Issue trees for student teams.

An issue tree is a map of the question. It keeps four people from researching the same paragraph and leaving the actual decision untouched. The Case Competition uses issue trees as a training object, not as a decoration for slide one.

Write the governing question at the top.

The question should be specific enough to fail. “How can the company grow?” is not a question. “Should Starbucks India accelerate store openings in 2015 given local coffee culture, joint-venture constraints, and competitor response?” can be analysed.

Then split. Each child should be necessary. Together the children should cover the question. That is the MECE test: mutually exclusive, collectively exhaustive. Students fail the test when two branches are synonyms, or when a critical constraint such as regulation or union response is missing.

Use business logic, not buzzwords.

Profit trees, 3Cs, and 4Ps are useful only when they match the decision. A people case such as Rob Parson at Morgan Stanley is not a market-sizing problem. A disaster-recovery analogue in public service is not a brand funnel. Choose the split that a manager in the room would recognise.

• Profitability: volume, price, mix, variable cost, fixed cost, one-off items.
• Growth: where to play, how to win, what to build, what to stop.
• Turnaround: cash, operations, customers, reputation, capital.
• Public service: users, reliability, cost recovery, equity, political feasibility.

Turn branches into workstreams.

Once the tree is drawn, assign owners. One person owns the quantitative branch, one owns customer or rider evidence, one owns competitive or institutional response, one owns the recommendation narrative. The tree then becomes a clock: after 90 minutes each branch must produce a claim, not a folder of screenshots.

Write the claim as a sentence. “Peak-hour crowding is the binding constraint, not average daily ridership.” That sentence can be true or false. A heading called “Market analysis” cannot.

Prune.

Students overgrow trees. If a branch cannot change the recommendation, cut it. If two branches lead to the same action, merge them. The final slide should still show the logic, but the live analysis should be ruthless about time.

Practise on the archive: build a tree for MTR fare policy, another for SK-II’s Japan-to-global question, another for Malaysia Airlines after a reputation shock. Compare the trees. The method is the same; the first split should not be.

Continue with case structure and market sizing.