Building an issue tree that survives the first fact

Most candidates can draw an issue tree. Far fewer can draw one that still makes sense after the interviewer hands over the first exhibit. The difference is whether the tree was built to answer the question or built to look complete.

Start from the question, not the framework

Take a simple prompt: a diagnostic centre in Mumbai has seen profits fall 20% in a year. The tempting move is to write down a profitability framework with every box you remember. The better move is to ask what would have to be true for profit to fall, and split along that.

Profit down 20%
├── Revenue fell
│   ├── Fewer patients
│   └── Less revenue per patient
│       ├── Fewer tests per visit
│       └── Lower price per test
└── Costs rose
    ├── Fixed (rent, equipment, staff)
    └── Variable (reagents, consumables)

Every branch is a yes-or-no question you can test with a fact. That's what makes a tree useful in a live case.

Three tests for a tree that holds up

Mutually exclusive

Could one cause sit in two branches at once? If "discounts" could appear under both price and promotions, the interviewer can't tell where your thinking is going, and neither can you.

Collectively exhaustive

Is there any way profit could have fallen that your tree doesn't cover? Revenue and cost together cover everything, which is why the profit split is such a dependable first cut.

Prioritised

Which branch would you test first, and why? Say it out loud. "Revenue first, because the centre's costs are mostly fixed and unlikely to have jumped 20% in a year" shows judgment, not just structure.

When the first fact arrives

Suppose the interviewer says footfall is flat but revenue per patient fell 18%. A good tree tells you exactly what to do next: close the "fewer patients" branch, go one level deeper on revenue per patient, and park costs for now. You don't need a new framework. You just walk down the one you already have.

In lowercase, this is literally what you see: branches you've justified turn blue, branches you haven't explored stay grey, and the next fact only unlocks when you ask the question that leads to it.

Practise the tree on its own

You don't need a full 30-minute case to practise structuring. Take ten prompts across different case types and give yourself two minutes to draw each tree. Check each against the three tests above. Then pick the best one and run it as a full case with a partner. Structure improves fastest when you practise it separately from everything else.