Abstract
A question like "why are profits falling?" is too big to answer in one go. An issue tree splits it into branches, then splits those, until each leaf is something you can look up or act on.
At each level the branches should be MECE, the principle Barbara Minto developed at McKinsey: they shouldn't overlap, and together they should cover the whole question. A diagnostic tree asks why something is happening. A solution tree asks how you could get somewhere.
Method
- 01
Write the key question
Put one specific question at the root. Decide whether you're looking for a cause (why) or for options (how).
- 02
Make the first split
Break it into two to five parts. A formula helps if you have one, such as profit equals revenue minus costs.
- 03
Check the split
Make sure the branches don't overlap and that nothing important is missing.
- 04
Keep splitting
Break each branch down until the leaves are concrete enough to research or act on.
- 05
Pick what to test
Mark the leaves you think matter most and decide how you'll confirm or rule out each one.
When to use it
- A broad problem needs dividing up between people.
- You want to be sure the analysis hasn't skipped a whole area.
- You need to show someone how you approached a problem.
Common pitfalls
- Branches that overlap, so two people end up investigating the same cause.
- Mixing kinds of split at one level, like regions next to product lines.
- Going deep everywhere when only a few branches matter.
- Treating the tree as the answer. It's a plan for finding one.
Origin and references
Attributed to McKinsey & Company; MECE principle by Barbara Minto (1960s).
- Minto, B. (1987). The Pyramid Principle: Logic in Writing and Thinking. Pitman.
- Rasiel, E. M. (1999). The McKinsey Way. McGraw-Hill.
- Conn, C., & McLean, R. (2018). Bulletproof Problem Solving. Wiley.