Consulting Case Study Examples: Real Firm Cases and Practice Cases With Worked Math
Where to read real McKinsey, BCG and Bain case studies, three summarized from the firms' own pages, plus four practice cases with the math worked out.
The best place to read real consulting case studies is the firms' own client-impact pages, because that's the only place where a firm's name, a client and a result are published together by someone accountable for them. If you're preparing for case interviews, read a few of those for business context, then spend most of your time on practice cases where you do the math yourself. This page gives you both: links to the published libraries, three real examples summarized exactly as the firms tell them, and four practice cases we wrote with the arithmetic worked out.
Soreno isn't affiliated with any firm named here.
Where to read real consulting case studies
Each of these pages is run by the firm itself:
- McKinsey case studies, a running list of named client stories across industries
- BCG client impact, BCG's library of client success stories
- Bain client results, Bain's collection of client stories
- Deloitte Stories of Impact
- Accenture reinvention client stories
- KPMG client stories
Read them with two things in mind. They're marketing: firms publish the projects that went well, and the client usually signed off on the wording. Many are also anonymized ("a global consumer durables company") or use made-up company names. If a blog post gives you a precise result for an unnamed client and doesn't link to the firm's page, treat it as unverified.
Three published cases, as the firms describe them
McKinsey and B.TECH (retail)
B.TECH is an Egyptian consumer electronics retailer. According to McKinsey's case study, the team narrowed more than 60 possible tech and AI use cases to five core initiatives on a two-year road map. The work covered web and app conversion, pricing, a fulfillment control tower and a new consumer-lending business called mylo. McKinsey reports about a 35 percent EBITDA uplift in the first year after the transformation, a 14 percent lift in conversion rates, a 4 percent gross profit uplift from AI-driven pricing, and e-commerce on-time delivery rising from 30 percent to 80 percent.
What to notice for interviews: the first real decision was prioritization, cutting 60 ideas to five. That's the same move an interviewer wants when you brainstorm 12 levers and then say which two you'd pursue first.
BCG and a consumer durables manufacturer (cost)
In BCG's write-up, the unnamed client had staff and product costs up to 20% higher than competitors. BCG ran a six-week diagnostic benchmarking costs, restructured overhead, moved component sourcing to lower-cost countries (potential savings of 8% to 10% in many categories) and redesigned products to strip features customers didn't value. BCG says the company is on track to realize over $1 billion in total cost-saving opportunities.
What to notice: the cost work was split into overhead, purchased materials and product design. That's a clean split you can reuse in any cost-reduction case.
Bain and "L-Box" (private equity due diligence)
On Bain's page, a buyout firm called L-Box (Bain says the names are changed but the results are real) wanted to buy a plastics packaging company, PackageMaster. Bain had four weeks for due diligence. L-Box made an early, preemptive bid that went unopposed. Bain reports EBITDA of 12% before the purchase and 28% two years later, and says EBITDA more than doubled.
What to notice: the question was whether the target could hold its lead in its core segment and compete in a new one. Diligence cases in interviews usually come down to the same thing: is the market growing, and can this company win in it.
How to turn a published case into practice
Published cases give you the answer and skip the reasoning, which is the part you're tested on. So reverse them. Write the client's problem as a one-line prompt, close the page, and spend five minutes building your structure. Then list the three or four numbers you'd ask for. Only after that, reopen the page and compare your levers with what the firm actually did. You'll find your weak spots fast: usually you forgot a cost bucket or never said which lever you'd pull first.
Four practice cases (fictional, written by us)
Everything below is invented for practice. The companies and numbers don't exist. Try each one before reading the solution.
Case 1: Coffee chain profit drop (profitability)
Prompt. A regional coffee chain with 40 stores earned $4.8M in profit last year on $48M in revenue. This year revenue was flat but profit fell to $2.4M. Why, and what should they do?
Structure. Profit = revenue minus costs. Revenue is flat, so go straight to costs: cost of goods, labor, rent and overhead.
Data you get when you ask. Cost of goods $14.4M both years. Rent and overhead $12.0M both years. Each store runs 40,000 staff hours a year, so 1.6M hours in total. The average wage rose from $10.50 to $12.00 an hour.
Math.
| Line | Last year | This year |
|---|---|---|
| Revenue | $48.0M | $48.0M |
| Cost of goods | $14.4M | $14.4M |
| Labor (1.6M hours) | $16.8M | $19.2M |
| Rent and overhead | $12.0M | $12.0M |
| Profit | $4.8M | $2.4M |
Labor rose $2.4M, which explains the whole drop.
Two fixes the client is considering:
- Cut 10% of hours in off-peak slots: 160,000 hours × $12 = $1.92M saved.
- Raise prices 5%, expecting 2% fewer transactions. Revenue becomes $48M × 1.05 × 0.98 = $49.39M, up $1.39M. Cost of goods falls with volume: $14.4M × 2% = $0.29M. Profit gain about $1.68M.
Recommendation. Start with off-peak scheduling. It recovers most of the gap and customers barely see it. Test the price increase in a handful of stores before rolling it out, because the 2% volume loss is a guess. If both work, profit reaches $2.4M + $1.92M + $1.68M = $6.0M.
Case 2: E-bike subscription launch (market sizing and break-even)
Prompt. An e-bike company wants to launch a $60-a-month subscription in a city of 3 million people. Should it?
Structure. How big is the market, what share can they win, and does that share cover the costs?
Market sizing.
- Adults 18 to 64: 65% of 3M = 1.95M
- Short commute and interested in a subscription: 5% = 97,500
- Share the client wins by year three: 10% = 9,750 subscribers
Economics. Revenue per subscriber is $60 × 12 = $720 a year. Bike depreciation and maintenance run $400 a year, so contribution is $320 per subscriber. Fixed costs (warehouse, staff, marketing) are $2.4M a year.
- Year-three profit: 9,750 × $320 = $3.12M contribution, minus $2.4M fixed = $0.72M
- Break-even: $2.4M ÷ $320 = 7,500 subscribers, about 7.7% of the 97,500 interested adults
Recommendation. Launch, but with a checkpoint. The business only works if it gets to 7,500 subscribers, which is 77% of the year-three target, so there isn't much room for error. Tell the client to track sign-ups against that 7,500 number and cut marketing spend if it's clearly out of reach after year one. More on this kind of math in our break-even analysis examples and market sizing questions.
Case 3: Warehouse capacity (operations)
Prompt. A distribution center ships 12,000 orders a day and demand will rise to 15,000. Should it add a night shift or buy a fourth packing line?
Current capacity. 3 lines × 2 shifts × 8 hours × 250 orders an hour = 12,000 orders a day. The gap is 3,000 orders, or 12 line-hours (3,000 ÷ 250).
Data. Day labor costs $300 per line-hour, night labor $360 (a 20% premium). A new line costs $1.2M. The center runs 250 days a year.
| Option | Daily cost of 12 extra line-hours | Yearly cost | Upfront |
|---|---|---|---|
| Night shift on existing lines | 12 × $360 = $4,320 | $1.08M | $0 |
| Fourth line on day shifts | 12 × $300 = $3,600 | $0.90M | $1.2M |
The new line saves $0.18M a year, so it takes $1.2M ÷ $0.18M, or about 6.7 years, to pay back.
Recommendation. Add the night shift. A 6.7-year payback is too long for a problem the night shift solves today. Revisit if demand keeps climbing: three lines running around the clock top out at 3 × 24 × 250 = 18,000 orders a day, so order the fourth line well before volume gets near that ceiling.
Case 4: Buying a pest control company (private equity)
Prompt. A private equity fund can buy a pest control company with $10M EBITDA for 9x, or $90M. It targets 2.5x its money in five years. Should it bid?
Deal math. The fund borrows $45M and puts in $45M of equity. Management's plan gets EBITDA to $14M in five years. Assume an exit at the same 9x multiple and debt paid down to $25M.
- Exit value: $14M × 9 = $126M
- Equity at exit: $126M minus $25M debt = $101M
- Return: $101M ÷ $45M = 2.24x, short of 2.5x
What would make it work. Two ways to reach 2.5x:
- Higher EBITDA. The fund needs $45M × 2.5 = $112.5M of equity at exit, so an exit value of $137.5M, which is $137.5M ÷ 9 = about $15.3M of EBITDA.
- Lower price. Keeping the $14M plan and $45M of debt, equity in must be $101M ÷ 2.5 = $40.4M, so the price is $85.4M, or about 8.5x.
Recommendation. Bid no higher than about $85M unless diligence finds a credible path to $15.3M of EBITDA. The case interview version of diligence is exactly the Bain question above: is the market growing, and can this company take share. Our M&A case study has a fuller deal example.
How to practice these
Do each case out loud and time yourself: two minutes to structure, then the math on paper without a calculator. Once you can get through them cleanly, move to cases where someone pushes back on your numbers, since real interviewers will. Our case interview examples and frameworks guide cover more formats, and there are firm-specific sets for McKinsey and Bain.
If you don't have a practice partner, Soreno runs full mock cases with an AI interviewer, either interviewer-led or candidate-led, asks follow-up questions, and gives you a feedback report at the end. It also has timed drills for structuring, market sizing, consulting math and chart reading. New accounts start with 2 free cases and 3 drills.