What is an issue tree in a case interview?

An issue tree is a visual decomposition of a case interview problem into MECE sub-problems. You draw it during the structure phase of the case to break a big, ambiguous question (e.g., 'why is profit down?') into specific, testable branches (revenue vs cost → price, volume, mix, etc.).

How is an issue tree different from a framework?

Frameworks (Profitability, Porter's 5 Forces, 4Ps) are generic templates. Issue trees are case-specific - you build them on the spot from the prompt. A framework can inspire your top-level split, but your tree should always be tailored to the actual case facts.

How MECE does an issue tree need to be?

Strictly MECE at the top level - if level 1 overlaps or has gaps, the whole analysis breaks. Level 2 and level 3 can be slightly softer in practice (interviewers understand you're working under time pressure), but mutual exclusivity at the top is the hard bar.

Should I draw the issue tree or keep it mental?

Draw it. Even in voice-only interviews, sketch it on paper for yourself. Candidates who verbalize a clear tree ('I see two buckets - revenue and cost - and under revenue I want to look at price, volume, and mix…') consistently score higher than candidates who think it through silently.

What's the best way to practice issue trees?

Quantity plus feedback. Start with the guided structure decisions on this page, then continue free to the real drill room and write your own tree for AI feedback on whether the top-level split is MECE and case-specific.

Issue Tree Examples: 10 Case Interview Structures Explained

Free · Issue Tree Drill

Ten fully worked issue tree examples for the most common case interview problems - profitability, market entry, M&A, pricing, and more. Each tree is MECE, explained step by step, and tied to the hypothesis you should test first.

Should a European coffee roaster enter the US ready-to-drink market?

Which of these is the strongest opening structure?

The quick version: An issue tree (or logic tree) is a visual decomposition of a case interview problem into mutually exclusive, collectively exhaustive (MECE) sub-problems so you can find the driver instead of guessing.

What makes an issue tree work

Interviewers do not grade your issue tree on how pretty it looks - they grade whether the top-level split actually tests the right hypothesis. A good issue tree has three properties working together.

MECE is non-negotiable. Every branch must be mutually exclusive (no overlap) and collectively exhaustive (no gaps). "Revenue vs cost" is MECE for profit. "North America vs retail" is not - you're mixing geography with channel.

The top-level split is your hypothesis. If you split profit decline into revenue vs cost, you're saying "I believe the answer lives on one of those two sides - let me find out which." That's issue tree drill thinking, and it's the highest leverage step in case interview structure practice.

Depth caps at three levels. Most 30-minute cases use a two-level tree with a few leaf investigations. We see 70%+ of candidates try to pre-build a five-level tree and burn their first 3 minutes of case time - don't.

Finally, an issue tree is not a framework. Frameworks like Profitability or Porter's are generic templates. An issue tree is case-specific: you build it on the spot from the prompt. Use a framework as inspiration for the top-level split, then branch out from there.

How to build an issue tree in 2 minutes

This is the 2-minute process top candidates use. It sounds simple because it is - the reason most candidates fail at it is that they skip step 1.

Step 1 - Restate the problem precisely. Write the root node as a specific, quantified question: "Why did $500M CPG client's profit drop 15% YoY?" - not "Profit problem." Specificity forces MECE downstream.

Step 2 - Pick your decomposition dimension. Profit → revenue vs cost. Growth → organic vs inorganic. Market share loss → our performance vs competitor vs market. One clean dimension per split - no mixing.

Step 3 - Branch level 1 into 2–4 MECE pieces. Two branches is the cleanest; four is the ceiling. If you need five, your split dimension is wrong.

Step 4 - Add level 2 under each level 1 branch. Drill down using the same discipline. Revenue becomes Price × Volume × Mix. Cost becomes Fixed + Variable.

Step 5 - Mark leaf nodes as "to investigate" and pick your first hypothesis. This is hypothesis-driven problem solving: don't investigate every leaf. Tell the interviewer which branch you suspect and why, then ask for data.

That's the full issue tree practice loop. The 10 examples below show you what the finished product looks like for each problem type.

10 real issue-tree drills

These are active Easy drills from the real curriculum. Open any case in its normal set, build the issue tree yourself, and get the same feedback as every other app drill.

Train issue trees with AI feedback

Build one with live AI feedback

5 mistakes that tank your issue tree

The same issue tree drill failures show up over and over in mock interviews. Here's the shortlist.

Overlapping branches (not mutually exclusive)
Branches share items - e.g., "revenue, cost, and operations." Fix: pick one clean dimension (math, lever, or customer segment) and split along it only.
Missing branches (gaps - not collectively exhaustive)
Leaving out a critical dimension - e.g., market entry without execution cost. Fix: ask "could something live outside these branches?" If yes, add a branch.
Splitting by symptom instead of driver
"Sales are down because fewer customers and lower spend" is a symptom description. Fix: split by acquisition vs retention vs pricing - the levers you can actually pull.
Going four or five levels deep on a 30-minute case
Over-engineering eats your clock. Fix: cap at level 3 during structure, then only drill further when the interviewer asks for it.
Using a generic framework as your issue tree
Porter's 5 Forces, 4Ps, or 3Cs are not MECE for every problem. Fix: build a case-specific tree; borrow from frameworks only for the top-level split inspiration.

Issue tree examples - FAQ

What is an issue tree in a case interview?
An issue tree is a visual decomposition of a case interview problem into MECE sub-problems. You draw it during the structure phase of the case to break a big, ambiguous question (e.g., 'why is profit down?') into specific, testable branches (revenue vs cost → price, volume, mix, etc.).
How is an issue tree different from a framework?
Frameworks (Profitability, Porter's 5 Forces, 4Ps) are generic templates. Issue trees are case-specific - you build them on the spot from the prompt. A framework can inspire your top-level split, but your tree should always be tailored to the actual case facts.
How MECE does an issue tree need to be?
Strictly MECE at the top level - if level 1 overlaps or has gaps, the whole analysis breaks. Level 2 and level 3 can be slightly softer in practice (interviewers understand you're working under time pressure), but mutual exclusivity at the top is the hard bar.
Should I draw the issue tree or keep it mental?
Draw it. Even in voice-only interviews, sketch it on paper for yourself. Candidates who verbalize a clear tree ('I see two buckets - revenue and cost - and under revenue I want to look at price, volume, and mix…') consistently score higher than candidates who think it through silently.
What's the best way to practice issue trees?
Quantity plus feedback. Start with the guided structure decisions on this page, then continue free to the real drill room and write your own tree for AI feedback on whether the top-level split is MECE and case-specific.

Timed drill

Practice issue tree examples with feedback

Get one scored rep, see what worked, and fix the weak spot before your interview.

Start a structure drill nowFirst attempt is free.