Bain Case Interview Examples: 10 Practice Cases With Worked Solutions (2026)

10 Bain case interview examples with full worked solutions, including Bain's own published coffee shop and fashion retailer cases, the SOVA screen, and the private equity diligence archetype Bain asks more than any other firm.

Updated Jul 31, 2026. Reviewed by Ned.

Bain case interviews are candidate-led and conversational, typically two first-round interviews of about 40 to 45 minutes each followed by two or three final-round interviews with senior interviewers, with most regions running an online screen before any live case. Bain publishes two of its own practice cases: Coffee Shop Co., a market sizing and break-even problem set in Cambridge, England, and FashionCo., a revenue decline problem. Bain's published coffee case answer is 204,675 cups to break even in year one, about 3 percent of a 7 million cup market.

Both of those cases are worked end to end below, alongside eight original Bain-style examples covering private equity diligence, channel profitability, market entry, plant throughput, and a pricing decision that reverses when you look past the subscription line. Bain states that it is testing whether you can make sensible assumptions, do quick math, and build constructively on other people's ideas, and that there is not necessarily a right answer.

What should a Bain case interview example actually show you?

Most example lists give you a prompt and a paragraph of "how to think about it". That is a prompt collection, not a worked case. A Bain example is only useful if you can grade yourself against it, and grading needs five things visible.

Every case below carries all five. Work each one with the answer covered, then compare. For the format and round structure behind these examples, the Bain case interview guide owns process; this page owns material.

How Bain differs from McKinsey and BCG in practice

The three firms score the same underlying skills and test them in visibly different formats. Practising only one firm's shape is the most common preparation error, because the transfer is worse than candidates expect.

The practical consequence: if your reps have all been interviewer-led, the first Bain case will feel disorienting for the first four minutes, because nobody tells you what to look at next. If your reps have all been exhibit-heavy, Bain will feel light on data and you will be tempted to keep asking for exhibits that do not exist. Bain wants you to reason forward from a thin fact base and commit.

The five question types inside a Bain case

A Bain case is a sequence of separately scored asks rather than one continuous conversation. Each of the five maps to a skill you can rehearse on its own, which is far more efficient than running full cases and discovering the same weakness five times.

The one you cannot self-grade is the first. Everyone believes their structure was mechanism-level. Get one scored before you read the ten examples, so you know which of the five you are actually spending your practice hours on.

Build a candidate-led structure and get it scored from the CoachNed drill engine. Answer a real prompt and get AI-scored feedback. Free accounts include daily drills.

The online screen that comes before the case

In most regions the case examples below are not the first thing standing between you and an offer. Bain screens candidates with an online assessment first, most commonly SOVA, and a failed screen means the case preparation never gets used. The screen tests numerical reasoning, logical pattern recognition, situational judgement, and behavioural consistency, which is a different skill set from case solving and needs its own reps.

The two items below are the real format: one numerical data-interpretation question and one abstract logical-pattern question, the two sections candidates most often report being caught out by.

Live free simulator from CoachNed for the Bain SOVA numerical and logical assessment. Practise the real format in the browser, no account needed to start.

Section-by-section detail, reported timings, and what each section is actually scoring live in the Bain SOVA test guide. The full simulator sits at the Bain SOVA practice page. Some offices use TestGorilla instead of or alongside SOVA, with a different section mix and a tighter clock; the Bain TestGorilla guide covers that variant. There is no simulator for the TestGorilla version, so prepare for it with timed numerical and logical reps rather than a branded practice console.

Once the screen is behind you, the rest of your preparation is cases, and that is where the next 4,000 words go.

The private equity diligence archetype, and why it is the Bain-specific one

Bain runs more private equity work than either of the other two MBB firms, and the case portfolio reflects it. A Bain candidate who has never worked an acquisition case is exposed in a way that a McKinsey candidate is not.

The archetype has a fixed shape and you should be able to run it from memory:

• The thesis. Why would this asset be worth more under this owner than the current one? Name the value creation lever before touching the numbers.

• The market. Is the pool growing, and is the growth structural or cyclical? A cyclical peak entry kills otherwise sound deals.

• The asset. Does the target win for a reason that survives the ownership change, or does its performance come from something the seller was doing that the buyer cannot continue?

• The returns. Entry multiple, the earnings bridge to exit, the exit multiple assumption, and whether the leverage works.

• The deal breakers. The two or three findings that would make you walk.

The number that decides these cases is almost always the earnings bridge, and it is where candidates lose the room: not because it is hard, but because they build it without stating which line is growth, which is margin, and which is multiple. The M&A case framework covers the full structure. Get the arithmetic reflexive first, because a diligence case gives you no time to be slow at it.

Rep the multi-step math a diligence case runs on from the CoachNed drill engine. Answer a real prompt and get AI-scored feedback. Free accounts include daily drills.

10 Bain case interview examples with worked solutions

The first two are Bain's own published cases, worked with the figures Bain published. The remaining eight are original Bain-style cases built on the archetypes Bain asks most.

Case 1: Coffee Shop Co. (Bain's published case)

Prompt. A friend wants to open a coffee shop in Cambridge, England, and asks whether it is a good business idea.

Published facts. Cambridge population 100,000. Price per cup GBP 3. Cost per cup GBP 1. Setup cost GBP 245,610. Annual operating cost GBP 163,740. Bain's published market estimate is approximately 7,000,000 cups a year.

How to open it. Name the decision first: this is a go or no-go on a single-site investment, so the test is whether achievable volume clears the year-one cost base. Two branches carry the whole case: how big is the demand pool, and how many cups does the shop need to sell to break even? Everything else, competition, location, and the friend's own risk tolerance, sits under those.

Sizing. Bain's stated market of about 7 million cups a year works out at roughly 19,000 cups a day across a population of 100,000, so about 19 cups per 100 residents per day. That figure is coffee bought out of home, not total consumption, which is the assumption most candidates state backwards. Say it out loud: a per-person daily consumption figure includes coffee made at home, and the addressable pool for a shop is the out-of-home share.

Break-even. Contribution per cup is GBP 3 minus GBP 1, so GBP 2. Year one must recover setup plus operating cost:

• GBP 245,610 + GBP 163,740 = GBP 409,350

• GBP 409,350 divided by GBP 2 = 204,675 cups

• 204,675 divided by 365 = about 561 cups a day

• 204,675 divided by 7,000,000 = about 2.9 percent market share

The judgment answer. Three percent of the market in year one is achievable for a well-sited independent shop, so the arithmetic does not kill the idea. The interviewer's push is what makes or breaks the answer: 561 cups a day is roughly one cup a minute across a nine-hour trading day, which is a real operational constraint on till throughput and staffing, not just a market share question. Bain's own guidance lands on investment timeline and competitive positioning as the next things to understand.

Recommendation. Proceed to a site-level test rather than a yes or no. Break-even needs about 3 percent share and one cup a minute of sustained throughput, which is plausible in a high-footfall location and not plausible in a secondary one. The next test is a two-week footfall count at the three candidate sites, because the site choice, not the market, decides this investment.

Where candidates lose it. Treating year one break-even as operating cost only. The setup cost is sunk in year one and Bain's published answer includes it. Excluding it gives 81,870 cups and a materially easier, wrong answer.

Sizing openers like this one are the fastest place to gain points, because the assumption chain is visible and gradeable. Do one timed rep now and see where your driver chain gets flagged.

Size a market and defend your assumption chain from the CoachNed drill engine. Answer a real prompt and get AI-scored feedback. Free accounts include daily drills.

Case 2: FashionCo. (Bain's published case)

Prompt. A women's fashion retailer has seen revenues decline for five consecutive years. The CEO wants to know why, and what can be done to drive revenue growth before the next management meeting.

Published facts. 10 million customers, flat. Average annual spend USD 100. Two options on the table.

• Option A, a rewards programme: 25 percent of customers participate in year one, each pays a USD 50 one-time activation fee, and each receives a 20 percent perpetual discount.

• Option B, intermittent sales: a 20 percent discount for three months a year, during which sales volume doubles.

How to open it. The diagnosis and the options are two separate jobs, and candidates who blend them lose control of the case. The diagnosis branch is market trends (where women's fashion spend has moved over five years, and which new entrants captured it) and consumer trends (what shoppers now want that this retailer does not offer). The options branch is arithmetic. Do the arithmetic, then bring the diagnosis back in, because the arithmetic alone cannot answer a five-year structural decline.

Baseline. 10,000,000 customers times USD 100 equals USD 1,000M a year.

Option A worked.

• Participants: 10M times 25 percent = 2.5M. Non-participants: 7.5M.

• Non-participant revenue: 7.5M times USD 100 = USD 750M.

• Participant spend after a 20 percent discount: USD 80 each, so 2.5M times USD 80 = USD 200M.

• Activation fees: 2.5M times USD 50 = USD 125M.

• Total: 750 + 200 + 125 = USD 1,075M, an uplift of USD 75M.

Option B worked.

• Baseline revenue in the three discount months: USD 1,000M divided by 4 = USD 250M.

• Volume doubles and price falls 20 percent: USD 250M times 2 times 0.8 = USD 400M.

• The other nine months are unchanged: USD 750M.

• Total: 750 + 400 = USD 1,150M, an uplift of USD 150M.

Sensitivity, and say this out loud. Option A's answer swings entirely on the activation fee. Strip it out and the programme is a USD 50M revenue loss, not a USD 75M gain, because you are handing a permanent 20 percent discount to a quarter of the base. Any answer that quotes a single number without naming that dependency is incomplete. Option B's answer depends on the doubling holding, and a doubling of volume at a 20 percent discount is an aggressive assumption to accept without testing.

Recommendation. On the stated assumptions Option B produces the larger one-year revenue lift, roughly double Option A's. But Option A is the only one of the two that builds a durable customer relationship, and Option B trains the base to wait for the discount window, which is how a five-year decline gets worse rather than better. The strongest answer says: run Option B this year for the cash, do not run it twice, and use the year to fix the assortment problem, because neither option addresses why customers left. The next test is whether the customers lost over five years went to a specific competitor or out of the category.

Where candidates lose it. Answering "B, it makes more money" and stopping. The published case is not a calculation exercise, it is a test of whether you notice that both options are promotions and the client's problem is five years of structural decline.

Case 3: Private equity diligence, boutique fitness roll-up

Prompt. A mid-market private equity fund is considering acquiring FlexLoop, a boutique fitness chain with 42 studios, for USD 180M. Should they proceed?

Facts given on request. Revenue USD 96M. EBITDA** USD 14.4M, a 15 percent margin. 84,000 members at an average USD 95 a month. Monthly churn 4.5 percent. New studio build cost USD 1.1M, reaching USD 420K of EBITDA by year three.

Exhibit. Same-studio revenue growth, last twelve months: 26 studios open more than three years grew 1.2 percent; 16 studios open less than three years grew 19 percent.

The structure. Thesis, market, asset quality, returns, deal breakers. State up front that the entry multiple is 180 divided by 14.4, which is 12.5 times EBITDA, and that a 12.5 times multiple on a chain growing at 1.2 percent in its mature base is only defensible if the buyer can build new units profitably.

The decisive read. The 1.2 percent versus 19 percent split is the whole case. Blended growth looks healthy, but all of it comes from immature studios ramping, not from the business getting better. So this is not a growth asset, it is a unit-development story, and the diligence question becomes whether new units work rather than whether the current ones do.

Churn math. At 4.5 percent monthly churn, the average member stays 1 divided by 0.045, about 22 months, and annual retention is 0.955 to the twelfth power, about 58 percent. The base loses roughly 42 percent of members a year and has to replace them just to stand still. That is normal for boutique fitness and it is also why mature-studio growth is 1.2 percent.

Unit economics. A new studio costs USD 1.1M and produces USD 420K of EBITDA at maturity, a 2.6-year payback. That is genuinely good, and it is the reason the deal is not dead.

Returns bridge. Assume the sponsor opens 10 studios a year for four years, 40 new units, USD 44M of capex funded from cash flow.

• EBITDA at exit: USD 14.4M plus 40 times USD 420K = USD 14.4M + USD 16.8M = USD 31.2M

• Exit at a deliberately lower 11 times: USD 343M enterprise value

• Entry EV USD 180M to exit EV USD 343M is 1.9 times over five years, about 14 percent EV growth a year

• With USD 90M of debt held flat, equity goes from USD 90M to USD 253M, 2.8 times, roughly a 23 percent IRR

Recommendation. Proceed, but not at USD 180M on this build plan. The returns clear a 20 percent hurdle only if all 40 units hit USD 420K, which assumes no cannibalisation and no site-quality decay across a doubling of the estate, and the best sites are usually taken first. The two deal breakers to test in diligence: whether the last cohort of studios is hitting the USD 420K number or a lower one, and whether membership pricing has been rising or the USD 95 average is flattered by a mix shift. If cohort EBITDA is decaying, the model is worth 10 times, not 12.5.

BainM&A · medium

Run a live acquisition case with feedback

Same shape as Case 3: build the thesis, work the earnings bridge, and defend the multiple out loud before you give the go or no-go.

Practice this case free

Case 4: Profitability, European apparel margin reset

Prompt. A European apparel retailer with EUR 814M of revenue has watched gross margin fall from about 45 percent to 41 percent over three years. EBIT has halved. The CEO wants the margin back.

Facts given on request. 15 million units sold, roughly flat. Cost of goods EUR 30 a unit, stable. Full-price average selling price EUR 58, stable. Markdown share of units has gone from 22 percent to 31 percent. Markdown average selling price has fallen from EUR 41 to EUR 35.

The structure. With units flat and unit cost flat, this is not a cost problem and it is not a demand problem. It is a price realisation problem, and there are exactly two mechanisms: more units going through markdown, and each markdown unit fetching less. Quantify both before recommending anything.

Worked math, three years ago.

• Full price: 15M times 78 percent = 11.7M units times EUR 58 = EUR 678.6M

• Markdown: 15M times 22 percent = 3.3M units times EUR 41 = EUR 135.3M

• Revenue EUR 813.9M, COGS** 15M times EUR 30 = EUR 450M, gross profit EUR 363.9M, margin 44.7 percent

Worked math, today.

• Full price: 10.35M units times EUR 58 = EUR 600.3M

• Markdown: 4.65M units times EUR 35 = EUR 162.8M

• Revenue EUR 763.1M, COGS EUR 450M, gross profit EUR 313.1M, margin 41.0 percent

The same 15 million units now generate EUR 51M less revenue and EUR 51M less gross profit. Blended selling price fell from EUR 54.3 to EUR 50.9.

The trap, and it is the whole case. The obvious lever is to cut the buy. Buy 12 percent fewer units, sell a higher proportion at full price, and margin percentage improves. Work it: 13.2M units, markdown back to 24 percent, markdown price recovering to EUR 38 gives revenue of about EUR 702M, COGS of EUR 396M, gross profit of EUR 306M at a 43.6 percent margin. The margin percentage went up and the gross profit euros went down. A candidate who recommends this because the percentage improved has failed the commercial judgment test that Bain weights most.

The lever that works. Attack the markdown rate at source: buy accuracy and newness cadence, so the same 15M units sell at a better mix. At 15M units, markdown share back to 24 percent, markdown price recovered to EUR 38, revenue is EUR 798M against EUR 450M of COGS, gross profit EUR 348M, margin 43.6 percent. Same margin percentage as the buy cut, EUR 42M more gross profit.

Recommendation. Fix markdown rate, do not cut the buy. The two paths land on the same 43.6 percent margin and are EUR 42M apart in gross profit, which is the number that pays for the business. The main risk is that the 24 percent markdown rate is not recoverable if the assortment problem is structural rather than a buying error. The next test is markdown rate by category: if it is concentrated in two or three departments, it is a buying fix; if it is even across the range, the brand has a desirability problem and this becomes a much longer conversation.

Case 5: Market entry, home battery storage in Texas

Prompt. A European home energy company wants to enter the Texas residential battery storage market. Should it, and how?

Facts given on request. 11.0M Texas households. About 62 percent owner-occupied single family. Roughly 4 percent of those have solar. About 95,000 new residential solar installations a year, with a 28 percent battery attach rate. Retrofit rate on the existing solar base about 6 percent a year. Installed system price USD 11,500. Contribution margin 22 percent. Fixed entry cost USD 14M for an installer network, permitting, and inventory. Plan assumes 8 percent share by year three.

Sizing the pool.

• Owner-occupied single family: 11.0M times 62 percent = 6.8M homes

• Solar-attached: 6.8M times 4 percent = 272,000 homes

• New-install attach: 95,000 times 28 percent = 26,600 units a year

• Retrofit: 272,000 times 6 percent = 16,300 units a year

• Addressable: about 42,900 units a year, or USD 493M at USD 11,500

The plan, tested. Eight percent share is 3,430 units, USD 39.5M of revenue, and USD 8.7M of contribution at a 22 percent margin. Against a USD 14M fixed cost, the plan does not break even in year three.

Break-even share. Contribution per unit is USD 11,500 times 22 percent = USD 2,530. USD 14M divided by USD 2,530 = 5,534 units, which is 12.9 percent share. The plan needs 61 percent more share than it assumes just to cover its own fixed cost.

The judgment move. Do not stop at "the entry fails". The number that fails is the fixed cost, not the market. A USD 493M market growing with solar attach is genuinely attractive; a USD 14M owned installer network in a state with an established independent installer base is the wrong way to reach it.

Recommendation. Enter, but through installer partnerships rather than an owned network. Converting the USD 14M of fixed cost into a per-unit channel margin makes the 8 percent plan profitable in year two rather than unprofitable in year three, at the cost of a lower contribution margin and less control of the customer relationship. The main risk is channel dependency: the same partners will carry competing batteries. The next test is whether the top 20 Texas installers will sign volume commitments, because the whole recommendation rests on that.

BainMarket entry · medium

Run a live market entry case with feedback

The same entry decision, live: size the pool yourself, find the break-even share, and defend the entry mode when the interviewer pushes back.

Practice this case free

Case 6: Operations, motorcycle plant throughput

Prompt. A motorcycle manufacturer's flagship plant is producing 176 units a day against a design capacity of 240. Demand exceeds supply and the client wants the gap closed without new capital lines.

Facts given on request. Three assembly lines, 480-minute shifts. Availability 84 percent, performance 91 percent, quality yield 96 percent. Unplanned changeovers consume 68 minutes a shift. Contribution per unit USD 1,450. 240 production days a year. A changeover reduction programme would cost USD 900K.

The structure. Throughput gaps decompose cleanly into availability, speed, and yield. Do not open with a cost-cutting tree; the client is supply constrained, so every recovered unit is a sold unit and the whole case is about the multiplication.

Worked math.

• Overall equipment effectiveness: 0.84 times 0.91 times 0.96 = 73.4 percent

• 240 times 73.4 percent = 176 units a day, which reconciles to the observed figure

Lever one, changeovers. Cutting changeover time from 68 to 34 minutes a shift returns 34 minutes of 480, about 7 percentage points of availability, taking it to 91 percent.

• 0.91 times 0.91 times 0.96 = 79.5 percent

• 240 times 79.5 percent = 191 units a day, plus 15 a day

• 15 times USD 1,450 times 240 days = USD 5.2M a year

• Against a USD 900K programme cost, payback is about 2.1 months

Lever two, yield. Lifting quality yield from 96 to 98.5 percent takes OEE to 81.6 percent and output to about 196 a day, another 5 units, worth USD 1.7M a year.

Recommendation. Run the changeover programme first. It returns USD 5.2M a year for USD 900K and pays back inside a quarter, and it is the only lever that needs no engineering change to the product. Sequence yield work behind it, because the scrap analysis will take a quarter to complete and the two programmes compete for the same industrial engineering team. The main risk is that the 34-minute changeover target is not achievable on the oldest of the three lines. The next test is a single-line trial before the programme is funded across all three.

BainOperations · medium

Run a live operations case with feedback

The same throughput problem live: decompose the capacity gap, quantify each lever, and sequence them when the interviewer asks which one you would fund first.

Practice this case free

Case 7: Channel profitability, consumer electronics distributor

Prompt. A consumer electronics distributor has seen EBIT fall from USD 28.7M to USD 16.4M over three years while revenue grew from USD 390M to USD 410M. Why, and what should they do?

Exhibit, EBIT by channel.

The obvious answer, and why it is wrong. Everyone says the same thing: e-commerce grew 71 percent at a low margin, so mix shift destroyed the P&L. Test it before you say it.

The decomposition, and this is the point of the case. Apply today's revenue mix at the old margins:

• Retail: USD 230M times 9.5 percent = USD 21.9M

• E-commerce: USD 120M times 3.0 percent = USD 3.6M

• B2B: USD 60M times 2.6 percent = USD 1.6M

• Total at old margins: USD 27.0M

So the mix shift costs USD 28.7M minus USD 27.0M = USD 1.7M. Margin erosion inside the channels costs USD 27.0M minus USD 16.4M = USD 10.6M. That is 86 percent of the decline, and most of it sits in retail, the channel everyone assumed was healthy.

Digging into retail. Of the 3-point retail margin drop, vendor rebate income fell from 3.1 percent to 1.4 percent of retail revenue, which is 1.7 points, USD 3.9M. The rest is freight and a returns rate that has drifted from 7 percent to 11 percent.

Recommendation. Fix retail before touching e-commerce. Renegotiating the rebate tiers and pulling the returns rate back to 8 percent recovers most of the USD 10.6M, and neither requires a strategic decision about channel mix. Re-pricing e-commerce is the second move, not the first, and cutting it would surrender 71 percent revenue growth to solve USD 1.7M of the problem. The main risk is that rebate tiers fell because volume commitments were missed, in which case they are not recoverable by negotiation. The next test is the rebate contract terms against actual volumes by vendor.

Why this is a Bain case. The exhibit invites the wrong answer and the interviewer will let you give it. The scored moment is whether you decompose before concluding.

BainProfitability · medium

Run a live profitability case with feedback

Same decomposition discipline: separate mix from within-channel erosion, then defend which one you would fix first.

Practice this case free

Case 8: Growth, adding daycare to a veterinary clinic chain

Prompt. A veterinary chain with 340 clinics is considering adding dog daycare and boarding. Is it worth doing, and where?

Facts given on request. 131M US households, 66 percent own a pet, roughly 51M own a dog. About 18 percent of dog-owning households use paid daycare or boarding, at an average USD 640 a year. Per clinic catchment: 42,000 households, 39 percent dog-owning. Buildout USD 185K a clinic. Contribution margin 31 percent. Only 120 of 340 clinics have spare floor area.

Sizing the market.

• Dog-owning households using paid care: 51M times 18 percent = 9.2M households

• Market: 9.2M times USD 640 = about USD 5.9B

Sizing a single clinic.

• Dog-owning households in catchment: 42,000 times 39 percent = 16,400

• Paid-care users: 16,400 times 18 percent = 2,950

• At 12 percent capture: 354 customers times USD 640 = USD 227K revenue

• Contribution: USD 227K times 31 percent = USD 70K a year

• Against USD 185K of buildout, payback is 2.6 years

Scaling it. Across the 120 eligible clinics: USD 8.4M of annual contribution against USD 22.2M of capex, the same 2.6-year payback. That is acceptable but not exciting for a services add-on, and the average hides the distribution.

The judgment move. Twelve percent capture in a catchment with 2,950 existing paid-care users assumes the chain wins against incumbent daycare operators on convenience alone. In high dog-density catchments the same 12 percent produces materially more customers on the same fixed buildout, so the payback is much better at the top of the distribution and much worse at the bottom.

Recommendation. Do not roll out to 120 clinics. Pilot in the 40 clinics with the highest dog-density catchments, where the same USD 185K of buildout serves a larger user pool, and set a capture-rate gate before funding the rest. The main risk is that veterinary clinics carry a clinical association that makes them a poor daycare brand. The next test is the pilot's capture rate at month six against the 12 percent assumption.

Case 9: Pricing, a grocery subscription

Prompt. An online grocer has 640,000 subscribers paying USD 12.99 a month for free delivery. Finance wants to raise it to USD 15.99. Should they?

Facts given on request. A test on 40,000 subscribers showed 9 percent cancelling within 90 days at the higher price. Subscribers generate USD 46 a month of gross profit on their grocery baskets. Non-subscribers who cancel drop to roughly a third of their previous order frequency.

The subscription line alone.

• Today: 640,000 times USD 12.99 times 12 = USD 99.8M

• After: 640,000 times 91 percent = 582,400 subscribers times USD 15.99 times 12 = USD 111.8M

• Uplift: plus USD 12.0M, almost all of it margin

At this point most candidates recommend the increase. It is the wrong answer and the interviewer is waiting.

The basket line.

• Subscribers lost: 640,000 times 9 percent = 57,600

• Gross profit they carried: 57,600 times USD 46 times 12 = USD 31.8M a year

• Even allowing that cancelled subscribers keep ordering at a third of their frequency, roughly USD 21M of that gross profit does not come back

Net. USD 12.0M gained on the subscription line against roughly USD 21M lost on baskets is a net loss of about USD 9M, and that is on the optimistic residual-ordering assumption. The subscription is not a profit centre, it is the mechanism that holds order frequency.

Recommendation. Hold the headline price. If the finance target is non-negotiable, raise it only for the cohort whose grocery gross profit is low enough that losing them is neutral, which the test data can identify directly. The main risk is that a segmented price becomes public and damages trust, which is a real reputational cost in grocery. The next test is a cohort analysis of gross profit per subscriber, because the answer changes entirely by decile.

Case 10: Growth, a specialty coffee wholesaler

Prompt. A specialty coffee roaster sells USD 62M a year, 70 percent wholesale to cafes and 30 percent direct to consumers online. The board wants to double revenue in four years. What is the path?

Facts given on request. Wholesale gross margin 26 percent, direct 58 percent. Wholesale accounts number 1,100 with 8 percent annual account churn. Direct: 84,000 customers, average order USD 48, 4.1 orders a year, customer acquisition cost USD 38, first-year retention 41 percent.

The structure. Doubling from USD 62M to USD 124M in four years is 19 percent compound growth. There are three arithmetic paths: more wholesale accounts, more revenue per wholesale account, or scaling direct. Size each before choosing, because the board will assume the answer is "do all three".

Wholesale path. 1,100 accounts producing USD 43.4M is about USD 39,500 an account. Growing wholesale to USD 87M would need roughly 1,100 net new accounts, doubling the account base while replacing 8 percent churn a year, so about 1,450 gross wins over four years, around 30 a month against a sales team that currently wins about 12. This is a hiring plan, not a strategy, and it doubles a 26 percent margin business.

Direct path. 84,000 customers at USD 48 times 4.1 orders is USD 16.5M. Scaling direct to USD 50M means about 254,000 customers, so roughly 170,000 net additions. At 41 percent first-year retention, that requires about 415,000 gross acquisitions, and at USD 38 of acquisition cost that is USD 15.8M of spend, against a first-year gross profit per customer of USD 48 times 4.1 times 58 percent, about USD 114. The unit economics work, comfortably. The volume assumption is the fragile part.

The judgment answer. The margin difference decides it. Every dollar of direct revenue carries 58 cents of gross profit against 26 cents wholesale, so a mix shift toward direct grows profit more than twice as fast as it grows revenue. The retention number is the real constraint: at 41 percent, the business refills two thirds of its direct base every year, and lifting retention to 55 percent cuts the required gross acquisitions from 415,000 to about 309,000 and the spend from USD 15.8M to USD 11.7M.

Recommendation. Lead with direct and fix retention before scaling spend. Growing direct from USD 16.5M to USD 50M while wholesale grows modestly to USD 74M reaches the USD 124M target and roughly triples gross profit, against a wholesale-led path that would hit the same revenue with far less profit. The main risk is channel conflict, because wholesale cafe accounts will not enjoy being undercut by the roaster's own site. The next test is a retention cohort analysis: if 41 percent is a product problem rather than a marketing problem, the whole plan changes.

Self-diagnostic: are you practising like a Bain candidate?

Work the ten cases, then answer these honestly. Each maps to a specific failure the examples above are built to expose.

• Did your recommendation carry a number every time? "Fix the markdown rate" is an observation. "Fix markdown rate rather than cutting the buy, the two paths land at the same margin percentage and are EUR 42M apart in gross profit" is a recommendation. Bain scores the second.

• Did you decompose before concluding? Case 7 hands you an obvious answer that is 14 percent of the problem. If you took it, you are pattern matching instead of analysing.

• Did you notice when the arithmetic answer was the wrong answer? Cases 4 and 9 both have a calculation that points one way and a commercial answer that points the other. That gap is the single most Bain-specific thing on this page.

• Did you state the sensitivity? Case 2's Option A swings from plus USD 75M to minus USD 50M on one assumption. Naming the swing is worth more than the point estimate.

• Did you name a next test? Every recommendation above ends with the thing you would check next. Bain interviewers ask "what would you want to know?" almost every time, and candidates who have not thought about it stall.

• Could you hold all of it together out loud, under time? Reading a worked case and delivering one are different skills, and the second is the one that gets scored.

That last one is the gap reading cannot close. The recommendation is the shortest part of a case and the highest-scoring one, so rep it separately from full cases.

Deliver a recommendation with a number attached from the CoachNed drill engine. Answer a real prompt and get AI-scored feedback. Free accounts include daily drills.

Bain's official practice resources

Bain publishes a small amount of material directly, and it is worth working through before anything else because it is the only content written by the people who will interview you.

• Coffee Shop Co., a written market sizing and break-even case, published on Bain's coffee case study page. Worked in full as Case 1 above.

• FashionCo., a written revenue growth case with two quantified options, published on Bain's fashion case study page. Worked in full as Case 2 above.

• Mock interview videos for the Associate Consultant and Consultant levels, linked from Bain's interviewing page. Watch these for pacing and for how Bain interviewers actually push, which written cases cannot show you.

• Bain's own statement of what it assesses: the ability to make sensible assumptions, do quick math, and build constructively on other people's ideas, with the note that there is not necessarily a right answer.

Bain does not publish a casebook PDF. Compilations circulating under that name are university consulting club material, and they are useful for prompt volume but carry no signal about how Bain scores.

How to use these ten cases

Reading worked cases is preparation for practice, not practice itself. Use each one twice: once cold with the answer covered, once as a reference after you have written your own.

For each case:

• Read only the prompt and the facts. Cover everything else.

• Write your structure in two minutes. Name mechanisms, not categories.

• Work the arithmetic on paper, keeping units attached at every step.

• Say your recommendation out loud in under 60 seconds, with a number, a risk, and a next test.

• Uncover the worked solution and find the first place your chain diverged. That divergence, not the final answer, is your practice target.

Then wire each weakness to a specific rep rather than running another full case and hoping.

Execution checklist

• ✓Structure came out as categories rather than mechanisms. Rebuild the tree using the profitability framework branches, then get one structure scored so you find out whether it reads as a template

• ✓Sizing assumptions felt arbitrary. Work the market sizing method and rep the driver chain until you can defend each assumption with a sentence

• ✓Arithmetic slipped or units drifted. Use case interview math practice for multi-step reps with the units carried through

• ✓Acquisition cases feel unfamiliar. Bain asks these more than any other firm. Work the M&A case framework, then run three acquisition cases from the case library

• ✓The recommendation rambled. Practice the four-part close: decision, number, main risk, next test. Sixty seconds, out loud, every single time

• ✓The online screen is still ahead of you. Run reasoning reps against the Bain SOVA guide format before it arrives, because a failed screen means none of the case work gets used

• ✓Experience interview stories are not built yet. Bain runs a dedicated experience interview alongside the case. Build four stories using the Bain behavioural questions guide

• ✓Application materials are not finished. Run yours through the consulting resume grader and check it against the Bain resume guide before the round starts

Two more things sit outside case practice and cost candidates offers anyway. The Bain cover letter is read more closely than at most firms because of how Bain screens for fit, and knowing the Bain salary bands before a recruiter call means you are not negotiating from a blank page. For worked examples across the other firm formats, see case interview examples.

Finally, a word on volume. Candidates ask how many cases they need. The honest answer is that ten cases solved properly, out loud, with a scored recommendation and a diagnosed weakness after each, beats fifty prompts skimmed. The ten above are enough material for three weeks of real practice if you use them that way. When you want live reps against a real interviewer voice, start a free case and deliver the recommendation out loud rather than writing it down.

Talk through a candidate-led case out loud

A conversational profitability case in the Bain mould: you open the structure, call for the facts you need, and get scored on your math, your judgment, and the recommendation you close with.

Start a free voice case

Sources and further reading (checked July 31, 2026)

• Bain & Company interviewing and hiring process: bain.com/careers/hiring-process/interviewing

• Bain Coffee Shop Co. published practice case: bain.com/careers/hiring-process/interviewing/coffee-case-study

• Bain FashionCo. published practice case: bain.com/careers/hiring-process/interviewing/fashion-case-study

• Bain careers, consulting roles and hiring overview: bain.com/careers/hiring-process

• Career in Consulting, Bain case interview assessment criteria and round structure: careerinconsulting.com/bain-case-interview

• Leland, Bain interview guide, round structure and evaluation criteria: joinleland.com/library/a/bain-case-interviews-a-comprehensive-preparation-guide

All worked solutions to Cases 3 through 10 are original practice material written for this article. Cases 1 and 2 use the figures Bain published on its own careers site, with the arithmetic and the recommendation worked here.