Case Interview Examples 2026: 43 on the Firms' Own Sites

The 43 case interview examples, firm by firm

Twenty-one plus ten plus six plus four plus two is 43, verified by addition on 5 October 2026. Kearney announces one more, a promotional planning case, whose content is not present in the archived snapshot of its page, so 44 are announced and 43 readable.

BCG's 13 sit outside that sum on purpose. Adding them gives 56, and a single figure of 56 would mean two things at once: 43 cases the firms serve themselves, and 13 modules BCG writes and a third party serves. The next section gives them their own line and their own wording.

Two counting notes, because the numbers are easy to inflate. Deloitte's tool shows three more titles in its filter menus with no page behind them. Those URLs return 404, so ten is the honest figure. The Accenture count comes from the list inside its workbook, which splits 13 and 8.

These 43 are more practice material than most candidates will use. Reading one carefully is worth more than skimming ten.

BCG's 13 modules, and why they sit apart

Until 5 October 2026 this page put BCG at zero. That was wrong. The firm announces the material from its own case interview preparation page, under "Put Your Skills to the Test": "Experience a day in the life of a consultant with OneDay@BCG, a free job simulation that provides insights into the challenges we solve."

Two facts go together here and each one alone misleads. The content is written by BCG and announced from the firm's own page. It is served by a third party, theforage.com, where the modules and their files actually sit.

The firm's virtual experience page gives the scale: "With 13 unique modules to choose from" and "Enroll today and join the 700,000+ participants". The strategy module is a full case in seven tasks. It opens "Today, you'll help us by stepping into the shoes of a BCG consultant to support our client, BeyondTech, a global technology and telecommunications leader" and asks "Is installment financing a viable option for BeyondTech?". Its attachments open without an account: a three page PDF of income statements and balance sheets, a workbook and a slide template. The payload carries Model Answer blocks and multiple choice questions with their corrections. At least eight of the thirteen modules name a fictional client, among them ClothingCo, RetailCo, CoffeeCo, PowerCo, SuperCo and NewCo.

These modules are the closest thing BCG publishes to a case interview example, and three things the firm does not say are the reason for a separate line. It never calls these modules case interviews, and never says they resemble its selection case. The pages print "No grades · No assessments". And the firm attaches a clause: "This virtual experience is not an employment relationship with BCG", followed by "It is our policy that you do not list your engagement as employment, work experience or an internship with BCG on your CV, LinkedIn profile, or any other social media or professional networking platform."

Which case interview cases to use, and what each adds

The cases differ less than the answer formats do. The answer format is what makes a case useful on your own.

McKinsey states what its four contain, on its interviewing page at mckinsey.com: "These cases contain logic, thought processes, and suggested answers." So you get the reasoning as well as the number. Beautify, Diconsa, Electro-Light and Talbot Trucks are the McKinsey case interview examples the firm itself serves, and the count above stops at four because no fifth sample McKinsey case appears on its own pages.

Bain serves two Bain case interview examples, Coffee Shop Co. and FashionCo., and publishes two tiers per question, labeled "A good answer" and "A better answer". That structure is itself the most useful thing on the page, because it shows you what separates adequate from strong on the same question.

Deloitte serves ten Deloitte case study examples, each publishing a suggested answer under every question and hints that name a structure. On the retail case below, its suggested answer for the first question is filed under Internal factors and External factors, which is the firm recommending a structure inside its own hint. Its tool states "Estimated completion time: 15-20 minutes" per case and that "This is not an assessment and we do not track your responses or results."

Oliver Wyman publishes something no other firm does: evaluation criteria printed under each individual question, and they change from question to question within one case.

Accenture publishes two fully worked examples inside its workbook, one sizing and one prepared-meals case, with the method spelled out as unit margin, fixed cost, division, then market share at breakeven.

So if you want graded tiers, use Bain. If you want criteria, use Oliver Wyman. If you want volume with answers, use Deloitte. If you want reasoning narrative, use McKinsey.

A published case worked to a 550 million dollar gap

Deloitte's Recreation Unlimited case, extracted from its page data at usrecruiting.deloitte.com, gives enough published data to produce a real claim.

What the firm publishes about the company. Page checked 30 September 2026. Founded in 1952 in the United Kingdom. Eleven billion dollars of global revenue. 32,800 employees. Two hundred owned stores. Five percent of sales online against an industry average of ten percent. Strongest market share in Europe. High turnover among younger staff, falling satisfaction, a highly hierarchical organization, and training spend halted some years ago.

What the firm publishes about the engagement. "The client has engaged Deloitte to drive top-line growth through eCommerce, specifically targeting an aggressive +40% annual growth in direct-to-consumer digital channels over the next five years."

Now the arithmetic, which is ours because the answer tables on that page are images. Recomputed by hand on 30 September 2026.

• Current online revenue: 5 percent of 11 billion dollars is 550 million dollars.
• At the industry average: 10 percent of 11 billion is 1.1 billion dollars.
• The gap to parity: 1.1 billion minus 550 million is 550 million dollars, so reaching the industry average doubles online revenue.

State that as one sentence and the case has a spine. The client is half as digital as its industry, and closing that gap alone is 550 million dollars of revenue moved into a channel, which is 5 percent of total revenue.

The five-year target implies 26.9 percent online

The published target is 40 percent annual growth for five years, and compounding it is the calculation the case is built around.

• 1.4 squared is 1.96.
• 1.96 squared is 3.8416, which is 1.4 to the fourth.
• Times 1.4 is 5.37824, which is 1.4 to the fifth.
• Applied to 550 million dollars: 550 times 5.37824 is 2,958 million, so about 2.96 billion dollars.

Recomputed by hand. Against 11 billion dollars of current revenue, 2.958 over 11 is 26.9 percent.

That figure is the so what, and it is worth saying carefully. If total revenue stayed flat, hitting the published target would put 26.9 percent of sales online, which is nearly three times the current industry average of 10 percent. Recomputed: 26.9 over 10 is 2.69.

So the target goes far past catching up. It is a plan to become the most digital player in the category by a wide margin. A candidate who says that in the first two minutes has read the prompt better than one who starts listing e-commerce levers.

Two caveats to say out loud, because both are real. Total revenue will not stay flat over five years, so 26.9 percent is the share only under that assumption. And a 40 percent compound rate for five consecutive years is a rate very few businesses sustain, which makes the feasibility question part of the answer rather than an aside.

The organizational data on the same page is there for a reason. High turnover among younger staff, falling satisfaction, a hierarchical structure and training spend stopped some years ago. A firm that halted training and cannot keep junior staff is short of the people who build a digital channel. The fifth question asks about workforce implications. Its suggested answer is filed under change management, then talent and organizational structure.

Where a published answer and its exhibit disagree

One published case has an internal gap, and reporting it honestly is more useful than smoothing it over.

Bain's coffee shop case at bain.com states: "Assume that Cambridge has a population of 100,000 people and, on average, each drinks 1 cup of coffee per day." That is 100,000 times 365, which is 36,500,000 cups a year. Its recommendation instead works from a 7,000,000 cup market, and the narrowing steps sit inside an image we could not read.

Both readings are arithmetically sound and they disagree. Its published breakeven is 204,675 cups. Against 7,000,000 that is 2.92 percent of the market. Against 36,500,000 it is 0.56 percent. Recomputed by hand: 204,675 over 7,000,000 is 0.02924, and 204,675 over 36,500,000 is 0.00561.

We are not going to decide which the firm intended. What a candidate should take from it is a habit: whenever you quote a share, say which base it is taken on. A candidate working only from the stated text of that case will not reproduce the published share, and saying which denominator you used protects you.

The same page publishes the assumption principle that makes this acceptable: "When making assumptions about market size, there is no 'right' or 'wrong' answer." The assumption is free. Naming the base is not optional.

What case interview cases do not contain

Six absences, and knowing them stops you looking for something that is not there.

1. No weights or scores. Not one of the 43 carries a marking scheme. BCG's modules carry Model Answer blocks and corrected multiple choice questions, and the pages that hold them print "No grades · No assessments".
2. No time per question. Deloitte gives a 15 to 20 minute total for a written case, and no firm publishes a per-question budget.
3. No readable exhibits in several cases. Bain's fashion tables, Deloitte's LeadAuto data sheet and Bain's coffee market derivation are all images, and we reproduce no figure from any of them.
4. No difficulty labels outside Oliver Wyman's six.
5. No indication that these are the cases you will be given. Accenture writes that cases are often drawn from the interviewer's own project experience.
6. No spoken version. All 43 are written, and the exercise they prepare you for is spoken.

That last gap is the largest. A written case read silently rehearses none of the narration, and narration is where the constraint actually bites.

How to work one so it counts

The weights below are ours, read from the grading code in src/lib/grading. No consulting firm publishes a marking scheme of its own.

A reading order

Eight cases in a sensible sequence, all free.

1. Oliver Wyman's Urban Grocers. The only case labeled easy, on a truck fleet decision.
2. Bain's Coffee Shop Co. A sizing and a threshold, with two answer tiers and a base problem to spot.
3. McKinsey's Electro-Light. A launch with a thin unit margin and a bottles-to-gallons conversion.
4. Deloitte's Recreation Unlimited. Retail growth with a compounding target, worked above.
5. Oliver Wyman's Dairy Farm. For the criteria printed under each question.
6. Deloitte's LeadAuto. For the good answer against great answer split, and a weighted average question.
7. Oliver Wyman's Oil Gas Price Strategy. A profitable entry refused on payback.
8. One of Deloitte's Valuation or Cyber cases. Because they use vocabulary the other seven do not.

Our math drill and chart drill score the components with the first correction named, and the case library collects the prompts we use in our own exercises. All are free and none needs an account. None of them replaces reading a firm's own case on the firm's own site, which costs nothing.

For the taxonomy these cases fall into, see case interview types. For the questions inside them, see case interview questions. For the criteria one firm prints beside them, see how cases are scored.

What changed on this inventory, and when

We correct in public and we date the correction. The body above carries the count as it stands today.

5 October 2026. BCG at zero was false, and the phrase "all on their own sites" was loose. The page said firms publish 43 readable cases with BCG absent from the list. BCG announces OneDay@BCG from its own case interview preparation page, 13 modules with 700,000+ participants, and its strategy module is a complete case in seven tasks on a named fictional client, with statements to download and Model Answer blocks inside. The nuance is published whole: the content is written by BCG and it is served by a third party, theforage.com. Either half alone replaces one error with another.

5 October 2026. The count was not recomputed at a guess. The 43 stands for the cases the firms serve on their own sites and the addition was redone by hand: 21 plus 10 plus 6 plus 4 plus 2 is 43. BCG's 13 are a separate line because they are hosted elsewhere and because the firm does not call them cases. A merged figure of 56 would hide both of those facts, so this page does not print one. The title and the summary carried the word all, which claimed completeness the page could not support; both have been rewritten.