumed;
what could change;
what would happen if the assumption is wrong;
how the organisation will respond.

A useful decision-making progression is: What We Know → What We Assume → What Could Change → What it Would Mean → How We Will Respond. The strongest recommendation doesn't say: "We know exactly what will happen." It says: "Based on the available evidence, this is the strongest decision and this is how we will manage what remains uncertain."

Risk, Uncertainty, Ambiguity, and Ignorance

These concepts are related but different.

Risk

Risk exists when the team can identify potential outcomes and reasonably estimate their likelihood. For example: Historical data suggest that 8% of shipments arrive late. The exact Outcome remains unknown, but the probability can be estimated. Risk can often be analysed using:

probabilities;
expected values;
ranges;
insurance;
reserves;
controls;
mitigation plans.

Uncertainty

Uncertainty exists when possible outcomes can be identified, but their probabilities cannot be estimated with confidence. For example, the company is entering a new market with limited evidence of customer adoption. The team may create scenarios, but attaching precise probabilities could create false confidence.

Ambiguity

Ambiguity exists when the problem, evidence, or meaning is unclear. For example: Sales are declining, but it is unclear whether the central problem is price, customer experience, distribution, or changing demand. Before evaluating possible outcomes, the team must clarify what problem it is solving.

Unknown Unknowns

Some events or consequences may not be identified in advance. Teams cannot model every possibility. They can, however, improve resilience through:

flexibility;
reserves;
monitoring;
diversified dependencies;
rapid feedback;
decision-making processes capable of adapting.

Incomplete Information

Information may exist but be unavailable within the case or competition period. For example:

customer willingness to pay;
partner capacity;
competitor cost structure;
final regulatory guidance.

The team must decide whether to:

research;
estimate;
use a range;
test;
defer;
proceed with safeguards.

Uncertainty Is Not the Same as Risk

The distinction matters because the analytical response should differ.

Condition	What is known?	Useful responses
Risk	Outcomes and approximate probabilities	Expected Value, insurance, reserves, mitigation
Uncertainty	Outcomes may be known, probabilities are weak	Scenarios, stages, flexibility, signposts
Ambiguity	The problem or meaning is unclear	Research, reframing, diagnosis, hypotheses
Unknown unknowns	Important events may not be anticipated	Resilience, buffers, monitoring, adaptability
Incomplete information	Relevant evidence is missing	Estimation, testing, ranges, information gathering

Don't assign precise probabilities when the evidence doesn't support them.

Types of Uncertainty

Uncertainty can enter a recommendation from several sources.

Customer Uncertainty

Will customers adopt the offering?
What will they pay?
How frequently will they use it?
Will they remain?
Which segment will respond?

Market Uncertainty

How quickly will the market grow?
How large is the addressable segment?
Will customer preferences change?
Will a substitute become more attractive?

Competitive Uncertainty

Will competitors reduce prices?
Will they copy the offering?
Will a new entrant appear?
Will a partner become a competitor?

Technical Uncertainty

Will the technology perform?
Can it integrate with existing systems?
Will it scale?
What cybersecurity or reliability problems may emerge?

Operational Uncertainty

Can the organisation deliver the required quality and volume?
Will suppliers perform?
Can employees adopt the process?
Will implementation remain on schedule?

Financial Uncertainty

Will cost, price, volume, margin, and investment match the forecast?
How will inflation, interest rates, or exchange rates change?
Is enough working capital available?

Regulatory and Political Uncertainty

Will approval be granted?
Will the rules change?
How long will compliance take?
Will public funding or policy support continue?

Stakeholder Uncertainty

Will employees support the change?
Will a community accept the project?
Will investors provide funding?
Will partners maintain their commitment?

Environmental and Social Uncertainty

Could extreme weather disrupt operations?
Will environmental expectations change?
Could the decision create an emerging social or reputational risk?

Identifying the type of Uncertainty helps determine the appropriate response.

Uncertainty Depends on the Decision

Not every case contains the same level of uncertainty.

Lower-Uncertainty Decision

Replacing proven equipment in an established process may have:

known costs;
measurable productivity effects;
experienced suppliers;
reliable implementation data.

The team may focus on:

financial risk;
implementation schedule;
operational controls.

Higher-Uncertainty Decision

Launching a new product in an emerging international market may involve Uncertainty about:

demand;
willingness to pay;
customer behaviour;
regulation;
competition;
distribution;
culture;
cost;
organisational capability.

The team should consider:

scenarios;
pilots;
stages;
partnerships;
decision triggers.

The level of Uncertainty should influence:

the depth of analysis;
size of the initial commitment;
reversibility of the strategy;
implementation design.

The Assumption Test

Every recommendation contains assumptions. The objective is not to eliminate all assumptions; it is to identify the assumptions that matter most. Examples include:

10% of target customers will adopt the service;
customers will pay $25 per month;
customer-acquisition cost will remain below $100;
the supplier can increase capacity;
regulatory approval will take six months;
employees can be trained within three months;
the partner will meet the required service level.

For each assumption, ask:

What must be true?
What evidence supports it?
How confident are we?
What happens if it is wrong?
Can we test it?
Can we reduce our exposure?
What result would cause us to change direction?

Assumptions Are Not Risks

An assumption is something the recommendation accepts as true. A risk is a possible event or condition that could affect the recommendation.

· Assumption: Customers will adopt the service at a rate of 10%.

· Response: Pilot the service and expand only if adoption and retention meet predefined thresholds.

This distinction creates a clearer implementation plan.

Prioritising Assumptions

Teams don't have time to test every assumption. Prioritise assumptions using two dimensions:

· Importance: How significantly would the assumption affect the recommendation if it were wrong?

· Uncertainty: How weak or incomplete is the evidence?

	Lower Uncertainty	Higher Uncertainty
Higher importance	Monitor carefully	Test first
Lower importance	Accept provisionally	Monitor if inexpensive

The most dangerous assumptions are both:

important; and
uncertain.

These are the critical assumptions.

Example

A subscription model may contain these assumptions:

Assumption	Importance	Confidence	Priority
Customers will pay $25 monthly	High	Low	Test immediately
Technology can support recurring billing	Medium	High	Confirm
Customer retention will exceed 75%	High	Low	Test immediately
Marketing design will be completed on time	Low	Medium	Monitor
Existing service employees can manage inquiries	Medium	Medium	Test during pilot

A strong team doesn't merely say: "The strategy contains risk." It identifies which assumptions create the risk.

Confidence Should Match Evidence

Teams should calibrate their language.

Strong Evidence

The evidence demonstrates…
Historical data show…
The result is consistent across…

Moderate Evidence

The evidence suggests…
The pattern indicates…
The most likely explanation is…

Weak or Emerging Evidence

We hypothesise…
Early evidence suggests…
The recommendation assumes…
This should be tested before scaling…

Acknowledging Uncertainty doesn't weaken the recommendation. Overstating confidence does.

Sensitivity Analysis

Sensitivity analysis tests how the Outcome changes when one assumption changes. It asks which variables influence the decision most. Suppose project value depends on:

customer adoption;
price;
customer-acquisition cost;
retention;
variable cost;
initial investment.

The team can vary each assumption while holding the others constant.

Example

Variable tested	Lower case	Base case	Higher case	Effect on break-even
Customer adoption	6%	10%	14%	Very high
Monthly price	$20	$25	$30	High
Acquisition cost	$140	$100	$70	High
Variable cost	$14	$12	$10	Moderate
Initial investment	$2.4M	$2.0M	$1.8M	Lower

The analysis reveals that adoption, price, and acquisition cost matter most. The team should therefore focus its:

research;
pilot;
monitoring;
contingency planning

on those variables.

Break-Even Sensitivity

A useful question is: At what Value does the recommendation cease to be attractive? Examples include:

minimum adoption;
maximum acquisition cost;
minimum price;
maximum implementation delay;
minimum retention;
maximum investment.

This threshold becomes an implementation trigger.

Sensitivity Analysis Is Not Scenario Analysis

The two tools answer different questions.

Sensitivity Analysis

Changes one variable at a time. It helps identify which assumption has the greatest effect.

Scenario Analysis

Changes several related assumptions together to create a coherent possible future. It helps examine how the strategy performs under different conditions. For example, weak economic conditions may simultaneously produce:

lower customer demand;
greater price sensitivity;
higher financing costs;
and slower adoption.

Changing only one variable would not capture the complete situation.

Scenario Analysis

Scenario Analysis examines several plausible futures. The objective is not to predict which scenario will occur. It is to understand:

how the strategy performs;
what breaks;
which actions remain valuable;
and what signals indicate that conditions are changing.

Base Scenario

Conditions develop approximately as expected.

Upside Scenario

Important conditions are more favourable than expected.

Downside Scenario

Important conditions are less favourable.

However, strong scenarios should be more than optimistic and pessimistic financial forecasts. They should describe internally consistent conditions.

Example: International Market Entry

Scenario 1: Rapid Adoption

strong customer demand;
moderate acquisition cost;
stable regulation;
limited competitor response;
strong partner performance.

Scenario 2: Competitive Pressure

moderate customer demand;
aggressive competitor pricing;
higher acquisition cost;
lower margins.

Scenario 3: Delayed Market Access

regulatory approval takes longer;
launch costs increase;
partner capacity is limited;
revenue begins later.

Scenario 4: Economic Contraction

demand shifts toward lower-priced options;
financing costs rise;
customer adoption slows.

These scenarios reveal different weaknesses and response requirements.

Building Useful Scenarios

A useful scenario should be:

plausible;
relevant;
internally consistent;
meaningfully different;
connected to the decision.

Step 1: Identify the Critical Uncertainties

Select two or three factors that are:

highly uncertain;
capable of materially changing the strategy.

Step 2: Define Distinct Conditions

Describe how each Uncertainty behaves within each scenario.

Step 3: Test the Strategy

Ask:

Does the recommendation remain attractive?
Which assumption fails?
What capability becomes important?
What financial effect occurs?
What should the organisation do differently?

Step 4: Identify Signposts

What evidence would indicate that the scenario is emerging?

Step 5: Create Contingencies

What action should the organisation take if the signpost appears? Scenarios become useful when they influence action.

Robust Strategies Versus Optimal Strategies

An optimal strategy may yield the highest return under a given forecast. A robust strategy performs acceptably across several plausible futures. Consider two alternatives.

· Alternative A

o very high upside;

o large investment;

o long commitment;

o significant loss if demand is weak.

· Alternative B

o lower maximum upside;

o smaller initial commitment;

o flexible expansion;

o limited downside.

Alternative A may be optimal in the base forecast. Alternative B may be more robust under uncertainty. The best choice depends on:

risk tolerance;
financial capacity;
reversibility;
strategic urgency;
the organisation's ability to absorb loss.

The highest expected return is not always the best decision.

Decision Trees

Decision trees are useful when:

decisions occur in stages;
several outcomes are possible;
one Outcome creates another decision;
probabilities can be estimated reasonably;
future flexibility has Value.

A decision tree distinguishes:

· Decision Nodes: Points where the organisation chooses among alternatives.

· Chance Nodes: Points where an uncertain outcome occurs.

· End Outcomes: The final financial or strategic result.

The structure is: Decision → Uncertain Outcome → Next decision → Final Outcome

Example

A company is considering a market-entry pilot.

Decision 1

launch a pilot; or
enter at full scale.

If it pilots, possible outcomes include:

strong demand;
moderate demand;
weak demand.

Decision 2

After the pilot, the company may:

expand;
redesign;
partner;
exit.

The Value of the pilot comes partly from the future choices it preserves.

Expected Value

When reasonable probabilities and financial outcomes are available, expected Value can help compare alternatives.

· For one uncertain Outcome: Expected Value = Probability × Outcome

· For several outcomes: Expected Value = Σ (Probability × Outcome)

Example

A pilot has two possible outcomes:

60% probability of a $5 million value;
40% probability of a $1 million loss.

Expected value: (0.60 × $5 million) + (0.40 × −$1 million) = $3 million − $0.4 million = $2.6 million

This doesn't mean the project will produce exactly $2.6 million. The actual Outcome may be $5 million or a $1 million loss. Expected Value is a weighted average across possible outcomes.

Limits of Expected Value

Expected Value is useful, but it should not decide automatically. It may overlook:

the organisation's ability to absorb a loss;
cash-flow timing;
risk tolerance;
irreversible consequences;
ethical concerns;
stakeholder impact;
strategic flexibility;
uncertainty in the probabilities themselves.

Two alternatives may have the same expected value but very different downside exposure. A small organisation may reject a positive-expected-value project if one Outcome could cause insolvency. Expected Value is one piece of evidence, not the answer.

Risk Appetite and Capacity

Risk appetite is the amount and type of risk the organisation is willing to accept. Risk capacity is the amount of risk it can absorb without threatening its survival or essential objectives. An organisation may be willing to take risk but lack the financial capacity to withstand a large loss. Ask:

How much loss can the organisation absorb?
How much cash is available?
What happens if implementation is delayed?
Could failure damage essential operations?
Is the decision reversible?
Which stakeholders bear the downside?
Does the potential return justify the exposure?

Recommendations should fit both strategic ambition and organisational capacity.

Value of Information

Additional information has Value when it could improve the decision. Examples include:

customer research;
a prototype;
technical testing;
regulatory consultation;
supplier assessment;
a market pilot;
a limited pricing experiment.

Ask:

What information is missing?
Could that information change the decision?
How much would it cost to obtain?
How long would it take?
Would waiting create an opportunity cost?
Is the information reliable?
Can the decision be staged while learning?

the choice;
design;
timing;
scale;
risk response.

Real Options Thinking

A real option preserves the right but not the obligation to take a future action. Instead of committing fully, the organisation may preserve flexibility through:

· Option to Delay: Wait until more information becomes available.

· Option to Stage: Invest in phases rather than all at once.

· Option to Expand: Begin small and increase commitment if results are strong.

· Option to Abandon: Stop if performance falls below a threshold.

· Option to Switch: Change suppliers, technology, channels, or operating models if conditions change.

· Option to Partner: Share investment, knowledge, or risk.

· Option to Pilot: Test the market or process before scaling.

These options create Value by allowing the organisation to learn before making a larger commitment.

Pilots Should Buy Information

A pilot is not simply a smaller launch. It should test the assumptions most important to the decision. A strong pilot defines:

the uncertainty being tested;
the target customer or location;
the scope;
the duration;
the resources;
the measures;
the minimum evidence required;
the expansion trigger;
the redesign trigger;
the exit condition.

For example: Launch a six-month pilot in one region to test whether adoption exceeds 8%, acquisition cost remains below $120, retention exceeds 70%, and contribution margin becomes positive by Month 6. If the pilot doesn't influence the next decision, it is not functioning as a real option.

No-Regret Moves, Hedges, and Big Bets

Actions can be grouped based on their relationship with Uncertainty.

No-Regret Moves

Actions that create Value across most plausible scenarios. Examples include:

improving data quality;
strengthening customer insight;
correcting a service failure;
building regulatory relationships;
reducing unnecessary complexity;
improving cash visibility.

Hedges

Actions that reduce exposure if conditions become unfavourable. Examples include:

diversifying suppliers;
maintaining reserves;
using flexible contracts;
preserving an alternative channel.

Options

Small investments that preserve future opportunities. Examples include:

pilots;
partnerships;
prototype development;
purchasing limited rights.

Big Bets

Large commitments justified only if particular assumptions prove correct. Examples include:

building a facility;
acquiring a company;
entering several markets simultaneously;
replacing an organisation-wide technology platform.

The implementation plan should identify which actions belong in each category.

Signposts, Triggers, and Contingencies

A resilient strategy defines what the organisation will monitor and how it will respond.

Signpost

Evidence that conditions may be changing. Examples include:

customer adoption slowing;
competitor prices falling;
regulation being delayed;
supplier performance deteriorating;
employee turnover increasing.

Trigger

A predefined threshold that causes a decision or action. Examples include:

expand if adoption exceeds 10%;
redesign pricing if acquisition cost exceeds $120;
pause implementation if error rates exceed 3%;
activate a second supplier if on-time delivery falls below 95%;
exit if contribution margin remains negative after six months.

Contingency

The action taken when the trigger occurs. A trigger without a contingency merely identifies a problem.

Decision Rules

A decision rule specifies in advance how evidence will shape the next choice. For example, if adoption is above 10% and the contribution margin is positive, expand to two additional regions. If adoption is between 6% and 10%, continue the pilot and revise the value proposition. If adoption is below 6%, stop expansion and reassess the target segment. Decision rules reduce the risk of:

emotional overreaction;
escalation of commitment;
changing criteria after results appear.

Managing Competitor and Stakeholder Response

Not all Uncertainty is passive. Competitors and stakeholders may respond to the organisation's decision. Ask:

What will competitors observe?
How quickly can they respond?
Could they reduce prices?
Could they copy the offering?
Could suppliers change terms?
Could regulators intervene?
Could employees resist?
Could customers use the new policy differently than expected?
Could a partner demand more favourable terms?

A strategy should account for adaptive behaviour, not only market forecasts.

False Precision

Uncertain inputs don't become certain because they appear in a spreadsheet. A forecast such as: Revenue in Year 3 will be $14,287,316. may imply more confidence than the evidence supports. A stronger presentation might say: Under the base assumptions, Year 3 revenue is approximately $14 million, with a reasonable range of $10 million to $17 million depending primarily on adoption and retention. Use:

ranges;
rounded numbers;
transparent assumptions;
sensitivity analysis;
scenario results.

Precision should reflect evidence quality.

A Competition Example

A company is considering entering a new international market. The initial model predicts:

$10 million in Year 3 revenue;
a 15% operating margin;
a three-year payback.

A weak team presents those numbers as the expected future. A stronger team identifies the critical assumptions:

customer adoption reaches 12%;
acquisition cost remains below $100;
regulatory approval occurs within six months;
the local partner meets service standards;
the competitor doesn't aggressively reduce prices.

Sensitivity Analysis

The team finds:

adoption below 8% causes the project to miss the return threshold;
acquisition cost above $140 delays payback beyond five years;
a six-month regulatory delay creates a significant cash requirement.

Adoption and acquisition cost are the most sensitive assumptions.

Scenario Analysis

Base Scenario

adoption reaches 12%;
acquisition cost is $100;
approval occurs on time;
the business reaches 15% margin.

Competitive Scenario

the competitor lowers prices;
acquisition cost rises to $145;
adoption reaches only 9%;
margin falls to 8%.

Delay Scenario

approval takes 12 months;
launch cost increases;
partner expenses begin before revenue;
the organisation faces a cash-flow gap.

Upside Scenario

the local partner provides lower-cost access;
adoption reaches 16%;
customer retention is strong;
payback occurs in two years.

Strategic Interpretation

The market remains attractive, but immediate full-scale entry creates substantial downside if adoption is weak or approval is delayed.

Revised Recommendation

The team recommends a phased market entry, beginning with a six-month pilot in one region through a local partner.

Stage-One Measures

customer adoption;
acquisition cost;
retention;
contribution margin;
partner service;
regulatory progress.

Expansion Triggers

Expand if:

adoption exceeds 10%;
acquisition cost remains below $120;
contribution margin is positive;
partner service exceeds 95%;
approval remains on schedule.

Redesign Trigger

Revise the value proposition, pricing, or channel if adoption is between 6% and 10%.

Exit Trigger

Stop additional investment if:

adoption remains below 6%;
approval is delayed beyond the defined limit;
the project cannot meet the minimum contribution threshold.

The Uncertainty has not disappeared. The strategy now manages it.

Decision-Making Under Uncertainty Is Not Indecision

Recognising Uncertainty doesn't mean refusing to choose. A recommendation such as "Conduct more research before deciding" may be appropriate if the missing evidence is essential and inexpensive to obtain, but it may also avoid the decision. Ask:

Can the organisation make a limited commitment?
What information is truly necessary?
What is the cost of delay?
Is the decision reversible?
Can the organisation act while learning?
What would waiting improve?
What opportunity may be lost?

The best response to Uncertainty may be:

proceed;
proceed conditionally;
pilot;
phase;
partner;
delay;
hedge;
redesign;
decline.

The team must still choose.

Winning the Room: Presenting Uncertainty Credibly

Judges don't need every scenario and assumption on the main slide. They need to understand:

what the recommendation depends on;
what could materially change it;
how resilient it is;
how implementation will adapt.

· State the Critical Assumptions: For example: Our recommendation depends on customer adoption above 10%, acquisition cost below $120, and regulatory approval within 12 months.

· Show the Most Important Sensitivity: Explain: Adoption is the most influential variable. Below 8%, the project fails to meet the return threshold.

· Present the Strategic Response: We therefore recommend a one-region pilot rather than immediate national entry.

· Define the Triggers: We will expand only if adoption, retention, contribution margin, and regulatory milestones meet the predefined thresholds.

· Explain the Downside Plan: If adoption remains below 6%, the company will stop expansion and reassess the segment and value proposition.

The analytical chain becomes: Critical Uncertainty → Financial or Strategic Impact → Staged Response → Trigger.

Coach's Lens

One of the strongest statements a team can make is: "Our recommendation depends on three critical assumptions." That statement demonstrates awareness. Even stronger is: "We designed the implementation to test those assumptions before making the full investment." Now the team is not merely acknowledging Uncertainty. It is managing it. The strongest leaders are not the ones who pretend to know the future. They know:

what they believe;
why they believe it;
what could make them wrong;
how they will respond.

Common Mistakes

· Pretending Uncertainty Doesn't Exist. The team presents one forecast as fact. Identify critical assumptions and ranges.

· Listing Risks Without Linking Them to Assumptions: The risk register becomes generic. Show which assumption underlies each risk and how the response addresses it.

· Treating Every Assumption Equally: The team tries to test everything but prioritise assumptions by importance and Uncertainty.

· Assigning Unsupported Probabilities: Precise probabilities create false confidence. Use qualitative scenarios when evidence is insufficient.

· Building Only One Scenario: The base forecast becomes the strategy. Test several plausible, internally consistent futures.

· Creating Scenarios That Differ Only Financially: The team changes revenue by ±10 %. Connect customer, competitive, regulatory, operational, and financial conditions.

· Confusing Sensitivity with Scenario Analysis: The team uses the terms interchangeably. Use sensitivity for individual variables and scenarios for connected futures.

· Overanalysing Every Possibility: The model becomes too complex to support a decision. Focus on uncertainties that could materially change the recommendation.

· Ignoring Downside Severity: The option has a positive expected value but could threaten organisational survival. Examine risk capacity, cash flow, and the worst credible outcomes.

· Relying Only on Expected Value: The weighted average hides variation and irreversible harm. Consider downside, timing, flexibility, ethics, and strategic fit.

· Using Fake Precision: Detailed forecasts imply more certainty than the inputs justify. Round values, show ranges, and identify drivers.

· Recommending More Research Without a Decision Purpose: Research becomes a delay tactic. Specify what will be learned and how that will change the decision.

· Recommending a Pilot Without Learning Objectives: The pilot is merely a small launch. Define assumptions, measures, thresholds, and next decisions.

· Using Vague Triggers: "Expand if successful" cannot guide action. Define measurable thresholds for scale, redesign, pause, and exit.

· Ignoring the Cost of Waiting: Delaying may reduce Uncertainty but lose an opportunity. Compare the Value of information to the cost of delay.

· Failing to Build Flexibility: The organisation makes a single, large, irreversible commitment. Use stages, options, flexible contracts, and partnerships.

· Treating the Downside as Failure: An unsuccessful pilot may still create valuable information. Define what the organisation expects to learn and limit the cost of learning.

· Refusing to Recommend: The team hides behind Uncertainty. Make the best defensible choice and explain how to manage the Uncertainty.

MAD Skills Drill

Choose a recommendation from a previous case.

Step 1: Define the Decision

State the recommendation and the major commitment required.

Step 2: Identify Critical Assumptions

List five assumptions. For each, assess:

importance;
confidence;
evidence;
impact if wrong.

Select the three most critical.

Step 3: Identify Related Risks

For each critical assumption, identify:

what could go wrong;
likelihood where estimable;
severity;
affected stakeholders.

Step 4: Conduct Sensitivity Analysis

Change the most important numerical assumptions. Identify:

which variable matters most;
the break-even threshold;
the Value at which the recommendation should change.

Step 5: Build Three Scenarios

Create:

a base scenario;
an upside scenario;
a downside scenario.

Ensure that the assumptions within each scenario fit together logically.

Step 6: Test Strategic Resilience

For each scenario, ask:

Does the recommendation still work?
What fails?
What changes?
Which action remains valuable?
What new risk appears?

Step 7: Identify No-Regret Moves

Select actions that create Value across all three scenarios.

Step 8: Design a Real Option

Choose whether to:

pilot;
stage;
partner;
delay;
expand conditionally;
switch;
preserve an exit.

Step 9: Define Signposts and Triggers

Identify:

the evidence to monitor;
expansion threshold;
redesign threshold;
pause threshold;
exit threshold.

Step 10: Create Contingencies

Specify the action that follows each trigger.

Step 11: Deliver the Recommendation

Prepare a 60-second explanation answering:

What is the recommended decision?
Which assumptions matter most?
Which variable creates the greatest sensitivity?
How does the recommendation perform across scenarios?
How will the organisation learn before committing fully?
What trigger will change the strategy?

Don't apologise for uncertainty; show how to manage it.

Reflection Questions

Which assumption in your last recommendation was most uncertain?
Which assumption had the greatest effect on the result?
Did the team test it?
Did your financial model contain false precision?
What happened under the downside scenario?
Could the organisation absorb that Outcome?
Did you assign probabilities that the evidence could not support?
Could implementation have been staged?
What information would have been worth obtaining before full commitment?
What was the cost of waiting?
Which signpost should the organisation monitor?
What trigger would cause you to expand, redesign, pause, or exit?

Chapter Summary

Decision-Making under Uncertainty is the ability to make responsible choices when important information about the future remains incomplete. Strong decision-making follows this progression: Decision → Critical Assumptions → Uncertainty Exposure → Scenarios & Sensitivities → Strategic Response → Triggers & Adaptation. The objective is not to eliminate Uncertainty. It is to:

identify it;
distinguish risk from deeper Uncertainty;
prioritise what matters;
measure what can be measured;
test important assumptions;
evaluate several plausible futures;
limit unacceptable downside;
preserve future choices;
adapt as evidence appears.

A weak recommendation hides Uncertainty behind one precise forecast. A strong recommendation explains why the decision is defensible and how the organisation will respond if important assumptions prove wrong.

Key Takeaways

✓ Risk involves outcomes whose probabilities can be estimated reasonably; Uncertainty involves outcomes whose probabilities are not confidently known.

✓ Ambiguity exists when the problem or meaning of the evidence is unclear.

✓ Different forms of Uncertainty require different analytical responses.

✓ Every recommendation depends on assumptions.

✓ Assumptions are statements accepted as true; risks are possible events or conditions that could affect the Outcome.

✓ Prioritise assumptions based on their importance and the weakness of supporting evidence.

✓ Sensitivity analysis identifies which individual variables have the greatest effect.

✓ Scenario Analysis examines coherent combinations of conditions and possible futures.

✓ Test break-even thresholds rather than relying only on base-case estimates.

✓ Use decision trees when decisions and outcomes occur in stages.

✓ Expected Value is useful when reasonable probabilities exist, but it doesn't account fully for downside severity, timing, ethics, flexibility, or risk capacity.

✓ The strategy with the highest forecast return may not be the most resilient.

✓ Additional information has Value only if it can improve or change the decision.

✓ Pilots should test critical assumptions and lead to defined next decisions.

✓ Real Options preserve the ability to delay, stage, expand, abandon, switch, or partner.

✓ Identify no-regret moves that create Value across several scenarios.

✓ Use signposts to monitor change, triggers to prompt decisions, and contingencies to define the response.

✓ Match precision to the quality of the evidence.

✓ Recognising Uncertainty doesn't mean refusing to recommend.

✓ In the presentation, identify the critical assumptions, show their impact, and explain how implementation will adapt.

Looking Ahead

Decision-Making Under Uncertainty completes Part II's focus on the thinking skills required for strong case solving:

Critical Thinking tests the quality of reasoning.
Strategic Thinking creates direction and choice.
Systems Thinking reveals relationships and consequences.
Analytical Thinking turns evidence into insight.
Creative Thinking expands the solution space.
Ethical Thinking examines what should be done.
Decision-Making Under Uncertainty makes action possible without perfect information.

The next part of the manual moves from these foundational thinking skills to practical tools for understanding the situation. It begins with SWOT Analysis, used not as a four-box laundry list, but as a way to connect internal capabilities with external conditions and generate strategic implications