Struggling with structuring + final recommendations in case interviews

Hey there,

Thanks for your question! First of all: I think you are trying to solve for 3 different problems, so let me go step by step.

1. Structuring: drills can indeed help. Try not only to build a structure, but also to communicate it (i.e., start with naming the buckets, then get into each bucket's bullet, be strategic on what to say regarding each bullet -- some of them are more important than others depending on the industry/company context, and on some of them you might want to showcase). Record yourself and then play Devil's advocate: try to find at least 2-3 ME issues and 2-3 CE issues. You'll get better, we've all passed this phase :)

2. Case communication: # of solved cases matters, but you might need a coach. This is where the sheer number of solved case matters the most. Keep in mind that the communication for, say, qualitative questions follows different rules than, say, maths. 

If you are not a natural communicator and don't see improvements after 5-10 cases, that's the area where a coach might help the most -- I think 70% of the work I do with my candidates is about communicating better concepts/ideas that they already have in their minds.

3. Recommendation: just pick a structure and iterate. This is the easiest part. I suggest you start following a simple structure, see the example below:

• A two-line recommendation. Think of this as the title of the final recommendation slide (and yes: in MBB, slide titles go over two lines!)
• At least two points of supporting rationale. Ideally, one quantitative and one qualitative
• Risks and mitigations: impossible to be exhaustive, focus on the top 1-2
• Immediate next steps: here, you should be pragmatic -- what can they do in next week, and in the next month?

Good luck on your journey! Feel free to DM me to continue this conversation with a free 15-min intro call :)

Best,

Tom

_____________
Recommendation Example

Leading Sentence: MediLink should implement a new inventory strategy to release NWC and improve commercial performance to generate $X in additional profit.

Supporting Rationale:

• Quantitative: Increasing Product A's service level to YY% generates €X.0M profit against €0.XM in holding costs (a >10× return). Reducing Product C’s over-servicing simultaneously releases tied-up NWC.
• Qualitative: The current uniform policy misallocates capital by over-serving low-value Product C and starving high-value Product A. A differentiated model aligns Ops with Sales objectives. 

Risks and Mitigations: Key implementation risks are IT system changes and team change management. We recommend mitigating this through a phased rollout and clear day-one KPIs.

Immediate Next Steps:

• Next week: Reclassify SKUs to correct categories and define differentiated service-level targets.
• Next month: Test the new model on top Product A SKUs (with IT update), measuring KPIs weekly to validate the business case for full rollout.

Hi,

You’ve only done 2–3 cases, so honestly — this is exactly where you should be. The issue isn’t that you’re stuck, it’s that you’re trying to fix everything just by doing more full cases. That rarely works at this stage.

Let me share what actually helped me when I was in a similar situation.

On structuring — the problem is usually over-relying on frameworks. Most candidates try to “fit” the case into something like profitability or market entry, and it ends up messy.

A better way is to start from the question itself.
Before writing anything, ask yourself: what would need to be true for me to answer this question confidently?
Then build 3–4 buckets from that.

For example, “Should we enter this market?” naturally leads to:

• is the market attractive
• can we win
• what could go wrong

That’s already clean enough. You don’t need perfect MECE — just logical and non-overlapping.

What helped me improve fast was doing quick structuring drills. Take a bunch of case prompts and spend 5 minutes on each just building the structure. No math, no solving. That’s way more effective than full cases early on.

On communication — what you’re describing is very common. It’s not that your thinking is messy, it’s that you’re not externalizing it properly.

The fix is simple but feels unnatural at first: talk while you think.

Instead of going silent, say things like:
“Let me break this into two parts…”
“I’ll first estimate volume, then look at pricing…”

It doesn’t need to be perfect. It just needs to be structured. This alone makes a big difference.

On final recommendations — most people overcomplicate this.

A solid answer is always the same pattern:

• clear answer
• 2–3 reasons
• key risks
• next steps

That’s it.

For example: “I’d recommend entering the market. It’s growing fast, we have a cost advantage, and competition is fragmented. The main risk is regulation, so I’d validate that first. As a next step, I’d run a pilot in one region.”

If you hit those four elements, you’re already ahead of most candidates.

What helped me here was simply practicing 1-minute summaries after each case. Even better if you record yourself — you’ll immediately notice what’s missing.

If I had to simplify everything:

• structuring: build from the question, not memorized frameworks
• communication: slow down and signpost your thinking
• recommendation: follow a simple, repeatable format

You’re early, so don’t worry — these are very fixable with a bit of targeted practice.

If you want, I’m happy to do a quick session focused just on structuring and recommendations. That’s usually where people see the fastest improvement.

Best,
Soheil