Reading a free sample case is not practice. In a typical first round of eight candidates, the ones who clear the bar have already done dozens of short, timed reps where someone—or a sheet—told them what they missed. Passively walking through McKinsey's Electro-Light or Bain's Coffee Shop Co. teaches the format; it does not train the muscles. This piece is the other thing: ten original drills you can run today, with answer keys, a scoring sheet, and a weekly rotation.
After you finish, you will be able to run a 45-minute skill block without a partner, know which skill leaked points, and know which free surfaces (firm pages and CoachNed) actually grade or model the work.
What free practice is actually worth
Firms publish free materials because they want prepared candidates, not because reading equals rehearsal. McKinsey's interviewing page links sample problem-solving cases with logic and suggested answers. BCG's case preparation page stresses structure, calculation, and practical recommendations, and points to free simulations such as OneDay@BCG. Bain's case interview page ships its own Coffee Shop Co. and FashionCo. walk-throughs and tells you to think aloud, structure, and adapt.
Use those pages for format fluency. Use drills for speed and accuracy. The interview sheet rewards both.
Ned's rule. If a free "drill" has no timer, no answer key, and no score, it is reading. Score every rep for 90 seconds. Patterns show up by rep four; volume without a sheet does not.
How to run a drill (90 seconds of process)
- Set a timer (times below).
- Speak aloud or type; silence hides sloppy math.
- Stop when the timer ends—even mid-sentence.
- Score yourself against the key using the sheet in the next section.
- Log one leak (units, MECE, so-what, rounding) and move on.
Do not spend ten minutes "checking" online. The answer key is enough.
Scoring sheet (print or copy)
Score each drill 0–2 per row. Max per drill = 8. A clean pass is 6+.
| Dimension | 0 | 1 | 2 |
|---|---|---|---|
| Structure / approach | No clear buckets or order | Partial buckets; some overlap | Clear, MECE, ordered |
| Accuracy (number or fact) | Wrong or missing | Close; one slip | Correct within tolerance |
| Units / labels | Missing or mixed | Mostly right | Explicit throughout |
| Insight / so-what | Restates the prompt | One implication | Client-ready implication |
Weekly target: average score ≥6 across the day's drills, and zero 0s on units.
Ten free drills (with keys)
Work these cold. Company names and figures are invented for practice.
Structure (two drills, 90 seconds each)
Drill 1 — HarborLink margin drop. HarborLink operates three regional ferry routes. Operating profit fell 18% year over year while revenue rose 4%. In 90 seconds, lay out how you would diagnose the drop. Do not solve it.
Key. Revenue up / profit down means costs or mix. Buckets that work: (1) volume vs. price vs. mix by route, (2) variable cost per trip (fuel, crew, maintenance), (3) fixed cost step-ups (new vessel lease, dry dock), (4) one-offs (strike, weather cancellations). State you would start with contribution margin by route. Overlap fails: "operations" and "costs" as separate top buckets.
Drill 2 — NovaShelf private label. A grocery chain wants to decide whether to expand private-label snacks into frozen. Structure the decision in 90 seconds.
Key. Decision tree or MECE: (1) demand / willingness to pay in frozen snacks, (2) unit economics (COGS, shelf life, spoilage, slotting), (3) capability (cold chain, supplier, brand trust), (4) competitive response and category cannibalization of national brands. End with a hypothesis gate: expand only if contribution after spoilage clears a hurdle and cannibalization stays below X%.
Mental math (two drills, 45 seconds each)
Drill 3 — Pack size. A snack SKU sells 2.4 million units at $1.25 wholesale. Gross margin is 40%. What is gross profit dollars?
Key. Revenue = 2.4e6 × 1.25 = $3.0M. Gross profit = 0.4 × 3.0M = $1.2M. Tolerance: ±$0.05M if you approximate 2.4 × 1.25 as 3.
Drill 4 — Rule of 72. Category volume grows 9% per year. About how many years to double?
Key. 72 ÷ 9 = 8 years. State the formula and the unit (years). Bonus: note Rule of 72 is an approximation.
Exhibit reading (two drills, 60 seconds each)
You do not need a chart image. Treat the table as the exhibit; say the so-what out loud.
Drill 5 — Region mix.
| Region | Revenue ($M) | YoY growth | Margin % |
|---|---|---|---|
| North | 40 | +3% | 22% |
| South | 25 | +18% | 9% |
| West | 35 | −2% | 19% |
Prompt: Where would you push growth next quarter, and why? Sixty seconds.
Key. South is the growth engine but thin margin (9%). North is large and profitable but slow. West is shrinking. Insight: do not blindly "invest in South"—test whether South's growth is discount-led. Better so-what: protect North margin while raising South price or mix; diagnose West churn. Fail: "grow South because growth is highest" with no margin comment.
Drill 6 — Channel shift. Online share of orders rose from 20% to 35% of units; online contribution margin is 8 points below store. Revenue flat. What happened to blended margin, directionally?
Key. Mix shift toward a lower-margin channel with flat revenue implies blended margin compressed. Rough math: online weight +15 points × (−8 pp) ≈ −1.2 pp on blended margin if other mix held. State assumptions (constant within-channel margins). Fail: claiming margin rose because "digital is efficient" without the given.
Market sizing (two drills, 3 minutes each)
Drill 7 — Campus laundry. Annual market for coin-operated laundry machines sold to U.S. university housing operators (machines, not wash cycles). Top-down.
Key (illustrative path). ~4,000 degree-granting institutions × ~40% with on-campus housing ≈ 1,600. Average housing beds ~2,500 → ~4M beds. Machines: ~1 per 20 beds → ~200,000 installed base. Useful life ~10 years → ~20,000 replacements/year. Plus new beds growth ~2% → ~8,000 incremental. Total units ~28,000/year. ASP ~$3,000 → ~$84M/year. Score the logic chain and stated assumptions harder than the final number. Tolerance: order of magnitude $50–150M.
Drill 8 — Pet dental chews. U.S. annual retail spend on dog dental chews.
Key. ~90M dogs (round) × 60% of owners buy treats regularly × 40% of those buy dental-specific × $8/month × 12 ≈ 90M × 0.24 × $96 ≈ ~$2.1B. State each filter. Fail: jumping to a number with one population × ASP and no purchase frequency.
Synthesis (two drills, 60 seconds each)
Drill 9 — FleetCo. You found: utilization 71% vs. 82% peers; maintenance cost/mile +18%; NPS flat. CEO asks for a recommendation. Sixty seconds, recommendation-first.
Key. Recommendation: prioritize utilization recovery before cutting maintenance rates. Rationale: idle fleet and peer gap dominate; higher maintenance may be a symptom of age mix or reactive repairs. Next steps: segment idle assets by region/age; pilot predictive maintenance on the worst 20% of assets. Risks: utilization push that spikes breakdowns. Fail: listing findings without a clear "do this / not that."
Drill 10 — BrightMeter pricing. Pilot raised price 12%; volume −4%; churn unchanged; competitor matched in two of five cities. Synthesize.
Key. Net revenue effect roughly +7.5% where unmatched (1.12 × 0.96 − 1). Recommend keep the increase where competitors have not matched; hold or add value in matched cities; monitor churn weekly for four weeks. Fail: "pricing worked" without geographic split.
A weekly rotation you can actually keep
| Day | Focus | Drills | Time |
|---|---|---|---|
| Mon | Structure | 1–2 | 15 min + score |
| Tue | Math | 3–4 + five more from /drills/quick-math | 20 min |
| Wed | Exhibits | 5–6, then one firm sample exhibit | 25 min |
| Thu | Sizing | 7–8 | 25 min |
| Fri | Synthesis | 9–10, then one full recommendation aloud | 20 min |
| Sat | Firm case | One McKinsey / BCG / Bain free case, timed | 30–45 min |
| Sun | Weak-spot only | Re-run any drill that scored under 6 | 20 min |
Swap Friday and Saturday if your mock partner is only free midweek. Do not skip the score log.
Where free graded practice lives (honest map)
CoachNed keeps /drills/quick-math free forever—timed mental math with no account wall. /start is a five-minute first case rep (three typed turns, instant scores) without an account. Everything else on the drill engine—/drills/structure, /drills/math, /drills/charts, /drills/market-sizing, /drills/synthesis, /drills/brainstorming—opens for a free week with no card; then $120 for a recruiting season or $49 a month.
Pair that with the firms' own free pages above. Work Electro-Light or Coffee Shop Co. for interviewer voice; use the ten drills here for reps per skill. CoachNed is independent and not affiliated with McKinsey, BCG, Bain, Deloitte, or any firm named in this article.
Practice this today (45 minutes)
- Run Drills 3, 4, 5, and 9 cold. Score each.
- Open
/drills/quick-mathand clear five prompts. - If you want one graded structure or chart rep under the free week, use the matching path under
/drills/. - Log your lowest dimension. Tomorrow's block starts there—not with a random full case.
Frequently asked questions
Are free firm practice cases enough to prepare?
No. They teach format and expectation. They do not give you enough timed, skill-isolated feedback. Use them once you can already structure, compute, and synthesize under a clock.
How many drills per day before a first-round case interview?
Aim for four to six short drills or one skill block from the weekly table, five or six days a week, for three to four weeks. Full cases two or three times a week once basics clear 6+ on the sheet.
Do I need a partner for these drills?
Not for the ten above. Partners help on full cases and pushback. For structure, math, exhibits, sizing, and synthesis micro-reps, a timer and a sheet beat an untrained partner who only says "sounds good."
What is free forever on CoachNed versus the free week?
/drills/quick-math stays free. /start works without an account. The six skill drills under /drills/ are open for seven days with no card, then paid access as above.
Should I memorize frameworks instead of drilling?
No. Interviewers spot recycled MECE trees in the first two minutes. Drill custom structures on prompts like HarborLink and NovaShelf until buckets fit the facts.
Sources
- Interviewing at McKinsey — official problem-solving interview guidance and sample cases. Checked 2026-09-24.
- BCG Case Interview Preparation — what BCG says cases test and free simulation pointers. Checked 2026-09-24.
- Bain Preparing for the Case Interview — Bain sample cases (Coffee Shop Co., FashionCo.) and interview tips. Checked 2026-09-24.
- Bain Coffee Shop Co. — worked market-sizing / break-even style practice case. Checked 2026-09-24.
- McKinsey Electro-Light sample case — official launch / break-even style practice case. Checked 2026-09-24.