If you’re a sports fan, perhaps you’ve wondered how NFL teams draft players? A University of Pennsylvania study found success in draft prediction with a simple tool—decision trees. When paired with machine learning, decision trees are also used to understand NBA statistics. So, what exactly is a decision tree?
Decision trees are tools mainly used for problem solving and classification. They lay out possible outcomes or subcategories, creating a visual map. This map is made up of a root, branches and leaves. Below is an example taken from the NFL study we mentioned above.
There are many decision tree examples for business, and their benefits are not exclusive to professional sports. The flexibility and simplicity of decision trees allow businesses to apply them to almost any scenario. Businesses in many sectors use them in a variety of ways to improve their services, make strategic decisions, and increase productivity.
Customer support agents are constantly faced with “what if?” situations. Decision trees can help agents navigate a number of scenarios, including customer returns, troubleshooting, and user authentication to name a few. Here's a great example of how a decision tree can streamline customer returns and remove confusion.
Without this process laid out, some support agents might reach out to the customer before printing the label, creating an unnecessary delay. The decision tree empowers all team members to handle returns in a consistent way.
Decision trees are great at breaking down phone call scripts, giving agents a concrete roadmap of solutions. Because conversations can be scattered, decision trees are used to contain all the possible directions a call might take. New employees who lack experience or training can still support customers with a decision-tree script on hand.
Perhaps the most crucial decision tree perk for customer support agents is process improvement. Finding faster and better ways to help customers is always a priority. Once a process chain is visually mapped, it’s much easier to spot bottlenecks and make fast revisions. Keeping a process chain updated through a living decision tree can make a huge difference and reduces, or even eliminates, the need to constantly update your team's internal documentation.
65% of customers want to buy from companies that offer quick and easy online transactions. As online shopping continues to take over, companies are using predictive analytics to get an edge on the market. Businesses can predict customer choices and preferences by pairing analytics technology with the capabilities of decision trees.
Customer choices, preferences, and actions are plugged into a decision-tree algorithm that generates outcomes. Previous purchases, search history, and link sharing are just a few components that could be used in the tree.
Decision tree algorithms are not strictly native to retail shopping or the customer experience. They’re also used to make internal predictions and answer questions. Which customers are likely to churn? How much can we upsell a customer? What products should we recommend?
Financial institutions rise or fall based on their ability to make accurate predictions. Every decision must be scrutinized by weighing expected gains against potential risk. Decision trees help by parsing out every variable and alternative possibility related to an investment decision.
In the financial world, decision trees are commonly used to calculate Net Present Value (NPV) from an investment opportunity. Data and statistics associated with the decision are plugged into the decision tree, revealing each possible outcome and the best way to generate value. In the hypothetical decision tree example below, a company creates a tree to decide whether they should take action or stand by.
What appears are four different outcomes and their potential value. Note the answer still isn’t 100% clear. The middle branches include economic probability, which is likely to fluctuate. This risk and reward assessment is left to the ultimate decision-maker. But even with this variability, the decision tree captures each scenario, making the final decision abundantly clear.
Weighing investment options, writing loans, and determining interest rates are all opportunities to use decision trees. Information like dollar amounts, stock market points, and credit scores can all be inserted into a decision tree. If your business is experimenting with decision trees, consider all the quantitative data you have at your disposal so you can get the most out of your results.
Making critical decisions that affect the entirety of a business is no easy task. Executive leadership must be able to navigate uncertainty accurately. Like throwing a stone into a lake, every high-level decision creates a ripple that influences all aspects of a business. Decision trees have a proven history of lighting the path forward for large corporate entities—helping them to navigate operational shifts such as location changes, team restructuring, acquisitions, and more.
Consider Gerber, the renowned baby products company. In 1999, the brand created a decision tree to navigate a complex situation. A possibly harmful chemical, phthalates, was reported to appear in the company’s products. The Consumer Product Safety Commission, Greenpeace, and the national media were all bearing down.
Should Gerber conduct a massive recall or await the results of an investigation? To make the right call, they used the decision tree below.
Although phthalates were eventually approved for use, Gerber wasted no time in devising a plan. If your business is up against a significant challenge like this, decision trees can clarify an emotionally charged decision by keeping things objective.
Decision trees aren’t limited to any specific business, obstacle, or purpose—and that’s what makes them so useful. The next time you’re faced with a high-stakes business decision, remember the value of decision trees. Their simple framework can bend and mold to the situation at hand, revealing every component you need to consider.
