Decision Trees: Why And How To Grow These Bad Boys

A decision tree is a thinking tool that uses a tree-like model to evaluate decisions and their possible consequences, including chance events, resource costs, and utility. It’s a good way to display a high-stakes decision with downstream effects. Using decision trees ensures we identify a strategy most likely to reach our goals. The tree is a flowchart-like structure. They are taught to undergrad students of business, health economics, and policy to help them predict decisions and their consequences.

As all thinking tools, decision trees come with ups and downs. Decision trees are simple. Easy to understand and interpret. A brief explanation is most often enough. Even with little data, they can be valuable and can provide important insights solely by virtue of exploration. They help determine the worst, best, and expected values for situations. Decision trees are widely applicable, so they fit with other models as well. But they are unstable, meaning a small change in the conditions or data can skew the results. Decisions trees are often inaccurate. It’s a trade-off: you get an easy, efficient, and simple tool, but you sometimes pay the price by making errors.

Use decision trees when you need to break down a large, complex problem, or when you need to predict outcomes. This is because they help you dilute multifaceted problems into it’s component action points and their outcomes. By definition, this results in a deeper understanding of the challenge at hand and it’ll give you a clearer view of the expected value behind a decision. The basic idea is to narrow down the options until you reach a destination that meets your desires. In the end, what could have been utter chaos is turned into an easy, visual path, helping you reach a consensus without endless debate and unproductive thinking.

They find their use everywhere where you may want to reveal potential outcomes and the overall utility of different strategies. The structure is rather simple: the parent node, or the decision itself, usually in the form of a square that represents one of the options that can be selected; the child node, or the chance, often depicted as circles, pointing to whether an outcome is dependent on chance or something outside the decision-makers control; the leaf nodes, or the triangle-shaped endpoints that represent the final outcomes of a choice. Grow your tree. From the root, to the branch, to the leaf. An example: “Should I start a business?” (root); “tech startup” or “retail firm” (decision); market “favorable” or “unfavorable” (chance); “invest in marketing” or “cut costs” (branches); “business succeeds with 500 thousand dollars in profit” or “business fails with 100 thousand dollars in losses” (leaves).

Real life implications of decision trees:

• Business: a company might be deciding on whether to expand internationally, and it may draw out a tree including the options of Europe and Asia, the outcomes of success or failure, and the probabilities and payoffs to make a viable choice;

• Healthcare: a doctor might be prescribing treatment plans for their patient, and the options could be surgery or alternative therapy, while the outcomes to evaluate are risks, costs, and recovery rates;

• Personal: you could be deciding between stocks and bonds, with your options being high-risk, high-return stock, and low-return, low-risk bonds, which you can evaluate based on your risk-aversity and market conditions;

• Career: you might be choosing between two job offers, one with a higher salary and hours, the other with a lower pay but better work-life balance, and you could weigh the financial and lifestyle outcomes with a decision tree to choose what’s suitable.

How you might use decision trees as a thinking tool: (1) clearly state the decision you need to make, such as whether you want to invest in a new product line; (2) list all the possible alternatives or actions you can take, like expanding current offerings or doing nothing; (3) for each decision, consider the potential outcomes, including odd events and uncertainties, like low demand or a spike in customers; (4) estimate the likelihood of each outcome and weigh the associated benefits and/or costs, choosing the decision path with the highest expected value. Extra tips to be effective: (1) quantify whatever is uncertain, trying to be as objective as possible; (2) facilitate communications, bringing others into the conversation if applicable; (3) break down what is complex and not easily understandable; (4) think long-term wins, not short-term gains.

Thought-provoking insights. “Every choice you make has a consequence; decision trees make the consequences visible.” the structured and visual nature of a decision tree helps avoid hasty, emotional decisions. “In uncertainty, clarity is the compass.” by visualizing outcomes, decision trees bring clarity to vagueness. “The best decision is not always the most obvious one.” quantitative evaluations in decision trees can reveal counterintuitive, but optimal solutions. Use them with care. Don’t oversimplify. Cut off unnecessary branches. Combine math and preference, objectivity and subjectivity, for best results. Grow those goddamn trees!

Grab the framework and start growing those decision trees today.