“Hypotheses are scaffoldings erected in front of a building and then dismantled when the building is finished. They are indispensable for the workman; but you mustn’t mistake the scaffolding for the building.”
– Johann Wolfgang von Goethe, 17th Century German Author
If you smash together problem statements, hypotheses and disaggregation you get hypothesis trees, one of the go-to tools of McKinsey problem solving.
What is a Hypothesis Tree?
A hypothesis tree takes a problem statement and comprehensively disaggregates potential solutions. Any time you are trying to understand a problem and opportunities better, a hypothesis tree is a great tool.
Most projects at strategy consulting firms start with the team spending a few hours brainstorming and aligning on the hypothesis tree for the defined problem statement. The hypothesis tree helps a team scope out the scale of the problem and potential solutions, build context, prioritize some of the analysis and facts needed, and communicate the problem-solving structure and path.
In my experience, people tend to zero in on a root cause of a problem or a solution a bit too quickly, without really understanding all of the potential causes or solutions. These situations often lead to misdiagnosis of a problem or not arriving at the optimal solution. Before zeroing in on what you think the root cause or solution is, take a step back and think through what all of the potential options could be. Hypothesis trees are useful in thinking through all the possible options.
Below is an example of a hypothesis tree for a store that is trying to grow its sales.
The problem statement is simply “How can this store grow sales?” And, the hypothesis tree disaggregates the options the store has to grow sales down to three levels and can serve as the architecture to further analyze and diagnose the options and ultimately create a strategy to drive store sales growth.
What are the best practices with hypothesis trees?
Any time you are trying to diagnose a problem or create the potential universe of opportunities, please pull out the hypothesis tree from your toolkit. You’ll look like a genius. Here are the best practices in hypothesis trees:
Start with the Problem Statement Question
What are you trying to solve? Start there and state what you are trying to solve in the form of a question. The brain likes to answer questions, not statements.
Disaggregate the Whole
Separate the problem or opportunity into its discrete parts. Think about your problem or opportunity like a math problem. Component A + Component B + Component C = The Universe of Potential Issues or Opportunities.
Make it MECE
As stated before, make sure the components at each level don’t overlap with each other (Mutually Exclusive), and that the component at each level encompasses the entire universe of root causes or opportunities (Collectively Exhaustive).
Brainstorm each branch
If you constrain problem solving to each branch, you can generate a lot of great ideas, many more then you typically think possible.
Get others involved
In consulting projects, we always pressure test hypothesis trees with clients to make sure we were covering all the basics, have the right structure and hypothesis for the problem statement, and to elicit the strong hypotheses. And, when creating a hypothesis tree, collaborating with a diverse group of people or stakeholders typically creates a much higher quality output.
DOWNLOAD THE HYPOTHESIS TREE POWERPOINT WORKSHEET
To get you going on creating a hypothesis tree, download the free and editable Hypothesis Tree PowerPoint Worksheet.[sociallocker] [/sociallocker]
Take an important problem your team or company is trying to solve.
Create a problem statement. You can utilize the problem statement module to do this.
Once you have a clear problem statement, disaggregate potential solutions into 2-5 major hypotheses. Make sure they are MECE and at the same level of specificity. Next, tree out sub-hypotheses or ideas that would drive the major hypotheses. Next, list out the analyses or fact base you need to prove or disprove the hypotheses or determine the magnitude of the ideas.