Practice cases: Japan and South Korea
Seven full cases: a rural railway in Japan, a Korean skincare brand growing in the US, a private-equity deal for a Japanese parts maker with no successor, a trucking firm under Japan's overtime cap for drivers, a Seoul gym chain's discount, merging two rural schools in Japan, and a Korean meal kit deal.
Key takeaways
- Where the structure comes from: it is built from the goal of this exact question (Loss = passengers x fare - operating cost), not taken from a list.
- Where the structure comes from: it is built from the goal of this exact question (Compare three-year contribution, then test how sure each option is), not taken from a list.
- Where the structure comes from: it is built from the goal of this exact question (Value today, value at exit, and the fund's return), not taken from a list.
Seven full cases: a rural railway in Japan, a Korean skincare brand growing in the US, a private-equity deal for a Japanese parts maker with no successor, a trucking firm under Japan's overtime cap for drivers, a Seoul gym chain's discount, merging two rural schools in Japan, and a Korean meal kit deal.
Cover the solution and run each case out loud, ideally with a partner playing the interviewer. Ask your own clarifying questions, state a hypothesis, build a structure from the maths of the goal (not from a memorised list), and do the math on paper before you look. Then compare your synthesis with the one given, and read the strong and weak candidate notes. Each case is labeled Starter, Standard, or Stretch.
Case 1: Kitase Line Railway: a rural line's loss nearly doubled
Where the structure comes from: it is built from the goal of this exact question (Loss = passengers x fare - operating cost), not taken from a list. Each branch is one driver of that goal, and the hypothesis above says which branch to test first.
Worked case
Starter: Kitase Line Railway: a rural line's loss nearly doubled
The prompt
The Kitase Line is a 40-kilometer rural railway in Japan. Its yearly loss has nearly doubled in five years, and the prefecture that covers the loss asks what it should do. The exhibit shows the key figures.
Difficulty: Starter. Format: interviewer-led, with an exhibit. Industry: Public transport. Region: Japan. Interview length: about 25 minutes. The company is fictional and all figures are illustrative.
Clarifying questions, with the interviewer's answers
- Who owns the line?Answer: A company owned jointly by the prefecture, the towns along the line, and local businesses. The prefecture covers the loss each year.
- Why are passengers falling?Answer: The towns are losing population, and most remaining riders are high-school students and older people.
- Can costs be cut on the railway itself?Answer: Only a little. Most of the JPY 1 billion a year is track, signals, and staff that are needed whatever the number of passengers.
A hypothesis to say out loud: Costs are mostly fixed and passengers are falling, so my hypothesis is that fare rises cannot close the gap and the prefecture will need to change how the service is run.
The structure
- Loss = passengers x fare - operating cost
- Key: Passengers: population and who rides
- Fare
- Operating cost, mostly fixed
- Other ways to run the service
The exhibit
| Measure | Five years ago | This year |
|---|---|---|
| Passengers a year (millions) | 2 | 1.6 |
| Average fare per trip (JPY) | 400 | 400 |
| Operating cost (JPY billion a year) | 1 | 1 |
Working it through
1. Revenue five years ago
2.0 million passengers at JPY 400.
Revenue five years ago (JPY):2,000,000 × 400 = 800,000,0002. Revenue this year
1.6 million passengers at JPY 400, 20 percent fewer.
Revenue this year (JPY):1,600,000 × 400 = 640,000,0003. Loss this year
Revenue minus JPY 1 billion of operating cost. Five years ago the loss was JPY 200 million.
Result this year (JPY):1,600,000 × 400 - 1,000,000,000 = -360,000,0004. Cost per passenger
Each trip now costs this much to provide, against a JPY 400 fare.
Cost per passenger (JPY):1,000,000,000 ÷ 1,600,000 = 6255. Raise fares 10 percent
Interviewer: "The line expects 3 percent fewer passengers after a 10 percent fare rise." Revenue becomes:
Revenue after the fare rise (JPY):1,600,000 × (1 - 0.03) × 400 × 1.1 = 682,880,0006. Curveball: switch to buses
Interviewer: "The prefecture could turn the track bed into a road only for buses, called bus rapid transit, a model some regional lines in Japan have used. It would cost JPY 550 million a year to run, and passengers would fall a further 10 percent."
Result with bus rapid transit (JPY):1,600,000 × 0.9 × 400 - 550,000,000 = 26,000,0007. Improvement against the railway
The bus result minus the railway result.
Yearly improvement (JPY):(1,600,000 × 0.9 × 400 - 550,000,000) - (1,600,000 × 400 - 1,000,000,000) = 386,000,0008. Payback of the conversion
Interviewer: "Converting the track bed would cost about JPY 2 billion once." Payback on the yearly improvement:
Payback (years):2,000,000,000 ÷ ((1,600,000 × 0.9 × 400 - 550,000,000) - (1,600,000 × 400 - 1,000,000,000)) = 5.18
What the exhibit shows
Costs stayed the same while passengers fell by a fifth, so every lost passenger adds to the loss.
The recommendation
The prefecture should study turning the line into bus rapid transit, rather than rely on fare rises. First, the loss grew from JPY 200 million to JPY 360 million a year because passengers fell 20 percent while costs stayed fixed; each trip now costs JPY 625 to provide against a JPY 400 fare. Second, a 10 percent fare rise adds only about JPY 43 million of revenue and hits students and older riders hardest. Third, bus rapid transit would turn the loss into a small surplus of about JPY 26 million, an improvement of about JPY 386 million a year, and would repay a JPY 2 billion conversion in about 5 years, even with 10 percent fewer riders. Hold public meetings early, because many residents see the railway as part of their town, and set bus times around school hours.
Risks: Residents may oppose losing the railway; Passengers could fall by more than 10 percent after the switch.
Next steps: Count passengers by station and hour to design the bus timetable; Ask national and prefectural offices which grants could fund the conversion.
A strong candidate
Showed that costs are fixed, compared the fare rise with a change in the service model, and thought about residents and students.
A weak candidate
Recommended a large fare rise to close the gap, without asking who the riders are or how many would stop riding.
Case 2: Gyeolhwa Skin Lab: online or a big retailer in the US?
Where the structure comes from: it is built from the goal of this exact question (Compare three-year contribution, then test how sure each option is), not taken from a list. Each branch is one driver of that goal, and the hypothesis above says which branch to test first.
Worked case
Standard: Gyeolhwa Skin Lab: online or a big retailer in the US?
The prompt
Gyeolhwa Skin Lab, a South Korean skincare brand, wants to grow in the United States. It can focus on online marketplaces or sign with a large US beauty retailer. Which should it choose? Figures are in Korean won (KRW) billions unless stated otherwise.
Difficulty: Standard. Format: candidate-led, with interviewer dialogue. Industry: Consumer goods (beauty). Region: South Korea and the US. Interview length: about 30 minutes. The company is fictional and all figures are illustrative.
Clarifying questions, with the interviewer's answers
- What does each option bring in by year three?Answer: Online marketplaces: sales of about KRW 30 billion a year. The retail chain: about KRW 50 billion a year across 1,000 stores.
- What are the margins?Answer: Online, gross margin after marketplace fees and delivery is 45 percent of sales, and advertising costs 20 percent. In retail, gross margin after the retailer's share is 35 percent, in-store marketing costs 10 percent, and a reserve for unsold stock that the retailer can return costs 5 percent.
- Any one-time costs?Answer: Retail needs KRW 6 billion up front for listing fees and displays. Online needs very little up front.
- Over what period should I compare?Answer: Three years.
A hypothesis to say out loud: Retail brings more sales but lower margins and an upfront cost. My hypothesis is that the two options are close over three years, so the answer depends on how sure we are about retail sales.
The structure
- Compare three-year contribution, then test how sure each option is
- Online contribution = sales x (gross margin - advertising)
- Retail contribution = sales x (gross margin - marketing - returns) - upfront cost
- Key: Sales retail needs to win
- Retailer rules and a smaller test
Working it through
1. Online, yearly contribution
Candidate: "KRW 30 billion at 45 percent margin minus 20 percent advertising."
Online contribution (KRW billion a year):30 × (0.45 - 0.2) = 7.52. Retail, yearly contribution
Candidate: "KRW 50 billion at 35 percent minus 10 percent marketing and 5 percent for returns."
Retail contribution (KRW billion a year):50 × (0.35 - 0.1 - 0.05) = 103. Online over three years
Candidate: "No real upfront cost."
Online, three years (KRW billion):3 × 30 × (0.45 - 0.2) = 22.54. Retail over three years
Candidate: "Minus the KRW 6 billion up front. Retail wins, but only by KRW 1.5 billion."
Retail, three years (KRW billion):3 × 50 × (0.35 - 0.1 - 0.05) - 6 = 245. Sales retail needs to match online
Candidate: "Retail must earn KRW 22.5 billion plus its upfront cost over three years at a 20 percent margin."
Break-even retail sales (KRW billion a year):(3 × 30 × (0.45 - 0.2) + 6) ÷ (3 × (0.35 - 0.1 - 0.05)) = 47.56. Curveball: the retailer's rule
Interviewer: "The retailer drops brands that sell less than KRW 40 million per store a year." Candidate: "Across 1,000 stores, that line is, in KRW billion:"
Delisting line (KRW billion a year):1,000 × 40 ÷ 1,000 = 407. Retail if sales land on that line
Candidate: "If retail sales are only KRW 40 billion, three-year contribution is:"
Retail at KRW 40 billion, three years (KRW billion):3 × 40 × (0.35 - 0.1 - 0.05) - 6 = 188. A smaller test
Candidate: "A test in 200 stores would need only this much of the upfront cost:"
Upfront cost of a 200-store test (KRW billion):6 × 200 ÷ 1,000 = 1.2
The recommendation
I recommend that Gyeolhwa start with online marketplaces and test retail in 200 stores before signing for all 1,000. First, over three years online earns about KRW 22.5 billion and retail about KRW 24 billion after its KRW 6 billion upfront cost, so retail is only slightly better. Second, retail wins only if sales reach about KRW 47.5 billion a year, close to the KRW 50 billion forecast, which leaves almost no room for error. Third, if sales per store sit at the KRW 40 million delisting line, retail earns only about KRW 18 billion. The 200-store test costs about KRW 1.2 billion.
Risks: The retailer may not offer the same terms later; Marketplace advertising costs may rise as more Korean brands compete online.
Next steps: Ask the retailer for a 200-store test on the same terms; Track repeat-purchase rates on the marketplaces for six months.
A strong candidate
Compared both options over the same period, found the sales retail needs, and used the retailer's rule to show how thin the margin for error is.
A weak candidate
Chose retail because its sales are larger, without subtracting the upfront cost or the returns reserve.
Case 3: Should a fund buy Tsukimine Precision, a parts maker with no successor?
Where the structure comes from: it is built from the goal of this exact question (Value today, value at exit, and the fund's return), not taken from a list. Each branch is one driver of that goal, and the hypothesis above says which branch to test first.
Worked case
Stretch: Should a fund buy Tsukimine Precision, a parts maker with no successor?
The prompt
Written case: a private-equity fund is considering buying Tsukimine Precision, a family-owned maker of precision metal parts in central Japan. Using the data pack below, prepare three slides: is the deal attractive at the asking price, what are the main risks, and what should the fund offer? Money figures are in JPY billion.
Difficulty: Stretch. Format: written case, with a data pack. Industry: Private equity and manufacturing. Region: Japan. Interview length: about 45 minutes. The company is fictional and all figures are illustrative. In a written case you usually get 30 to 60 minutes with a data pack, then present your slides and answer questions.
Clarifying questions, with the interviewer's answers
- Why is the company for sale?Answer: The founder is 74 and has no family member or manager ready to take over. Many mid-sized Japanese companies face the same succession question.
- What return does the fund need?Answer: About 20 percent a year over five years, which is about 2.5 times the money.
- Who are the customers?Answer: About 40 percent of revenue comes from one car maker. The rest is spread across machinery makers.
A hypothesis to say out loud: The price looks modest at about 6 times EBITDA, but the plan depends on margin gains and one large customer. My hypothesis is that the deal returns a bit below target at the asking price, and that customer concentration is the key risk.
The structure
- Value today, value at exit, and the fund's return
- EBITDA today and the entry multiple
- Plan: revenue growth and margin gains
- Key: Exit value, MOIC, and IRR
- Risks: largest customer, founder handover
The exhibit
| Item | Value |
|---|---|
| Revenue today (JPY billion) | 12 |
| EBITDA margin today (%) | 12.5 |
| Asking enterprise value (JPY billion) | 9 |
| Debt available at entry (JPY billion) | 4.5 |
| Revenue in year 5 of the plan (JPY billion) | 14 |
| EBITDA margin in year 5 of the plan (%) | 15 |
| Debt at exit after paydown (JPY billion) | 2 |
| Expected exit multiple (EV/EBITDA) | 6 |
| Share of revenue from the largest customer (%) | 40 |
Working it through
1. EBITDA today
JPY 12 billion of revenue at a 12.5 percent margin.
EBITDA today (JPY billion):12 × 0.125 = 1.52. Entry multiple
Asking price divided by EBITDA.
Entry multiple (EV/EBITDA):9 ÷ (12 × 0.125) = 63. Year-5 EBITDA
JPY 14 billion of revenue at a 15 percent margin.
Year-5 EBITDA (JPY billion):14 × 0.15 = 2.14. Exit value
Year-5 EBITDA at 6 times.
Exit EV (JPY billion):14 × 0.15 × 6 = 12.65. MOIC
Equity at exit (exit value minus JPY 2 billion of debt) divided by equity invested (JPY 9 billion minus JPY 4.5 billion of debt).
MOIC:(14 × 0.15 × 6 - 2) ÷ (9 - 4.5) = 2.366. IRR check
1.19 to the power of 5 is about 2.39, a little above the MOIC, so the IRR is a little under 19 percent a year.
1.19 to the power of 5:1.19 × 1.19 × 1.19 × 1.19 × 1.19 = 2.397. Price that meets the target
For 2.5 times the money, equity invested must be JPY 10.6 billion divided by 2.5, plus the JPY 4.5 billion of debt.
Maximum EV for 2.5 times (JPY billion):(14 × 0.15 × 6 - 2) ÷ 2.5 + 4.5 = 8.748. Curveball: the largest customer wants a price cut
The car maker tells all suppliers it wants a 5 percent price cut. If Tsukimine accepts and its costs do not change, year-5 EBITDA falls by 5 percent of 40 percent of JPY 14 billion, and the return becomes:
MOIC after the price cut:((14 × 0.15 - 14 × 0.4 × 0.05) × 6 - 2) ÷ (9 - 4.5) = 1.98
What the exhibit shows
The price is 6 times EBITDA and the exit multiple is also 6, so the return must come from debt paydown and EBITDA growth, much of it from raising the margin, while 40 percent of revenue depends on one customer.
The recommendation
Slide 1, the answer: the fund should not pay the asking price of JPY 9 billion, but the deal works at about JPY 8.7 billion with protection on the largest customer. At the asking price the entry multiple is 6 times today's EBITDA of JPY 1.5 billion. The plan lifts revenue to JPY 14 billion and the margin from 12.5 to 15 percent, giving year-5 EBITDA of JPY 2.1 billion and an exit value of JPY 12.6 billion at the same 6 times. After debt falls from JPY 4.5 billion to JPY 2 billion, the fund gets about 2.36 times its JPY 4.5 billion of equity, a little under 19 percent a year, below its 20 percent target. Slide 2, the reasons and risks: first, the exit multiple equals the entry multiple, so the return depends on the margin plan and debt paydown, with no help from a higher multiple; second, 40 percent of revenue comes from one car maker, and a 5 percent price cut from that customer would drop the return to about 1.98 times; third, the founder holds key customer relationships and know-how, and skilled engineers may leave with the founder, so the handover must be planned. Slide 3, the offer: bid about JPY 8.7 billion, the price at which the plan returns 2.5 times, ask the founder to stay as an adviser for two years, and tie part of the price to keeping the largest customer's volumes and prices. As a next step, review the supply contract and price history with the car maker.
Risks: The largest customer may move work to a rival during the change of owner; Skilled engineers may leave with the founder; Automation savings may take longer than planned.
Next steps: Review the supply contract and price history with the car maker; Interview the five most senior engineers about their plans; Test the margin plan line by line with the operations team.
A strong candidate
Built the return step by step, compared it with the target, found the price that works, and tested the customer risk with numbers.
A weak candidate
Said the deal is attractive because the multiple is low, without calculating the return or looking at the largest customer.
Case 4: Hokuyo Freight: keep its capacity under the drivers' overtime cap
Where the structure comes from: it is built from the goal of this exact question (Trips a year = drivers x hours per driver / driver hours per trip), not taken from a list. Each branch is one driver of that goal, and the hypothesis above says which branch to test first.
Worked case
Standard: Hokuyo Freight: keep its capacity under the drivers' overtime cap
The prompt
Hokuyo Freight runs trucks between Tokyo and Osaka. Since April 2024 the law has capped its drivers' overtime, and it now runs fewer trips than it used to. How should it keep serving its customers?
Difficulty: Standard. Format: candidate-led, with interviewer dialogue. Industry: Logistics. Region: Japan. Interview length: about 30 minutes. The company is fictional. All figures are illustrative, except the facts that name their source.
Clarifying questions, with the interviewer's answers
- What does the law say?Answer: Since 1 April 2024, drivers' overtime in Japan is capped at 960 hours a year (Japan Ministry of Health, Labour and Welfare, checked 2026-10-02). Hokuyo's 200 drivers used to work about 1,200 hours of overtime a year on top of 2,000 regular hours (illustrative).
- How long is a trip?Answer: A one-way trip between Tokyo and Osaka, with loading and rest, takes about 10 hours of a driver's working time. Each trip earns JPY 150,000 (illustrative).
- What can Hokuyo do?Answer: Hire more drivers at about JPY 6 million a year each, though drivers are hard to find. Or run a relay: drivers from each city swap trailers at a hub halfway and go home the same day. A trip then uses about 8 hours of driver time. The hub would cost JPY 60 million a year (illustrative).
A hypothesis to say out loud: The cap removes driver hours, so Hokuyo loses close to a tenth of its trips. My hypothesis is that using each driver hour better beats hiring in a market short of drivers.
The structure
- Trips a year = drivers x hours per driver / driver hours per trip
- Hours per driver before and after the overtime cap
- Trips and revenue lost
- Key: Option 1: hire drivers; option 2: a relay hub that cuts hours per trip
- Cost of each option and what customers need
Working it through
1. Hours per driver before
Candidate: "2,000 regular hours plus 1,200 hours of overtime."
Hours per driver before the cap:2,000 + 1,200 = 3,2002. Hours per driver now
Candidate: "Overtime is capped at 960 hours."
Hours per driver under the cap:2,000 + 960 = 2,9603. Trips before
Candidate: "200 drivers, 10 hours a trip."
Trips a year before the cap:200 × (2,000 + 1,200) ÷ 10 = 64,0004. Trips now
Candidate: "The same drivers with fewer hours."
Trips a year under the cap:200 × (2,000 + 960) ÷ 10 = 59,2005. Revenue at risk
Candidate: "4,800 lost trips at JPY 150,000 each."
Revenue at risk (JPY a year):(200 × (2,000 + 1,200) ÷ 10 - 200 × (2,000 + 960) ÷ 10) × 150,000 = 720,000,0006. Drivers to hire
Candidate: "Hours for 64,000 trips, divided by 2,960 hours a driver, minus today's 200."
Extra drivers needed:200 × (2,000 + 1,200) ÷ (2,000 + 960) - 200 = 16.227. Cost of hiring
Candidate: "Rounded up to 17 drivers at JPY 6 million each."
Cost of 17 more drivers (JPY a year):17 × 6,000,000 = 102,000,0008. Trips with the relay
Candidate: "At 8 driver hours a trip, the same 200 drivers can run:"
Trips a year with the relay:200 × (2,000 + 960) ÷ 8 = 74,0009. Curveball: what customers need
Interviewer: "Customers want today's 64,000 trips, no more. Which option costs less?" Candidate: "Both restore the trips. Hiring costs JPY 102 million a year and the hub JPY 60 million, so the relay saves:"
Yearly saving of the relay over hiring (JPY):17 × 6,000,000 - 60,000,000 = 42,000,000
The recommendation
Hokuyo should set up the relay hub rather than try to hire its way out. First, the 960-hour cap cuts each driver from 3,200 to 2,960 hours a year. So trips fall from 64,000 to 59,200, and JPY 720 million of revenue is at risk. Second, hiring would need about 16.216 more drivers, so 17, costing JPY 102 million a year in a market where drivers are scarce. Third, the relay cuts each trip from 10 to 8 driver hours, so the same 200 drivers can run 74,000 trips. That is more than customers need, and it costs JPY 42 million a year less than hiring. Pilot the relay on one customer's daily loads and agree delivery times with customers.
Risks: One late truck at the hub delays the drivers on both sides of the relay; Some customers may not accept trailer swaps for fragile or high-value loads.
Next steps: Pilot the relay on one customer's daily loads for two months; Agree delivery windows with the largest customers before switching routes.
A strong candidate
Turned the overtime cap into lost trips and revenue, compared hiring with using driver hours better, and checked the answer against what customers need.
A weak candidate
Said Hokuyo should hire 17 drivers, without asking whether drivers can be found or whether trips can use fewer driver hours.
Case 5: Seorin Fitness: more members, less profit
Where the structure comes from: it is built from the goal of this exact question (Monthly profit = members x average fee + personal training - staff - rent and other fixed costs), not taken from a list. Each branch is one driver of that goal, and the hypothesis above says which branch to test first.
Worked case
Starter: Seorin Fitness: more members, less profit
The prompt
Seorin Fitness runs 20 gyms in Seoul. It has 6,000 more members than a year ago, but its monthly profit has fallen. The exhibit shows the figures. Why did profit fall, and what should Seorin do?
Difficulty: Starter. Format: interviewer-led, with an exhibit. Industry: Fitness. Region: South Korea. Interview length: about 25 minutes. The company is fictional and all figures are illustrative.
Clarifying questions, with the interviewer's answers
- What changed in the offer?Answer: Last year everyone paid KRW 70,000 a month. This year Seorin added a 12-month prepaid plan at KRW 45,000 a month, and 40 percent of members now use it (illustrative).
- Did costs change?Answer: Staff cost rose because the gyms are busier at peak times. Rent and other fixed costs did not change.
- Is personal training important?Answer: Yes. It is the main extra spending per member.
A hypothesis to say out loud: More members but less profit points to the price paid by new members. My hypothesis is that the prepaid discount brought in members who pay less and spend less on training.
The structure
- Monthly profit = members x average fee + personal training - staff - rent and other fixed costs
- Key: Members and the mix of standard and prepaid plans
- Average monthly fee per member
- Personal training spending per member
- Staff cost at busy peak hours, and fixed costs
The exhibit
| Measure | Last year | This year |
|---|---|---|
| Members | 30,000 | 36,000 |
| Average monthly fee per member (KRW) | 70,000 | 60,000 |
| Personal training revenue (KRW millions a month) | 500 | 450 |
| Staff cost (KRW millions a month) | 900 | 1,000 |
| Rent and other fixed costs (KRW millions a month) | 800 | 800 |
Working it through
1. Profit last year
Fees of 30,000 members at KRW 70,000, plus training, minus staff, rent and other costs, in KRW millions a month.
Profit last year (KRW millions a month):30,000 × 70,000 ÷ 1,000,000 + 500 - 900 - 800 = 9002. Profit this year
Fees of 36,000 members at an average of KRW 60,000.
Profit this year (KRW millions a month):36,000 × 60,000 ÷ 1,000,000 + 450 - 1,000 - 800 = 8103. Where the average fee comes from
60 percent pay KRW 70,000 and 40 percent pay KRW 45,000.
Average monthly fee (KRW):0.6 × 70,000 + 0.4 × 45,000 = 60,0004. Effect of more members
6,000 extra members, valued at the old fee, in KRW millions a month.
Effect of more members (KRW millions a month):(36,000 - 30,000) × 70,000 ÷ 1,000,000 = 4205. Effect of the lower fee
All 36,000 members at KRW 10,000 less on average. Training fell by KRW 50 million and staff cost rose by KRW 100 million, which explains the rest.
Effect of the lower fee (KRW millions a month):36,000 × (60,000 - 70,000) ÷ 1,000,000 = -3606. Training spending per member
Last year it was about KRW 16,667 a member. This year:
Training revenue per member (KRW a month):450,000,000 ÷ 36,000 = 12,5007. Curveball: end the prepaid plan
Interviewer: "If we stop selling the prepaid plan, its 14,400 members could renew at KRW 70,000, but 30 percent of them would leave." Change in their fees, in KRW millions a month:
Change in monthly fees (KRW millions):(36,000 × 0.4 × 0.7 × 70,000 - 36,000 × 0.4 × 45,000) ÷ 1,000,000 = 57.6
What the exhibit shows
Members rose 20 percent, but the average fee fell by KRW 10,000 and staff cost rose, so profit fell.
The recommendation
Seorin should stop selling the prepaid plan at KRW 45,000 a month, because it brought in members who pay less and spend less. First, profit fell from KRW 900 million to KRW 810 million a month. The 6,000 extra members added KRW 420 million, but the lower average fee took away KRW 360 million. Second, staff cost rose KRW 100 million, while training spending fell to KRW 12,500 a member from about KRW 16,667. Third, if prepaid members renew at KRW 70,000 and 30 percent leave, fees still rise about KRW 57.6 million a month. Offer a smaller discount for off-peak hours instead, and test it in three gyms.
Risks: More prepaid members may leave than the 30 percent expected; Rivals may copy the discount and win the members who leave.
Next steps: Test an off-peak plan in three gyms for two months; Track training spending by type of plan each month.
A strong candidate
Built a bridge from last year's profit to this year's, split the fee change into more members and a lower price, and tested ending the discount.
A weak candidate
Said more members is good news and suggested more marketing, without seeing that each new member pays less.
Case 6: Should a town council in rural Japan merge its two primary schools?
Where the structure comes from: it is built from the goal of this exact question (Value of merging = yearly running saving + building work avoided, weighed against travel time for pupils), not taken from a list. Each branch is one driver of that goal, and the hypothesis above says which branch to test first.
Worked case
Standard: Should a town council in rural Japan merge its two primary schools?
The prompt
A town council in a mountain area of Japan runs two public primary schools. The number of children is falling. Should the council merge the two schools into one, and how?
Difficulty: Standard. Format: candidate-led, with interviewer dialogue. Industry: Public sector and education. Region: Japan. Interview length: about 30 minutes. The company is fictional and all figures are illustrative.
Clarifying questions, with the interviewer's answers
- How many pupils are there, and where is this heading?Answer: The larger school has 120 pupils. The smaller one, in a valley 12 km away, has 60, and births suggest it will have 35 in six years (illustrative).
- What does the town pay for each school?Answer: For this case, the town pays for buildings, support staff and running costs. That is JPY 80 million a year for the larger school and JPY 60 million for the smaller one. Teachers are paid from other budgets (illustrative).
- What would a merged school cost?Answer: A merged school at the larger site would cost JPY 95 million a year, plus 3 school buses at JPY 8 million a year each. The smaller school also needs earthquake strengthening costing JPY 300 million if it stays open (illustrative).
A hypothesis to say out loud: Each school has costs that do not shrink with fewer pupils, so merging should save money. My hypothesis is that bus times and the valley community decide how to do it.
The structure
- Value of merging = yearly running saving + building work avoided, weighed against travel time for pupils
- Key: Running cost of two schools against one merged school
- School buses for pupils from the valley
- Earthquake strengthening avoided
- Pupil numbers and cost per pupil over time
- Bus ride length and the valley community
Working it through
1. Cost today
Candidate: "JPY 80 million plus JPY 60 million, in JPY millions a year."
Cost of two schools (JPY millions a year):80 + 60 = 1402. Cost after merging
Candidate: "The merged school plus 3 buses."
Cost of the merged school (JPY millions a year):95 + 3 × 8 = 1193. Yearly saving
Candidate: "The difference."
Yearly saving (JPY millions):80 + 60 - (95 + 3 × 8) = 214. Ten-year value
Candidate: "Ten years of savings, plus the JPY 300 million of strengthening avoided. I will ignore discounting for a first view."
Ten-year value of merging (JPY millions):(80 + 60 - (95 + 3 × 8)) × 10 + 300 = 5105. Cost per pupil today
Candidate: "At the smaller school, JPY 60 million over 60 pupils."
Cost per pupil at the smaller school today (JPY):60,000,000 ÷ 60 = 1,000,0006. Cost per pupil in six years
Candidate: "The same cost over 35 pupils."
Cost per pupil at the smaller school in six years (JPY):60,000,000 ÷ 35 = 1,714,2867. Curveball: the bus ride
Interviewer: "Parents say the longest bus ride, 50 minutes, is too long. A fourth bus would cut it to 30 minutes." Candidate: "The yearly saving becomes:"
Yearly saving with four buses (JPY millions):80 + 60 - (95 + 4 × 8) = 138. Ten-year value with four buses
Candidate: "Ten years of the smaller saving, plus the strengthening avoided."
Ten-year value with four buses (JPY millions):(80 + 60 - (95 + 4 × 8)) × 10 + 300 = 430
The recommendation
The council should merge the two schools at the larger site and run four buses, not three. First, merging cuts the town's yearly cost from JPY 140 million to JPY 119 million, a saving of JPY 21 million. Second, it avoids JPY 300 million of earthquake strengthening, so the ten-year value is about JPY 510 million. Third, the smaller school's cost per pupil would rise from JPY 1 million to about JPY 1.71 million as pupils fall to 35. A fourth bus keeps the longest ride to 30 minutes and still saves JPY 13 million a year, or JPY 430 million over ten years. Hold meetings with valley parents before deciding, and offer the old building to the community.
Risks: Young families may leave the valley if its school closes; Bus routes may be cut by snow in winter.
Next steps: Hold meetings with parents in the valley before the council votes; Test the bus routes in winter conditions.
A strong candidate
Compared the running cost of two schools with one, counted the building work avoided, and paid for a shorter bus ride rather than ignoring the families.
A weak candidate
Said small schools are part of the community and should stay open, with no view on cost, pupil numbers or the building.
Case 7: Should Hanbit Foods acquire Dasom Kit, a meal kit maker?
Where the structure comes from: it is built from the goal of this exact question (Highest price = EBITDA with savings x Hanbit's limit of 8 times, today and if sales keep falling), not taken from a list. Each branch is one driver of that goal, and the hypothesis above says which branch to test first.
Worked case
Stretch: Should Hanbit Foods acquire Dasom Kit, a meal kit maker?
The prompt
Hanbit Foods, a Korean maker of sauces and ready meals, has been offered Dasom Kit, a meal kit maker, for KRW 90 billion. The data pack shows the numbers. Write a one-page recommendation: should Hanbit buy, and at what price?
Difficulty: Stretch. Format: written case, with a data pack. Industry: Packaged food. Region: South Korea. Interview length: about 40 minutes. The company is fictional and all figures are illustrative.
Clarifying questions, with the interviewer's answers
- How does Hanbit judge a price?Answer: It pays no more than 8 times EBITDA (earnings before interest, tax, depreciation and amortisation), counting the savings it can bring (illustrative).
- What could Hanbit bring?Answer: Its own sauce factories would cut Dasom's ingredient costs by about 3 percent of sales. Its sales team supplies convenience stores across Korea and could sell KRW 30 billion of kits a year there, at a 10 percent EBITDA margin (illustrative).
- How is the market doing?Answer: Dasom's sales fell from KRW 140 billion two years ago to KRW 120 billion this year (illustrative).
A hypothesis to say out loud: Dasom is expensive on its own. My hypothesis is that Hanbit can justify the price only through the savings it brings, and that falling sales could push the price above its limit.
The structure
- Highest price = EBITDA with savings x Hanbit's limit of 8 times, today and if sales keep falling
- Dasom on its own: EBITDA, margin and asking multiple
- Key: Savings only Hanbit can bring: sauces and convenience-store sales
- Falling sales and the margin next year
- Bid price and payment terms
The exhibit
| Item | Value |
|---|---|
| Sales this year (KRW billions) | 120 |
| Sales two years ago (KRW billions) | 140 |
| EBITDA this year (KRW billions) | 6 |
| Asking price (KRW billions) | 90 |
| Ingredient savings from Hanbit sauces (% of sales) | 3 |
| Extra sales through convenience stores (KRW billions a year) | 30 |
| EBITDA margin on extra sales (%) | 10 |
| Hanbit's highest multiple of EBITDA with savings | 8 |
Working it through
1. Margin today
KRW 6 billion of EBITDA on KRW 120 billion of sales.
EBITDA margin (%):6 ÷ 120 × 100 = 52. Asking multiple
The asking price over today's EBITDA.
Asking price as a multiple of EBITDA:90 ÷ 6 = 153. Ingredient savings
3 percent of sales, in KRW billions.
Ingredient savings (KRW billions a year):120 × 0.03 = 3.64. EBITDA with savings
Plus 10 percent on KRW 30 billion of convenience-store sales.
EBITDA with savings (KRW billions):6 + 120 × 0.03 + 30 × 0.1 = 12.65. Multiple with savings
The asking price over EBITDA with savings.
Asking price over EBITDA with savings:90 ÷ (6 + 120 × 0.03 + 30 × 0.1) = 7.146. Curveball: sales keep falling
Interviewer: "Sales could fall another 10 percent next year, and Dasom's own margin could slip to 4 percent." Sales next year, in KRW billions:
Sales next year (KRW billions):120 × 0.9 = 1087. EBITDA with savings next year
4 percent margin and 3 percent ingredient savings on the lower sales, plus the convenience-store EBITDA.
EBITDA with savings next year (KRW billions):120 × 0.9 × 0.04 + 120 × 0.9 × 0.03 + 30 × 0.1 = 10.568. Multiple next year
The asking price over next year's EBITDA with savings.
Asking price over next year's EBITDA with savings:90 ÷ (120 × 0.9 × 0.04 + 120 × 0.9 × 0.03 + 30 × 0.1) = 8.529. Highest price at 8 times
Hanbit's limit applied to next year's EBITDA with savings, in KRW billions.
Highest price at 8 times (KRW billions):8 × (120 × 0.9 × 0.04 + 120 × 0.9 × 0.03 + 30 × 0.1) = 84.48
What the exhibit shows
The price only works with Hanbit's savings, and sales have fallen for two years, so the bid should rest on next year's lower EBITDA.
The recommendation
Hanbit should buy Dasom Kit, but offer no more than about KRW 84.48 billion, not the KRW 90 billion asked. First, the asking price is 15 times today's EBITDA of KRW 6 billion, on a 5 percent margin, which is expensive on its own. Second, Hanbit can bring savings no other buyer has: its sauces cut ingredient costs by KRW 3.6 billion, and its convenience-store channel adds KRW 3 billion of EBITDA. With these, EBITDA is KRW 12.6 billion and the price is about 7.143 times, inside the limit of 8 times. Third, sales are falling. If they drop another 10 percent to KRW 108 billion and the margin slips to 4 percent, EBITDA with savings is KRW 10.56 billion. The price then becomes about 8.523 times, above the limit. So the bid should rest on next year's lower EBITDA: 8 times KRW 10.56 billion is KRW 84.48 billion. Part of the price could be paid later if sales stop falling. The main risk is that convenience-store sales take longer to build than planned. Next, test Dasom kits in 200 convenience stores before signing.
Risks: Convenience-store sales may take longer to build than planned; Meal kit demand may keep falling beyond next year.
Next steps: Test Dasom kits in 200 convenience stores before signing; Check in due diligence why sales fell, by channel and by product.
A strong candidate
Valued Dasom with the savings only Hanbit can bring, tested the price against falling sales, and turned that into a bid and payment terms.
A weak candidate
Rejected the deal at 15 times EBITDA without counting the savings, or accepted it at KRW 90 billion without testing falling sales.
Sources for this lesson (2)
- Recognized public explanations of case-interview concepts and terms
- Japan Ministry of Health, Labour and Welfare: overtime caps for construction, drivers and doctors (960 hours a year for driving work from 1 April 2024), checked 2026-10-02
My notes on this lesson
0 of 5,000 characters. Saves automatically.
Completing lessons builds your skill levels and your readiness.
Next: put this module into practice
You have read every lesson in this module. A full case puts them to work, with a model answer to compare against.
Spotted something wrong or out of date? Report a mistake. We check every report and correct the page.