Anatomy of a Profitability Case: A Full Walkthrough

Reading about frameworks teaches you the buckets. Watching a case played out turn by turn teaches you how the buckets actually work in real time — and where strong candidates diverge from average ones. This is a complete walkthrough of a profitability case at the difficulty level of an MBB round-2 interview.

The prompt

Inflection 1: Clarification (60–90 seconds)

The candidate's first move is to clarify scope. Strong candidates ask 2–4 specific questions; weak candidates either skip clarification or ask vague questions ("can you tell me more about the company"). For this prompt the high-value clarifications are:

• "You said same-store revenue is flat — has the store count grown over those 18 months, or is the network stable?"
• "Is the margin decline uniform across stores or concentrated in a subset?"
• "Are we benchmarking the 17%→11% decline against industry peers, or is this BrewHouse-specific?"

For this case, the interviewer responds: store count has grown from 540 to 605 over the period, the decline appears across most stores, and peer chains have seen smaller (1–2 point) declines.

Inflection 2: Structure (60 seconds of silence, then deliver)

Take 60 seconds. Write your structure on paper, top-down. Then deliver out loud, one branch at a time. Here is what a partner-level structure looks like for this case:

"I see three drivers I'd want to test: First, revenue per store — has price, mix, or volume per visit declined? Second, variable cost per drink — coffee bean cost, packaging, dairy, labor per cup. Third, fixed cost allocation — new-store buildout, headquarters overhead, underperforming new stores dragging the average. I'd start with revenue mix and cost-of-goods, since you mentioned peers are seeing smaller declines, suggesting an input-cost shock plus a BrewHouse-specific revenue mix issue."

Inflection 3: First drilldown — revenue mix

Interviewer: "Good. Let's start with revenue mix. Here is the data."

Strong candidates do two things here: (1) immediately compute the blended margin shift, (2) identify the mix story.

Quick math:

• Old blended GM: 0.60 × 76 + 0.22 × 70 + 0.12 × 38 + 0.06 × 55 = 45.6 + 15.4 + 4.6 + 3.3 ≈ 68.9%
• New blended GM: 0.48 × 76 + 0.27 × 70 + 0.19 × 38 + 0.06 × 55 = 36.5 + 18.9 + 7.2 + 3.3 ≈ 65.9%
• Mix-driven gross margin decline: ~3 percentage points.

Out loud: "On the data: blended gross margin moved from about 69% to 66% — a 3-point decline driven entirely by mix. Specialty coffee, the highest-margin category, lost 12 points of share. That mostly went into food, which is half the margin. Before drawing conclusions, I'd want to know whether specialty coffee volume actually declined or just share-of-mix declined."

Inflection 4: Second drilldown — cost

Interviewer: "Specialty coffee volume is roughly flat — food and brewed coffee volume grew faster. Now let's look at cost."

Quick math:

• Old per-drink contribution margin: ($5.40 − $2.35) / $5.40 ≈56%
• New per-drink contribution margin: ($5.40 − $2.90) / $5.40 ≈46%
• Per-drink contribution margin decline: ~10 percentage points

Out loud: "Per-drink contribution margin dropped 10 points — much larger than the 3-point mix effect. COGS is up 32% and labor up 18%. Bean prices industry-wide are up roughly 25% per public reporting, so most of the COGS increase is probably input-driven. The labor increase is harder to read without knowing whether it's wage rate or labor-per-drink inefficiency."

Inflection 5: Third drilldown — new-store drag

Interviewer: "Labor inefficiency has been a problem in newer stores — they take 12–18 months to reach steady-state labor productivity. Here's the new-store data."

Out loud: "So the new-store drag accounts for roughly 2 points of the margin decline on its own. Combined with the 3-point mix shift and the per-drink contribution margin pressure (mostly bean-cost driven), we have a picture: about half the margin decline is industry-wide cost pressure, and half is BrewHouse-specific from new-store cannibalization and adverse mix shift toward lower-margin food."

Inflection 6: Recommendation

Interviewer: "Good. The CEO is in the room. What do you recommend?"

Out loud: "I'd recommend a two-track plan. First, on the cost side: BrewHouse should hedge bean exposure with 12–24 month forward contracts — peer chains have done this and recovered 1–2 points of margin within a year. Second, on the revenue mix: rather than pushing back on food, which customers clearly want, BrewHouse should reformulate the food offering toward higher-margin items — pastries with gross margins in the 50–55% range — and bundle them with specialty drinks. That should recover ~1–2 points of blended margin without losing the food traffic. Third, on new stores: pause expansion for 6 months and focus on driving the underperforming cohort to steady-state. We can revisit unit economics for new openings once the existing base recovers.

Combined, this gets us back to ~14–15% EBITDA margin in 12–18 months. The key risk is bean prices continuing to rise — if the hedge does not work, we revisit pricing. That's a difficult conversation in this category, but a 5–10% price increase is defensible if peers move first."

What an interviewer writes on the rubric

The interviewer's rubric (paraphrased from a real MBB evaluation form) typically scores on four dimensions:

• Structure (1–5): Clear branches, prioritized, supported by reasoning. Above-bar = candidate explicitly names a hypothesis.
• Math (1–5): Arithmetic accuracy, framing of the question, sanity-checking. Above-bar = candidate states assumptions and corrects rounding.
• Communication (1–5): Top-down, concise, follows interviewer cues. Above-bar = candidate paces with the interviewer rather than driving.
• Recommendation (1–5): Specific, hedged correctly, includes risk. Above-bar = candidate names what would invalidate the recommendation.

For this case, a candidate who delivers the analysis above would typically score 4/5 or 5/5 on all four — a clear pass. A candidate who skipped clarification, used a generic revenue/cost tree, did the math without framing, and ended with "I would investigate further" would score 2/5 or 3/5 — a clear fail.

How to use this walkthrough

Read it twice. The first time, follow the logic. The second time, run the case yourself out loud — pretend you are the candidate, narrate each step, then check against the walkthrough. The gap between your version and this one is your actual prep target.