Skip to main content

Chapter 13

Chapter 13: Decision Making Under Uncertainty - Making Strong Decisions When the Future Cannot Be Predicted With Confidence

Learning Objectives

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

  • distinguish risk from uncertainty

  • recognise different types of uncertainty in a case

  • identify the genuinely unknown variables

  • use expected value to evaluate uncertain outcomes

  • understand probability-weighted decision making

  •   build decision trees

  • compare decisions across multiple possible futures

  • consider the value of information

  • recognise the importance of flexibility and real options

  • incorporate risk into investment decisions

  • design decisions that preserve flexibility

  • communicate uncertainty confidently to judges

  • make a recommendation even when the future is unclear

Why This Matters

Business decisions are rarely made with perfect information.

A company may be considering:

  • entering a new market

  • launching a new product

  • acquiring another company

  • building a new facility

  • investing in technology

  • expanding internationally

  • changing its business model

Management may have forecasts.

  • It may have historical data.

  • It may have industry benchmarks.

  • It may have consultants and analysts.

But nobody knows exactly what will happen. 

  • The market may grow faster than expected.

    Or slower.

  • Customers may adopt the product.

    Or they may not.

  • A competitor may respond aggressively.

    Or do nothing.

  • A regulation may change.

    Or remain unchanged.

  • Interest rates may rise.

    Or fall.

This creates a fundamental challenge:

How do you make a good decision when you cannot know the future with certainty? 

This is one of the most important skills in case solving.

Discover Your Mad Skills Principle

Good decision making is not about predicting the future perfectly. It is about making a decision that remains intelligent across plausible futures.

Risk vs. Uncertainty

These terms are often used interchangeably.

They are not exactly the same.

Risk

Risk exists when you can identify possible outcomes and assign reasonable probabilities to them.

For example:

  • There is a:

    a

    60% probability 

    that sales will be strong.

  • And a:

    a

    40% probability 

    that sales will be weak.

You can model those outcomes.

Uncertainty

Uncertainty exists when the possible outcomes—or their probabilities—are much less certain.

For example:

  • A company is entering an entirely new market.

  • There may be little historical information.

  • Customer behaviour is unknown.

  • Competitor reactions are difficult to predict.

You cannot confidently say:

"There is a 65% probability of success."

The information simply isn't strong enough.

A Useful Distinction

Think of the spectrum this way:

Known -->

↓

Risk -->

↓

Uncertainty -->

↓

Unknown

As you move to the right, prediction becomes more difficult.

Your decision-making approach needs to adapt.

The Problem With False Precision

One of the biggest mistakes in case competitions is assigning probabilities simply because the model requires them.

For example:

  • 70% probability of success

  • 20% probability of moderate success

  • 10% probability of failure

    • The numbers look sophisticated.

    But where did they come from?

    If the case provides no evidence for those probabilities, you may simply be creating false precision.

    Discover Your Mad Skills Principle

    A probability is only useful when you can explain why you believe it.

    If you cannot defend the probability, don't pretend it is more precise than it is.

    Part 1

    1:

    Identify the Uncertainty

    Before trying to calculate uncertainty, identify it.

    Ask:

    What don't we know? 

    Typical areas include:

    Market

    • market size

    • growth

    • customer adoption

    • competitive response

    Financial

    • revenue

    • costs

    • investment requirements

    • interest rates

    Operational

    • implementation timing

    • capacity

    • productivity

    • supply chain

    Strategic

    • competitor behaviour

    • technological change

    • regulatory changes

    Human

    • employee adoption

    • management capability

    • customer switching behaviour

    Build an Uncertainty Map

    A simple way to organizeorganise uncertainty is:

    Uncertainty Impact Ability to Influence
    Customer adoption High Moderate
    Competitor response High Low
    Implementation cost Medium High
    Regulatory change High Low
    Employee training Medium High

    This immediately creates an important distinction.

    Some uncertainties are:

    are

    outside our control. 

    Others can be:

    be

    actively managed. 

    That distinction should influence the recommendation.

    Control What You Can Control

    Suppose customer demand is uncertain.

    You cannot force customers to buy.

    But you can:

    • run a pilot

    • test pricing

    • conduct customer research

    • launch in one market first

    • create a minimum viable product

    Similarly, if implementation costs are uncertain, you can:

    • obtain supplier bids

    • negotiate contracts

    • stage the investment

    • establish spending limits

    The goal is not to eliminate uncertainty.

    It is to reduce the uncertainty that matters most.

    Part 2

    2:

    Probability-Weighted Decisions

    When probabilities are reasonably defensible, expected value can be useful.

    The basic concept is:

    Expected Value = Σ (Probability × Outcome).

    For example:

    Outcome Probability NPV
    Strong market 30% $8M
    Base market 50% $4M
    Weak market 20% -$2M

    Expected NPV:

    (30% × $8M) +

    (50% × $4M) +

    (20% × -$2M) 

    =

    =

    $4.0M

    The expected NPV is therefore:

    $4.0M

    What Expected Value Does—Does and Doesn't—t Tell You

    Expected value provides a useful mathematical summary.

    But it does not mean:

    "We will make exactly $4 million."

    There may be no actual outcome of $4 million.

    The organizationorganisation might make:

    $8M8M,

    $4M 

    or

    or

    -$2M2M.

    The expected value simply represents the probability-weighted average.

    Discover Your Mad Skills Principle

    Expected value is a decision tool—not a prediction of what will actually happen.

    Expected Value and Risk

    Now compare two investments.

    Investment A

    Outcome Probability NPV
    High 50% $8M
    Low 50% $0M

    Expected NPV:

    $4M

    Investment B

    Outcome Probability NPV
    High 50% $5M
    Low 50% $3M

    Expected NPV:

    $4M

    Both have the same expected value.

    But they are very different decisions.

    Investment A has:

    higher upside 

    and

    and

    greater downside risk.

    Investment B has:

    lower upside 

    but

    but

    greater stability.

    Which is better?

    That depends on the organization'organisation's:

    • risk tolerance

    • financial capacity

    • strategic priorities

    • ability to absorb losses

    Risk Tolerance Matters

    Not every organizationorganisation should make the same decision.

    • A startup with substantial growth ambitions may accept greater risk.

    • A mature company with tight cash constraints may prioritizeprioritise stability.

    • A family-owned business may have different risk preferences than a publicly traded company.

    • A company facing financial distress may prioritizeprioritise cash preservation over long-term upside.

    The financial result is therefore only part of the decision.

    Part 3

    3:

    Decision Trees

    Decision trees are particularly useful when a decision involves:

    • multiple possible outcomes

    • sequential decisions

    • probabilities

    • future choices

    A simplified structure looks like:

    Decision --> Market Outcome.

    ↓

     

    Market                                                                                 Outcome

    ↙ ↓ ↘

                                                                            Strong Base Weak

                                                                                        ↓

                                                                             Financial Result

    The important idea is that the organizationorganisation may not have to make every decision today.

    Example: Market Entry

    Imagine a company is considering entering a new market.

    It has two choices.

    • Option A

    A:

    Launch immediately.

  • Option B

    • B:

    Conduct a $1M pilot first.

    The pilot provides information about customer demand.

    The decision tree might look like:

    Pilot -->

    ↓

    Strong demand → Expand

    | Weak demand → Stop

    This changes the nature of the decision.

    Instead of asking:

    "Should we invest $10 million?"

    we ask:

    "Should we spend $1 million to learn whether the $10 million investment is worthwhile?" 

    That is a much more sophisticated decision.

    Discover Your Mad Skills Principle

    Sometimes the best first investment is not the final investment. It is the investment that gives you better information.

    Part 4

    4:

    The Value of Information

    Information has value when it can change a decision.

    Suppose you are considering a:

    a

    $10M investment. 

    You are uncertain about market demand.

    You could spend:

    spend

    $500,000 

    on a market test.

    • If the test tells you that demand is weak, you avoid the larger investment.

    • If the test shows strong demand, you proceed with greater confidence.

    The $500,000 is not simply a research expense.

    It may be:

    be

    an insurance policy against a much largerbigger mistake.

    When Is Information Worth Buying?

    Information is valuable when:

    1. the decision is significant

    2. uncertainty is material

    3. the information can reduce uncertainty

    4. the information can change the decision

    5. the cost of obtaining the information is less than its potential benefit

    This creates an important question:

    What would we need to know to make this decision with greater confidence?

    The Value of a Pilot

    A pilot can provide information about:

    • customer demand

    • willingness to pay

    • operational feasibility

    • employee adoption

    • implementation costs

    • technical performance

    For example:

    Instead of:

    of

    a $10M full rolloutrollout,

    consider

    consider:

    a

    $1M pilot

    ↓

    • Measure:

      • adoption

      • cost

      • customer satisfaction

      • operational performance

      ↓

    • Then decide:

      Scale 

      or

      or

      Stop

    This creates flexibility.

    Part 5

    5:

    Real Options and Flexibility

    Some decisions have value because they preserve future choices.

    Consider:

    "Should we build a $20 million facility immediately?"

    versus:

    versus:

    "Should we lease temporary capacity while we test demand?"

    The second option may not have the highest immediate NPV.

    But it preserves flexibility.

    • If demand increases:

      Expand.

    • If demand disappoints:

      Exit.

    That flexibility has value.

    Option Thinking

    A decision can create future options such as:

    • expand

    • delay

    • abandon

    • scale

    • acquire

    • partner

    • switch suppliers

    • change pricing

    • enter another market

    When uncertainty is high, ask:

    Which decision gives us the most valuable future choices?

    Discover Your Mad Skills Principle

    Under uncertainty, flexibility can be worth more than maximizing the base-case forecast.

    Part 6

    6:

    Reversible vs. Irreversible Decisions

    This is another useful way to think about uncertainty.

    Reversible Decision

    The company can change direction relatively easily.

    Examples:

    • pilot program

    • temporary staffing

    • limited marketing campaign

    • short-term supplier agreement

    These decisions can often be made with less information.

    Irreversible Decision

    The company commits substantial resources that are difficult to recover.

    Examples:

    • major factory

    • acquisition

    • long-term infrastructure

    • large specializedspecialised technology investment

    These decisions deserve much greater analysis.

    The Uncertainty–Uncertainty/Commitment Matrix

    Think about decisions using two dimensions:

    • Level of Uncertainty


      • and

    • Size of Commitment

    Low Uncertainty + Low CommitmentCommitment:

    Move quickly.

     

       

    		Low Uncertainty + High CommitmentCommitment:

    AnalyzeAnalyse carefully.
    High Uncertainty + Low CommitmentCommitment:

    Experiment.

     

     

    		High Uncertainty + High CommitmentCommitment:

    Slow down, gather information, and stage the investment.

    This is a powerful framework for case solving.

    Part 7

    7:

    Scenario Planning

    When probabilities are difficult to establish, scenarios may be more appropriate.

    Instead of saying:

    "There is a 60% chance of high growth.growth,"

    you might develop:

    Scenario 1

    1:

    Rapid market growth

    Scenario 2

    2:

    Moderate market growth

    Scenario 3

    3:

    Market stagnation

    Then ask:

    How does our strategy perform under each scenario?

    Robust Strategies

    Imagine three strategies.

    Strategy High Growth Moderate Growth Low Growth
    A $10M $5M -$2M
    B $8M $6M $3M
    C $6M $5M $4M

    • Strategy A has the highest upside.

    • Strategy C has the strongest downside protection.

    • Strategy B performs reasonably well across all three futures.

    Now ask:

    Which strategy is most robust? 

    That may be Strategy B.

    The Robustness Test

    A robust strategy doesn't have to be the absolute best in every scenario.

    Instead, it performs acceptably across a broad range of plausible futures.

    This is particularly valuable when:

    • uncertainty is high

    • the decision is difficult to reverse

    • downside risk is significant

    • information is limited

    Deciphering Cases

    When you encounter a case with substantial uncertainty, ask five questions:

    1.
    • What do we know?

    Separate facts from assumptions.

    2.
  • What don't we know?

    • Identify the major uncertainties.

    3.
  • What could change the decision?

    • Focus on material uncertainties.

    4.
  • Can we learn more?

    • Look for:

    • research

    • pilots

    • experiments

    • customer testing

    • partnerships

    5.
  • Can we preserve flexibility?

    • Consider:

    • staging

    • delaying

    • scaling

    • exiting

    • partnering

    This turns uncertainty into something you can manage.

    Part 8

    8:

    Decision Making Under Uncertainty in a Case Competition

    You rarely have enough time in a case competition to conduct perfect analysis.

    That is okay.

    You don't need perfect information.

    You need a defensible decision process.

    A practical approach is:

    • Step 1

    1:

    Identify the key uncertainty.

  • Step 2

    • 2:

    Estimate the range of plausible outcomes.

  • Step 3

    • 3:

    Determine the financial impact.

  • Step 4

    • 4:

    Identify the downside risk.

  • Step 5

    • 5:

    Determine whether more information could change the decision.

  • Step 6

    • 6:

    Identify a way to reduce the risk.

  • Step 7

    • 7:

    Build flexibility into the implementation.

  • Step 8

    • 8:

    Make the decision.

    A Case Example

    Imagine an airline is considering launching a new international route.

    The company estimates:

    Investment = $5M5M.

    But passenger demand is uncertain.

    • High Demand

    Demand:

    NPV = $6M

  • Moderate Demand

    • Demand:

    NPV = $2M

  • Low Demand

    • Demand:

    NPV = -$3M

    The team could recommend:

    "Launch immediately."

    But aA stronger team might ask:

    "Can we test demand before making the full investment?"

    Perhaps the airline can:

    • begin with limited seasonal service

    • use existing aircraft capacity

    • partner with another carrier

    • test promotional pricing

    • measure advance bookings

    Now the strategy reduces uncertainty.

    The recommendation might become:

    "Proceed with a six-month pilot using existing capacity. Expand to a permanent route only if advance bookings exceed the required threshold." 

    That is a much stronger recommendation.

    Turning Uncertainty Into Implementation

    Notice what happened.

    The financial analysis identified:

    High uncertainty -->

    ↓

    Sensitivity analysis identified:

    Demand as the key driver -->

    ↓

    Decision analysis identified:

    Information has value -->

    ↓

    Strategy identified:

    Pilot first -->

    ↓

    Implementation created:

    A decision threshold -->

    ↓

    Recommendation became:

    Conditional but actionable

    This is sophisticated case solving.

    Decision Thresholds

    A threshold tells management when to continue, change, or stop.

    Examples:

    • Expand if customer adoption exceeds 10%.

  • Continue if gross margin remains above 35%.

  • Proceed to Phase 2 if NPV remains positive.

  • Exit if monthly losses exceed $500,000.

    • Thresholds make uncertainty manageable.

    They also make implementation measurable.

    Discover Your Mad Skills Principle

    Don't simply predict what will happen. Define what you will do when it happens.

    Part 9

    9: 

     Communicating Uncertainty to Judges

    Teams sometimes avoid discussing uncertainty because they fear it will make the recommendation look weak.

    Don't.

    Executives understand uncertainty.

    What they dislike is:

    is

    unmanaged uncertainty. 

    Compare:

    Weak

    • Weak:

      "There is some risk that adoption could be lower."

    • Better:

      Better

      "Customer adoption is our primary risk."

    • Strong:

      Strong

      "Customer adoption is the primary risk. If adoption falls below 7%, the investment no longer creates value. We therefore recommend a pilot with a 7% adoption threshold before full rollout."

    The third answer demonstrates:

    • awareness

    • analysis

    • quantification

    • mitigation

    • decision discipline

    The Executive Uncertainty Statement

    A useful communication structure is:

    "Our key uncertainty is ________. If ________ happens, the impact is ________. Therefore, we will ________."

    For example:

    "Our key uncertainty is customer adoption. If adoption falls below 7%, the investment becomes value value-destroying. Therefore, we will launch in two pilot markets first and proceed to full rollout only if adoption exceeds 7%."

    This is concise and powerful.

    Common Mistakes

    Mistake
      1 —
    • Pretending the future is certain

    certain.

    Forecasts are estimates.

    Mistake 2 —
  • Making up probabilities

    • probabilities.

    Don't create precise probabilities without a defensible basis.

    Mistake 3 —
  • Treating all uncertainty equally

    • equally.

    Focus on uncertainty that could change the decision.

  • Overanalysing

    Mistakeuncertainty. 4 — Overanalyzing uncertainty

    The goal is not to eliminate uncertainty.

    The goal is to make a decision despitedecide it.

    • Mistake 5 —
  • Ignoring flexibility

    • flexibility.

    A staged decision may be better than a binary yes/no decision.

    Mistake 6 —
  • Ignoring information value

    • value.

    Sometimes the best next step is to learn.

    Mistake 7 —
  • Ignoring downside

    • downside.

    High upside does not automatically make a good investment.

    Mistake 8 —
  • Failing to define thresholds

    • thresholds.

    If the strategy depends on certain conditions, specify them.

    Mistake 9 —
  • Giving an "it depends" recommendation

    • recommendation.

    Executives still need a decision.

    Your recommendation should be clear even when conditions matter.

    Mad Skills Drill

    Take a case with a major uncertain decision.

    Complete the following:

    Decision

    • Decision: What are we deciding?

       

      ____________________________

    • Key Uncertainty

      • Uncertainty:

      What don't we know?

       

      ____________________________

    • Key Driver

      • Driver:

      What assumption matters most?

       

      ____________________________

      Downside

    • Downside:

      What happens if we're wrong?

      • ____________________________

    • Information:

      Information

      What could we learn?

      • ____________________________

    • Flexibility:

      Flexibility

      How could we stage the decision?

      • ____________________________

    • Threshold:

      Threshold

      What condition determines whether we continue?

      • ____________________________

    • Recommendation:

      Recommendation

      What should management do?

      • ____________________________

    Advanced Mad Skills Drill

    Now challenge your recommendation.

    Ask:

    What would have to be true for our recommendation to be wrong? 

    List three conditions.

    • Condition 1

  • Condition 2

  • Condition 3

    • Now ask:

    Can we monitor those conditions? 

    If yes, create a metric.

    If no, ask whether you can reduce the uncertainty before committing significant resources.

    This is how you move from forecasting to active risk management.

    Building the Full Financial Decision Story

    At this point, your financial analysis should connect the entire Part IV toolkit.

    ROI

    ROI:

    How attractive is the return?

    ↓

    -->

    NPV

    NPV:

    How much value does it create?

    ↓

    -->

    IRR

    IRR:

    What rate of return does it generate?

    ↓

    -->

    Comparison

    Comparison:

    How does it compare with alternatives?

    ↓

    -->

    Sensitivity

    Sensitivity:

    How robust is the result?

    ↓

    -->

    Scenario Analysis

    Analysis:

    What happens under different futures?

    ↓

    -->

    Uncertainty

    Uncertainty:

    What don't we know?

    ↓

    -->

    Decision Design

    Design:

    How can we reduce risk and preserve flexibility?

    ↓

    -->

    Recommendation

    Recommendation:

    What should management do?

    This is the complete financial decision-making process.

    The Part IV Financial Decision Framework

    When facing an investment decision, use:

    1.

    Calculate: Calculate

    ROI / NPV / IRR -->

    Compare:

    ↓

    2. Compare

    Alternatives -->

    Test:

    ↓

    3. Test

    Sensitivity / Scenarios -->

    Identify:

    ↓

    4. Identify

    Uncertainty / Risk -->

    Decide:

    ↓

    5. Decide

    Invest / Stage / Test / Wait / Don't Invest -->

    Protect:

    ↓

    6. Protect

    Thresholds / Mitigation / Flexibility

    This framework provides a practical structure for case competition financial decisions.

    What Judges Are Really Looking For

    Judges are rarely impressed simply because a team knows how to calculate XNPV.

    They are looking for evidence that the team understands:

    • what drives the economics

    • what could go wrong

    • what matters most

    • what the organizationorganisation can control

    • what the organizationorganisation cannot control

    • what information is missing

    • what decision should be made

    • how the decision can be protected

    In other words:

    They are evaluating your judgment. 

    The spreadsheet is evidence of that judgment.

    It is not the judgment itself.

    Coach's Lens

    When coaching teams, I often find that students want certainty before making a recommendation.

    That is understandable.

    But business leaders rarely get certainty.

    They get:

    • incomplete information

    • competing priorities

    • imperfect forecasts

    • limited resources

    • time pressure

    The skill is not waiting until uncertainty disappears.

    The skill is learning how to make a defensible decision because of how you manage uncertainty.

    Chapter Summary

    The future is uncertain.

    Financial models cannot eliminate that uncertainty.

    But they can help us understand it.

    • Sensitivity analysis shows us:

      How much does the outcome change when assumptions change?

    • Scenario analysis shows us:

      What happens under different plausible futures?

    • Decision analysis goes one step further:

      What should we do given what we know—and what we don't know?

    The strongest decision makers don't attempt to predict everything.

    • They identify what matters.

    • They test what they can.

    • They learn where possible.

    • They preserve flexibility.

    • They establish thresholds.

    And then they decide.

    Key Takeaways

    ✓ Risk and uncertainty are related but different.

    ✓ Identify what you know and what you don't know.

    ✓ Focus on uncertainties that could change the decision.

    ✓ Don't create false precision with unsupported probabilities.

    ✓ Use expected value when probabilities are defensible.

    ✓ RecognizeRecognise that two investments with the same expected value can have very different risk profiles.

    ✓ Consider the organization'organisation's risk tolerance.

    ✓ Use decision trees when decisions occur sequentially.

    ✓ RecognizeRecognise the value of information.

    ✓ Use pilots and experiments to reduce uncertainty.

    ✓ Consider the value of flexibility.

    ✓ Distinguish reversible from irreversible decisions.

    ✓ Stage investments when uncertainty is high.

    ✓ Build decision thresholds into implementation.

    ✓ Design strategies that remain effective across plausible futures.

    ✓ Don't simply predict what might happen—decide what you will do if it happens.

    ✓ The strongest recommendation is not the one that assumes the future will cooperate. It is the one that is designed to succeed even when the future doesn't.

    Part IV Closing

    Financial analysis is ultimately not about Excel.

    • Excel is a tool.

    • ROI is a tool.

    • NPV is a tool.

    • IRR is a tool.

    • Sensitivity analysis is a tool.

    • Scenario analysis is a tool.

    Their real purpose is to help you answer one question:

    "What should we do?" 

    A strong case team can take a large amount of financial information and turn it into a clear decision.

    They can explain:

    • What the organizationorganisation can afford.

    • What creates value.

    • What the alternatives are.

    • What could go wrong.

    • What assumptions matter.

    • How robust the recommendation is.

    • How the organizationorganisation can protect itself.

    And ultimately:

    What management should do next. 

    That is the difference between financial analysis and financial decisiondecision-making. making.

    You are not trying to prove that your spreadsheet is correct.

    You are trying to demonstrate that your decision is defensible.

    Discover Your Mad Skills

    The deeper lesson of Part IV is simple:

    Numbers should reduce uncertainty—not create the illusion of certainty.

    • Use the numbers to understand.

    • Use the analysis to compare.

    • Use the scenarios to test.

    • Use your judgment to decide.

    And use your implementation plan to manage what happens next.

    Looking Ahead

    You now have the financial tools to support a case recommendation.

    But a great recommendation still has to be communicated.

    The judges need to understand:

    • what you recommend

    • why you recommend it

    • why it is financially credible

    • what risks you considered

    • how you will implement it

    The next challenge is therefore no longer:

    "Can we prove that the recommendation works financially?" 

    It is:

    "Can we make the judges believe this is the right decision?" 

    That takes us from financial decision-making into the next stage of case mastery: From Analysis to Action. Building the recommendation, implementation plan, and business case that turns financial insight into a decision.

    Back to top