How To Create and Use a Business Decision Tree

Next time you need to make an important decision about your business, imagine a tree. The trunk is the new decision, while the branches are the various choices you might make and their probable outcomes. When this chain of events is represented graphically, it’s called a decision tree, a key tool for understanding how each option might play out.

Here’s how decision trees work, when to use one, and how to create your own, alongside an example use case.

What is a business decision tree?

A business decision tree is a diagram or visual decision-making tool, similar to a flowchart, that maps out potential courses of action, their anticipated outcomes, and the statistical probability of those outcomes actually occurring. It gets its name from its visual resemblance to a tree, with different branches representing the range of possible actions and outcomes for a business decision.

Business decision trees provide visual clarity to the decision-making process. When you create a decision tree, you’re better able to conceptualize and analyze your options. Seeing the different paths and where each decision leads can help you decide the best course of action. They are used in numerous industries to evaluate decisions in operations, manufacturing, marketing, human resources, finance, and more.

Components of a decision tree

• Root node
• Decision nodes
• Chance nodes
• End nodes

Decision tree diagrams have four main nodes that, when connected, make up the branches.

Root node

The starting point for your decision tree is the root node—the trunk of the tree—which represents the core decision you plan to make. For example, an ecommerce business planning for the holiday season might consider a limited-time sale on select products. 

Decision nodes

Decision nodes (also called decision points) connect by a line to the root node and are typically represented as squares in a decision tree, indicating a point where you must make a choice. Branches radiate out from decision nodes, each representing different options available to the business. 

For example, the ecommerce business mentioned above would have at least two branches connecting to the root node, comparing expected outcomes from offering a sale versus no price break. As this business evaluates product options, they could add a subsequent branch and decision node that reads: “Should we offer free shipping?”

Chance nodes

Chance nodes are typically shown as circles and represent uncertain outcomes outside the team’s control, such as market trends, supplier behavior, or world events. Each branch from a circle represents a possible scenario (e.g., high customer demand, TV product placement, or oil embargo) and is typically assigned a probability percentage.

End nodes

End nodes (also called endpoint nodes or leaf nodes) are typically shown as triangles and represent potential outcomes of a specific path, often expressed as a financial value, such as a projected profit or loss.

The sequence flows from left to right: You start with a root node, then branch out into decision nodes and chance nodes to account for market variables, and potentially pass through more decision and chance nodes to model the complexity of your problem set. Eventually, you reach the end nodes, where you can compare the expected value of each choice.

When to use a business decision tree

A carefully constructed decision tree helps companies assess multiple options, weigh advantages and risks, and make informed decisions. Here are some scenarios where decision trees can help.

• Launching a new product line. Before committing to a large production run, use a decision tree to map out high-demand, medium-demand, and low-demand scenarios. This helps you calculate the technical feasibility of a big launch and its associated costs. You can also explore the probabilities of profits both from sales that top expectations or losses due to unsold inventory.

• Determining marketing spend. If you are torn between investing your budget into paid search, SEO, webinar marketing, or influencer partnerships, a decision tree can help you weigh the expected customer acquisition cost (CAC) and conversion rates of each.

• Vetting a new supply chain partner. When considering switching manufacturers, a decision tree’s chance nodes can account for unexpected events, such as shipping delays, quality control shortfalls, or rising raw material costs.

• Setting a pricing strategy. You can use a decision tree to compare premium pricing versus competitive pricing and see how each choice may affect conversion rates, profit margins, and total sales.

• Evaluating software options. If you are deciding whether to build a custom solution or pay for existing software (e.g., a Shopify app), a decision tree can help you compare the upfront costs against long-term maintenance and potential revenue gains. This visual data analysis enables more effective decision-making than going off instinct or anecdotal evidence.

• Choosing to invest in automation or keep a manual process. A decision tree can compare the upfront time and cost of automation (such as with Shopify Flow) versus retaining manual workflows and their long-term operational impact. You can predict how much implementation will cost, but also plan for expected efficiency gains and error reduction.

• Navigating legal or regulatory hurdles. If new shipping regulations or tax laws affect your business, a decision tree makes it easy to map out various compliance paths and the potential financial impact of each on your business vision.

How to make a business decision tree

1. Define your business decision and metrics for success
2. Map your controllable choices as decision nodes
3. Chart your chance nodes
4. Quantify outcomes using a mix of hard data and informed assumptions
5. Determine the expected value of each outcome and choose your path

Here’s how to make a business decision tree:

1. Define your business decision and metrics for success

Start with a single, clear question. This will function as the root of your business decision tree. You’re identifying the initial choice you need to make, from which all other choices will stem. For example, you might ask: “Should we launch our new product in Q2 or Q3?” You will depict this as the initial square decision node on your business decision tree.

Every choice must be mutually exclusive, meaning you can’t pick both paths at the same time. This clarity is essential for a successful logical flow. In this case, you can either launch in Q2 or Q3, not both.

You also need to define your measurements for success. This might mean profit, growth, risk reduction, or customer impact. You want every branch of the tree to be evaluated the same way.

2. Map your controllable choices as decision nodes

Your root node (square) should have branches representing your options. If your core query is when to launch a new product, your branches could represent a Q2 launch, a Q3 launch, or alternatives like a pilot program. Only chart the options that are operationally realistic.

3. Chart your chance nodes

For every path you’ve identified, look for the points of uncertainty—the factors you cannot control. These are your chance nodes, and they’re represented as circles. Identify the range of alternative outcomes that could happen for each choice. For example, if you choose to launch a new product in Q2, the uncertainty might be market demand, which you would mark with a circular chance node. Then the possible outcomes—“high demand,” “medium demand,” and “low demand”—would branch off from the circular node.

4. Quantify outcomes using a mix of hard data and informed assumptions

To predict outcomes, assign percentages to each branch splitting off a node. Make sure the outcomes add up to 100%. Access to historical data, such as past sales, and research on market trends, can help as you evaluate and assign probabilities. If you want to add additional options (e.g., “initial high demand followed by sustained low demand”), you can, as long as the total percentages add up to 100.

5. Determine the expected value of each outcome and choose your path

It’s now time to assign values to the leaf nodes at the end of every branch. These are frequently represented in terms of either financial or operational impact, depending on the question you posed in your root node. For example, how much revenue you would take in based on the identified possibilities of a Q2 launch, or the rate of total sales volume growth compared to a year earlier.

At this point, you need to calculate the expected value of each possible path. Follow the example below to learn how to calculate this expected value. 

Business decision tree example

Here’s a basic hypothetical example of how a business decision tree can aid you in charting a path for your business. Imagine you’re exploring new markets for your US ecommerce business, and you’re considering making a play for customers in Mexico.

You start your decision tree with a square decision node that asks: “Should we expand to Mexico?” Three options branch from that node—each with associated costs related to operations and advertising: 

1. Launch a full-scale expansion into Mexico. (Cost: $400,000)

2. Run a pilot program targeting a small segment of the Mexican market. (Cost: $75,000)

3. Do nothing. (Cost: $0)

The third branch is a dead end with no financial changes. However, the other two branches lead to circular chance nodes representing an external factor that will affect your success. In this simplified example, the chance nodes represent market demand. You can create branches from each node that offer the following options, each accompanied by your estimated probability of them happening:

1. High market demand (20%)

2. Medium market demand (50%)

3. Low market demand (30%)

From there, you can create triangular end nodes with predicted financial outcomes. Those nodes might be:

1. Full-scale expansion + high market demand ($1.5 million in sales, 20% probability)

2. Full-scale expansion + medium market demand ($800,000 in sales, 50% probability)

3. Full-scale expansion + low market demand ($400,000 in sales, 30% probability)

4. Pilot program + high market demand ($200,000 in sales, 20% probability)

5. Pilot program + medium market demand ($100,000 in sales, 50% probability)

6. Pilot program + low market demand ($48,000 in sales, 30% probability)

You’re then looking at a decision tree featuring your potential courses of action, predicted likelihood, and predicted sales revenue in each scenario. Next, you need to calculate the expected value of both possible choices.

Expected value (EV) is calculated by taking the sum of all outcome values identified in your decision node multiplied by their assigned probabilities: In this example, we’re examining three potential outcomes for a full-scale expansion and three separate potential outcomes for a pilot program: 

EV = (Probability x Outcome 1) + (Probability x Outcome 2) + (Probability x Outcome 3) 

For example, the expected value of a full-scale expansion is:

EV = (0.2 x $1.5 million) + (0.5 x $800,000) + (0.3 x 400,000) = $820,000 in expected sales

$820,000 - $400,000 (cost of full expansion) = $420,000 net gain, representing a return on investment of 105%.

The expected value of the pilot program is:

EV = (0.2 x $200,000) + (0.5 x $100,000) + (0.3 x $48,000 = $104,400 in sales

$104,400 - $75,000 (cost of pilot) = $29,400 net gain, representing a return on investment of 39.2%.

In this example, a full-scale launch is expected to result in a $420,000 net gain, while the pilot program is expected to result in a net gain of $29,400. Although the pilot program might have seemed like a prudent first step into the Mexican market because it requires far less upfront investment, this strategy ultimately would have produced a significantly lower return on investment in this example. You can see why, as a business manager, taking the time to craft a thoughtful business decision tree that models the expected value of likely outcomes is a powerful tool for decision-making.