Using Decision Trees to Solve Low Feature Adoption
Published in
3 min readJul 3, 2020
Decision trees provide a systematic way of looking at a problem you need to solve. It provides a step-by-step method for breaking down big problems into smaller, more manageable ones while allowing you to prioritize the problems and providing clear “decision gates” at each step.
There are two basic kinds of decision trees:
- Problem trees — created by answering “Why?”
- Solution trees — created by answering “How?”
How to create a decision tree
Problem tree
A decision tree should cover the whole problem from a high level. Here are a few basic principles to keep in mind
- Start breaking down the problem into separate categories/branches.
- Use the MECE principle: mutually exclusive, collectively exhaustive: Mutually exclusive means there is no overlap between different parts of the tree; Collectively exhaustive means they cover the whole problem.
- Do not go into specific hypotheses: focus on capturing the broad categories that make up the problem.
- Apply the 80/20 rule: focus on the few parts of the problem that are most impactful.
- Whenever possible, base your branches on data, rather than…