Decision trees are a visual and numerical technique used to support decision-making by mapping out possible choices, their associated risks, probabilities, and financial outcomes, allowing businesses to make informed and rational choices under uncertainty.
What are decision trees?
A decision tree is a diagram that helps businesses make structured decisions where there is uncertainty about outcomes. It shows different courses of action, chance events, and the resulting consequences in terms of potential financial gains or losses.
Each decision tree is built using a combination of:
Decision nodes (represented by a square): These indicate points at which the business has to make a choice between different options.
Chance nodes (represented by a circle): These show where an outcome is uncertain and each possible result has an associated probability.
Branches: These extend from decision and chance nodes to show the various possible options or outcomes.
Outcomes: These are the final results of a sequence of decisions and chance events, each with an expected monetary value (EMV).
Decision trees are useful in any situation where a business must evaluate multiple possible scenarios. This could include whether to launch a new product, invest in new equipment, enter a foreign market, or change suppliers. They allow decision-makers to compare the likely financial returns of different options while accounting for the risks involved.
Constructing a decision tree
Creating a decision tree involves building a clear structure that shows all possible outcomes and decisions in a logical sequence. Below is a step-by-step guide for constructing a simple decision tree:
Step 1: Define the decision to be made
Start by identifying the decision that the business faces. For example, a firm may be deciding whether to launch a new digital product or stick with its existing product line.
Step 2: Draw the decision node
Represent this choice as a square (decision node). From this square, draw branches to show each available option (e.g. 'Launch New Product' and 'Maintain Current Product').
Step 3: Add chance nodes for uncertain outcomes
At the end of each decision branch, draw a circle (chance node) to show that the result of the choice is not guaranteed. For example, launching a new product may result in high sales or low sales.
Step 4: Add outcome branches and assign probabilities
From each chance node, draw branches for the different possible outcomes. Assign each outcome a probability, based on market research, historical data, or expert opinion. The sum of all probabilities at each chance node must equal 1.
Step 5: Add monetary values for each outcome
Each branch leads to an outcome, which must be given a monetary value. These are typically financial returns, such as forecasted profits or losses.
Step 6: Calculate the expected monetary value (EMV)
At each chance node, calculate the expected monetary value. This is a weighted average of all possible financial outcomes, using the probability of each outcome as the weighting.
Formula for EMV:
EMV = (Probability of Outcome 1 × Value of Outcome 1) + (Probability of Outcome 2 × Value of Outcome 2) + ...
Repeat this calculation for all branches.
Step 7: Choose the best decision path
Once the EMVs have been calculated for all branches, compare them. The branch with the highest EMV represents the most financially advantageous option. If costs are involved (such as initial investment), subtract these to calculate net gains.
Calculating EMV and net gain
Worked example
A business is considering launching Product A or Product B. The financial outcomes and probabilities are as follows:
Product A:
60% chance of success → £100,000 profit
40% chance of failure → £20,000 loss
EMV for Product A:
EMV = (0.6 × 100,000) + (0.4 × -20,000)
EMV = 60,000 - 8,000 = £52,000
Product B:
50% chance of success → £80,000 profit
50% chance of failure → £10,000 loss
EMV for Product B:
EMV = (0.5 × 80,000) + (0.5 × -10,000)
EMV = 40,000 - 5,000 = £35,000
Assuming both options require the same initial investment, Product A would be the better choice, as it has a higher EMV.
If the investment cost for Product A is £30,000, the net gain is:
Net gain = EMV – Investment
Net gain = 52,000 – 30,000 = £22,000
This shows the financial advantage of Product A after accounting for the cost of implementation.
Identifying the best decision path
Once all the EMVs are calculated, the decision-maker should compare the net EMVs (i.e., EMV minus any costs or investments associated with that path). The option with the highest net EMV should be chosen, assuming the business is comfortable with the level of associated risk.
However, the choice of path also depends on the business's risk tolerance:
Risk-averse businesses may choose a lower EMV with lower risk.
Risk-neutral businesses focus solely on the highest EMV.
Risk-seeking businesses may choose higher-risk, higher-reward options.
It is essential to understand that the EMV only tells us the average return over time. It does not guarantee any specific result.
Advantages of using decision trees
Decision trees offer several advantages that make them a useful tool for A-level Business students and real-world companies alike:
1. Structured decision-making
Decision trees provide a logical and systematic framework for analysing complex choices. This structure helps ensure all possible scenarios are considered, and the consequences of each are clearly understood.
2. Incorporation of risk and probability
They allow businesses to include risk assessment by assigning probabilities to different outcomes. This enables the decision-maker to account for uncertainty and make choices based on expected values rather than guesswork.
3. Visual clarity
By presenting information in a diagrammatic format, decision trees make it easier to communicate the logic behind decisions. Stakeholders can visually track the sequence of decisions and outcomes.
4. Financial focus
As each option is linked to a monetary value, decision trees help businesses evaluate profitability and make financially sound choices. This is especially useful for capital investment decisions.
5. Comparison of alternatives
Different decision paths can be directly compared using EMVs, allowing businesses to rank options based on their expected financial outcomes.
Limitations of decision trees
Despite their usefulness, decision trees have important limitations which must be understood in order to use them effectively.
1. Data quality and reliability
Decision trees rely heavily on accurate data for probabilities and financial outcomes. If the probabilities are based on guesswork or biased assumptions, the EMVs will not be reliable. Businesses may end up making poor decisions based on flawed inputs.
For example:
Market research data may be outdated or unrepresentative.
Sales forecasts may not account for seasonal changes or competitor actions.
Cost estimates may be too optimistic, leading to inaccurate EMVs.
2. Oversimplification
While decision trees are designed to simplify decisions, they may oversimplify complex real-world problems. Not all business decisions can be reduced to a set of quantifiable outcomes and fixed probabilities.
Some common oversimplifications include:
Ignoring interactions between decisions
Assuming that all outcomes are known
Treating complex human behaviour as binary outcomes
As a result, decision trees may provide a false sense of certainty.
3. Ignores qualitative factors
Decision trees focus entirely on quantifiable data, particularly monetary values. They do not account for qualitative considerations that may influence a decision, such as:
Brand reputation
Customer loyalty
Employee satisfaction
Ethical concerns
Environmental impact
A project with a high EMV may still be undesirable if it damages the business’s image or goes against its values.
4. Dynamic and uncertain business environments
Decision trees are static tools. Once constructed, they do not automatically update when conditions change. However, the business environment is often dynamic, and key variables may shift quickly due to:
New regulations
Competitor actions
Exchange rate fluctuations
Shifts in consumer preferences
As a result, the probabilities and values used in a decision tree may become outdated very quickly, making the decision less relevant or appropriate.
5. Can become complex and unwieldy
For decisions involving many stages or options, decision trees can become very large and complex, making them hard to read and interpret. This is especially true when:
There are multiple decision points
Each choice has several possible outcomes
The tree spans several years or product lines
In such cases, it may be difficult to spot the best path or to understand the overall picture, which undermines the tree’s usefulness.
Making effective use of decision trees
To make the best use of decision trees, businesses and students should:
Use reliable and current data to assign probabilities and financial outcomes
Be transparent about assumptions and limitations
Combine decision trees with other tools like SWOT, PESTLE, or scenario planning
Update decision trees regularly as new information becomes available
Use them alongside qualitative judgement, not in isolation
Decision trees are a powerful decision-making aid when used with care and insight, providing structure, clarity, and quantifiable evidence to support complex business decisions.
Practice Questions
Assess the usefulness of decision trees to a business planning to launch a new product in a competitive market.
Decision trees are useful for helping a business assess risk and forecast the financial outcomes of launching a new product. They provide a structured, visual layout of possible choices and outcomes, making it easier to compare options. By including probabilities and expected monetary values, firms can make data-informed decisions. However, decision trees rely on accurate estimates for probabilities and revenues, which can be difficult in competitive markets where outcomes are uncertain. They also ignore qualitative factors like brand impact or customer loyalty. Overall, while useful, decision trees should be used alongside judgement and market insight to guide final decisions.
Explain how a business would calculate the expected monetary value (EMV) of a decision and how this might influence its choice.
To calculate EMV, a business multiplies the monetary value of each possible outcome by its probability, then adds the results. For example, if there is a 70% chance of earning £100,000 and a 30% chance of losing £20,000, the EMV would be (0.7 × 100,000) + (0.3 × -20,000) = £70,000 - £6,000 = £64,000. This figure gives the average return expected. A business can use EMV to compare decisions, choosing the one with the highest EMV if it wants to maximise profit. However, risk tolerance and other qualitative factors may still influence the final decision.
FAQ
Expected monetary value (EMV) represents the average return a business can expect from a decision if the decision were repeated many times under the same conditions. It is a theoretical estimate, calculated by multiplying each possible outcome by its probability and summing the results. It helps businesses compare options using objective, data-driven analysis. However, it is important to understand that EMV is not the actual outcome. In reality, only one of the possible outcomes will occur, and it may be better or worse than the EMV. For instance, a decision with an EMV of £50,000 could still result in a loss if the less likely negative outcome occurs. EMV is useful for planning and evaluating risk but should not be mistaken for a guaranteed result. The actual outcome depends on chance events and external factors that may or may not align with the probabilities assumed during the analysis.
Biases can significantly affect the accuracy of a decision tree by distorting the inputs used for probabilities and financial outcomes. Confirmation bias may lead managers to overestimate the likelihood of a favourable outcome or underestimate potential risks because they subconsciously seek evidence that supports their desired course of action. Overconfidence bias can result in assigning unrealistic probabilities or revenue projections, especially if based on gut feeling rather than market research. Anchoring bias might cause a business to rely too heavily on initial data or outdated figures, skewing the decision tree’s objectivity. Additionally, if financial values are based on past experiences in different market conditions, the model may not reflect the current reality. These biases can lead to misleading EMVs and poor strategic choices. To mitigate this, businesses should use third-party data, challenge assumptions through peer review, and update models regularly with new information to ensure objectivity and relevance.
Despite having access to data, a business might avoid using a decision tree for several reasons. First, the complexity of constructing a detailed tree for a multi-stage or high-stakes decision can be time-consuming and resource-intensive, especially if the decision involves numerous variables. Second, while quantitative data may be available, decision trees do not incorporate qualitative factors like brand perception, employee morale, or customer relationships, which may be crucial to the business. Third, if the decision involves a high degree of uncertainty or volatility, such as entering an emerging market, the assigned probabilities and outcomes may lack reliability, undermining the usefulness of the EMV. Moreover, businesses with strong experience or intuition in a particular area may prefer to rely on strategic judgement or alternative tools like SWOT or PESTLE. Finally, the senior decision-makers may not be familiar with interpreting EMVs or may view the decision tree as too rigid or reductionist in approach.
Yes, decision trees can be used for non-financial decisions, but their effectiveness may be limited depending on the context. For example, they can help evaluate options like recruiting methods, expansion strategies, or partnerships, where outcomes are not strictly financial but still involve risk. In these cases, branches of the decision tree may reflect outcomes such as customer satisfaction, market share gains, or staff retention rates. However, since decision trees rely on assigning numerical values and probabilities, applying them to non-financial metrics can be challenging. Estimating the probability of reputational damage or employee dissatisfaction involves a degree of subjectivity, which can reduce the reliability of the model. Additionally, the inability to quantify all possible consequences may lead to an oversimplification of the decision. To be effective, decision trees for non-financial decisions should be combined with qualitative analysis and should make assumptions explicit, so decision-makers are aware of their limitations.
To improve the reliability of probabilities used in a decision tree, businesses should adopt a data-driven and systematic approach. First, they should collect comprehensive historical data on similar decisions, such as sales figures, market performance, or project outcomes, to base probability estimates on actual trends rather than guesses. Secondly, market research can help identify consumer behaviour patterns, competitive dynamics, and industry benchmarks that can support more accurate forecasts. Consulting external experts or using forecasting software can also improve the robustness of the inputs. Sensitivity analysis can be used to test how changes in probabilities affect the final EMV, revealing which variables have the greatest impact. Additionally, businesses should avoid over-reliance on internal estimates by using cross-functional teams to review assumptions and provide diverse perspectives. Regularly updating the decision tree when new data becomes available ensures that the model remains relevant and reduces the risk of basing strategic decisions on outdated or biased information.
