The Written Case Interview: From Data Pack to Three Slides

What is a written case interview?

A written case interview hands you a client problem and a pack of documents, typically 20-40 pages of charts, tables and articles, and gives you a fixed time to produce a short slide deck. You then present the deck to one or two interviewers, often a partner, and answer their questions. It compresses a week of project work into an hour or two.

Firms use it because it shows what a live case hides: whether you can find the relevant 20% of the data without help, calculate without an interviewer correcting you and write a conclusion someone else can read. The core skill is the one McKinsey's former written test, the PST, rewarded, which I took as a candidate: deciding quickly what not to read. A written case adds slides and a live defense on top.

Which firms use a written case, and in which round?

Mainly BCG and Bain in selected offices, plus several tier-2 firms. Formats change by office and year, so the labels below matter.

For the rest of each firm's process, see the BCG case interview guide and the Bain case interview guide.

How should you split the prep time?

Allocate roughly 10% to orientation, 40% to reading and analysis, 35% to slide building and 15% to rehearsal. For a 50-minute window that gives this plan:

Scale the plan for longer formats, but keep the order. Candidates who start slides before the analysis is finished rebuild them later.

Reading a 20-40 page data pack without drowning

Start from the question and work backward to the data. Write the client's decision at the top of your notes, then list the two or three numbers that would answer it. Scan exhibit titles and sort pages into core, supporting and ignore. Most packs include deliberate distractors: an industry article, a competitor profile or a chart that is accurate but irrelevant.

Read core exhibits fully, including units and footnotes. Exhibit traps are the same ones online tests set: thousands mixed with millions, percentages mixed with percentage points, and totals that do not match the sum of the segments. Our case interview math guide covers the calculator-free techniques that keep this step fast.

Building three slides: action titles, one chart, one recommendation

Each slide should carry one message, written as a full-sentence action title that states the finding. The body holds the single chart or table that proves it. The final slide states the recommendation, the expected impact, the main risk and next steps. This is the Minto Pyramid Principle applied to a deck: the conclusion leads, the evidence supports.

Presenting and defending under partner questions

Open with the answer in one sentence, then walk through the slides in two to three minutes each. Partners will probe your weakest number, ask what you would do with more time and test whether you will change your mind under pressure. Defend conclusions backed by data, concede quickly when they are not, and say how the concession changes the recommendation.

Practice mini data pack: a Dutch bakery chain

The client is a fictional Dutch bakery chain with 60 stores. Its operating margin fell from 9.0% to 4.0% in two years. The CEO asks why, and what to do. Give yourself 50 minutes without a calculator.

Exhibit 1: P&L, EUR millions

Exhibit 2: Store formats, 2025

Exhibit 3: Since 2023, butter prices rose 45%, flour 25% and eggs 30%. The average ticket rose 5%.

Model answer: three slides

Slide 1. "Ingredient inflation drives the margin drop: ingredient costs grew 22% while revenue grew 5%." Chart: a margin bridge from 9.0% to 4.0%. Ingredients moved from 30.0% to 34.9% of revenue (minus 4.9 points), staff from 40.0% to 42.1% (minus 2.1), rent minus 0.2, and other costs improved by 2.2 points. The verification step: the four bars must sum to the 5.0-point drop, and they do.

Slide 2. "The 15 supermarket-attached stores lose money while station kiosks earn 18%." Chart: contribution margin by format. Supermarket stores lose about EUR 0.2 million in contribution; kiosks earn about EUR 1.6 million on less revenue.

Slide 3. "Raise prices about 4% to lift the margin to about 7.7%, and renegotiate supermarket leases before deciding to exit." A 4% price increase at stable volume adds about EUR 2.0 million, giving EUR 4.0 million of profit on EUR 52.4 million of revenue. At the new prices, supermarket stores turn slightly positive (about EUR 0.2 million, or 2% contribution), so closing them would now cost profit; better lease terms come first. Main risk: volume loss at high-street stores. Next step: test the price increase in five stores for six weeks.

The trap in this pack sits in slide 3. A candidate who recommends closing the supermarket stores without recalculating their margin at the new prices puts a wrong conclusion on the final slide.

Cracking the Case includes a further written-case sample question. For timed practice on the core skill, reading dense exhibits fast and checking numbers without a calculator, The Key to the PST offers 26 questions across several business cases with a full answer guide, built on the same skills as the PST guide describes.