Study a worked case and log your own · How to use this page

Follow the seven tiles in order; each one opens the matching part of the worked case. The emulation section lists what to copy from a strong caser. Log your own cases in My Harper.

RECRUITING · CASE LOG

How does a case become a recommendation?

One case so far. Follow the decisions, see the work on paper, then practice saying it aloud.

Choose a view. Detailed notes stay available without crowding the worked example.

Practice hearing and setting up a prompt →

The one-page setup

How the caser laid out his first sheet before the prompt was read, and what he said about it afterward.

  • A notes area where the prompt is written down as the interviewer reads it. The caser replays the prompt back before asking anything, both to confirm he heard everything and to buy time to think about which questions to ask.
  • Four reminders kept beside the notes: what is the business, what is the geography, what is the objective, and whether there is a time frame. Objective is the one he asks about every single time; the others are prompts to check only when the answer is not obvious.
  • The right-hand side of the first page is reserved for the framework.
  • The hypothesis goes at the very top of the page. His reasoning: driving the case forward with the hypothesis and returning to it keeps the candidate oriented and lets the interviewer see where the candidate's head is.
  • Data pages are treated like an Excel sheet: numbers go into a table (seats, occupancy, price, per-show revenue) so the arithmetic is easy to do and easy to follow.
  • On how much space to leave, he admitted he feels it out with practice. The interviewer suggested keeping one separate sheet for final answers so values are easy to find even when working pages get messy; another facilitator said to use as many sheets as needed, write large, and not cram.
  • In a real interview he spends about a minute thirty on the framework; here he took longer on purpose so the room could read his handwriting.

Case 1 · Wharton 2017 casebook, Penn and Teller

Should Caesars re-sign Penn & Teller?

Seen live at a club workshop on October 9, 2026.

A consulting club case workshop for first-year MBA students: a facilitator framed what to watch, a second-year student ran a live demonstration case while narrating his own thinking out loud, a classmate played the interviewer, and the room then split into small-group breakout practice.

Prompt

Our client is Caesars, one of the largest gaming and resort companies in the world. Penn and Teller have headlined at the Rio Hotel since 2001 and are the longest-running headlining show in Las Vegas history. Caesars owns the Rio, where the duo performs in the 1,500-seat Penn and Teller Theater, and is wondering whether the act has become stale. The duo's annual contract is about to expire. Before Caesars meets with them, it has asked for a recommendation on whether or not to re-sign the act for another year.

Clarify

What the caser asked, and what came back. The later questions arrived mid-case, as the data did.

What exactly is Caesars solving for? What matters most as they weigh whether to re-sign the act?
Caesars cares solely about the company's total bottom-line profitability.
Is there a time frame to keep in mind for profitability, or is it the foreseeable future?
Think about the length of the contract, which, like all Vegas contracts, is a one-year period.
You said the act was stale. Can you say more about why Caesars believes that? (He flagged that odd words in a prompt are worth picking on.)
The show has plateaued. There has been no change in profitability in the last five years.
Mid-case, on revenue: beyond the number of seats, how often are they filled, what are the prices, and how many shows per year?
Occupancy 100% for A, 80% for B, 50% for C. Prices $120, $75, $55. Six shows a week for 40 weeks a year, with one day off per week and three months off for TV appearances and book deals.
Mid-case, on costs: I think I missed one of the six cost buckets. Which one?
Props, such as doves and playing cards, at $200 per show.
Housekeeping is $1,000 a week. Is that for the entire year or only the 40 weeks the show is performing?
Only the 40 weeks the show is live. Treat every other weekly or monthly expense as running 12 months a year.
After the interviewer flagged a cost line: the $2 million for Penn and Teller, is that per person or in aggregate?
Per person, so $4 million in total.
For the alternative act: is it still the same 240 shows per year, and the same contract length?
No changes to the number of shows, and the contract is the same one-year term as Penn and Teller's.

Hypothesis

Initial: because profitability has plateaued for five years, there is probably an opportunity to find something better, so Caesars should look at switching to a new, more popular show; customer preferences may be shifting. Revised after the cost math showed about $18 million of profit at nearly 80% margin: he said plainly that his hypothesis had been to switch, but the show looked pretty good and he was interested in re-signing while still testing alternatives. After the alternative act showed a four-year payback for about $600,000 a year of extra profit, he said he was leaning toward staying with Penn and Teller, and the recommendation followed that lean.

Framework

Three areas, read across and then down.

Paper lab · worked example

Build the case, one thought at a time

Follow the connections. Rehearse what you would write and say, then move to the next decision.

One-year profit · question → evidence → decision

01 / 06 · Decision

Start with the client’s decision

Write on the pageRenew the act for one year? Maximize Caesars’ total bottom-line profit.

Say it aloud

“Before I structure this, I want to confirm the objective and the one-year time horizon.”

What to watch

Underline the objective. Put the time horizon beside it, not in a separate bucket.

02 / 06 · Structure

Give each branch a question

Write on the pageCurrent economics · Alternatives · Risks

Say it aloud

“I’d establish what the current show earns, compare the alternatives, and check what could change the decision.”

What to watch

Draw three branches. Write questions under each heading; do not fill the page with a memorized list.

03 / 06 · Baseline

Put the equation under the first branch

Write on the page$22.8m revenue − $4.8m costs = $18m annual profit

Say it aloud

“The current show produces $18m on the supplied figures. That is the baseline an alternative must beat.”

What to watch

Keep units next to each number. Wider hotel and gaming effects remain open questions.

04 / 06 · Compare

Use the same time horizon

Write on the page$21.6m − $3m − $2.4m upfront = $16.2m in year one

Say it aloud

“The alternative improves recurring profit, but switching costs make it $1.8m worse in the first year.”

What to watch

Box one-time costs separately. A four-year payback does not answer a one-year decision.

05 / 06 · Brainstorm

Split ideas before listing them

Write on the pageImprove the show → pricing, seat mix, demand tests Build around it → hotel, gaming, concessions

Say it aloud

“I’d group ideas into improving the show itself and increasing its contribution to the wider property.”

What to watch

Add two short idea branches. Circle one promising test; explain impact, feasibility, and risk.

06 / 06 · Recommend

Return to the decision

Write on the pageRenew for one year on the supplied economics. The initial hypothesis weakens: the current show earns $18m, while switching yields $16.2m after upfront costs in the first year.

Say it aloud

“I recommend renewing for one year, then testing the demand and property effects that could change the answer.”

What to watch

Draw an arrow back to the objective. State the answer, the evidence, and one next step.

Compare with the complete driver map

See the reasoning unfold

Read across the questions, then follow the evidence down to the decision. This is one worked approach, not the only valid framework.

Client decisionRenew the act for one year? Maximize Caesars’ total bottom-line profit.

Starting hypothesis · to testA new, more popular show might improve profit. Test this against the current show and the cost of changing.

Start here → Current economics: establish the profit the alternative must beat before deciding whether the act is stale.

Bucket 1

What does the current show earn?
  • Revenue: seats × occupancy × ticket price × shows
  • Costs: performers, crew, housekeeping, props, utilities and SG&A
  • Wider property: hotel, gaming and concessions contribution

Test / evidence

Tickets: $95,000 per show × 240 shows = $22.8m per year. Annual costs: $4.8m, including $2m for each performer.

So what?

The supplied ticket and cost figures imply $18m annual profit. Stagnant growth alone does not prove the show should be replaced; wider-property effects still need checking.

Bucket 2

Would an alternative do better?
  • Compare another act on the same one-year horizon
  • Separate ongoing costs from one-time switching costs
  • Consider other uses for the theater

Test / evidence

New act: $21.6m revenue − $3m ongoing cost = $18.6m, before $2.4m upfront switching cost. First-year profit after switching: $16.2m.

So what?

Switching is $1.8m worse in year one on these figures. The $0.6m recurring improvement needs four years to recover the upfront cost, beyond the one-year decision window.

Bucket 3

What could change the decision?
  • Customer preferences and brand fit
  • Performer availability and retirement risk
  • Uncertainty in the alternative’s sales forecast

Test / evidence

The current show has been stable for five years. Category A sells out; category C fills only half its seats. Demand for the new act is a forecast.

So what?

Test demand and seat-mix improvements rather than assume a newer show will win. Brand effects and future availability remain questions, not measured benefits.

Revised hypothesis → recommendation

Renew for one year on the supplied economics. The initial hypothesis weakens: the current show earns $18m, while switching yields $16.2m after upfront costs in the first year.

Next step: Validate changing customer preferences, test pricing and seat mix, and check material hotel or gaming effects before finalizing the contract.

Full framework notes

Financials of Penn and Teller

He keeps revenue and cost in one bucket; splitting them is a matter of preference.

  • Revenue: tickets, meaning number of seats, capacity filled, and price
  • Revenue: other sources such as concessions
  • Revenue: connections to other parts of the property, such as hotel and gambling
  • Costs: the contract with Penn and Teller
  • Costs: variable costs per show, such as stagehands, utilities, cleaning
  • Costs: fixed costs, such as marketing and SG&A

Alternatives to Penn and Teller

Kept shorter, since much of it mirrors the first bucket.

  • Other shows and their profitability
  • Switching costs
  • Repurposing the space: hotel, gaming, arena

Other considerations

A catch-all for risks and benefits.

  • Brand, for both Caesars and Penn and Teller
  • Will the duo retire soon; do we need to think about a new act beyond the one-year window
  • Customer preferences and how they are evolving (added last, before he read the framework out)

Numbers

Every exhibit in this case was given verbally. Each card is one data drop: what it showed, the arithmetic as the caser talked it through, and the takeaway.

One-year profit bridge · case assumptions
Revenue$22.8m−Costs$4.8m=Current profit$18m

Alternative: $18.6m annual operating profit − $2.4m upfront switch = $16.2m in year one.

Decision: the alternative earns $0.6m more annually, but takes four years to recover the switch. Compare on the client’s one-year horizon.

Revenue inputs, given verbally

1,500 seats in three categories: 300 category A (best seats), 800 category B, 400 category C (balcony and back). Occupancy averages 100% for A, 80% for B, 50% for C. Prices are $120, $75, and $55. Six shows a week, 40 weeks a year.

Walk through the calculation

He first called out that A sells out while C is only half full, a possible opportunity for later. Then per show: A, 300 seats times $120 is 30,000 plus 6,000, so $36,000. B, 800 times 80% is 640 seats; 640 times 100 is 64,000 and 75% of that is $48,000. C, 400 times 50% is 200 seats; 200 times 50 is 10,000 plus 200 times 5 is 1,000, so $11,000. Total $95,000 per show. Shows per year: 6 times 40 is 240. Annual revenue: 95,000 times 240, multiplied out longhand, is $22.8 million.

Takeaway. A strong-looking number, but meaningless without the cost side. He asked to move to costs to judge whether $22.8 million is good or improvable. Method notes: talk the math through out loud, use shortcuts where they exist, and lay the data out as a table.

Cost inputs, six buckets, given verbally

Penn and Teller each earn $2 million a year. Crew (stagehands, showgirls, ushers) $2,000 per show. Housekeeping $1,000 per week, only for the 40 live weeks. Props $200 per show. Utilities, mostly lights, $52,000 per year. SG&A including marketing and box office staff $15,000 per month.

Walk through the calculation

Crew: 240 times $2,000 is $480,000. Housekeeping: 40 times $1,000 is $40,000. Utilities $52,000. SG&A: 15 times 12 is $180,000. Props: a tenth of crew, so $48,000. The non-contract items sum to $800,000. He deliberately booked the duo at $2 million total, reaching $2.8 million, and said that seemed crazy low against $22.8 million. The interviewer said one cost line was off; he rechecked, asked whether $2 million was per person, learned it was, and corrected to $4.8 million total.

Takeaway. Profit of about $18 million on $22.8 million, nearly an 80% margin. He said his hypothesis to switch now looked shaky and the show looked great. On the mistake: do not apologize, thank them for the clarification, and keep moving; on the job there is a spreadsheet in front of you.

The alternative act, given verbally after the brainstorm

Caesars could replace the show with one featuring the recent winner of America's Got Talent. Same one-year contract. The new show costs $3 million a year all in. Switching requires reconfiguring the theater plus a one-time marketing blitz, $2.4 million, incurred before the first show. The new show would sell $90,000 of tickets per show; gaming and concessions revenue would be insignificant. Still 240 shows.

Walk through the calculation

He worked in deltas rather than redoing everything. Revenue: $95,000 to $90,000 is $5,000 less per show; 5,000 times 200 is 1 million and 5,000 times 40 is 200,000, so $1.2 million less, taking revenue from $22.8 million to $21.6 million. Costs: $3 million against $4.8 million saves $1.8 million a year. Net: about $600,000 more profit per year. Payback: $2.4 million divided by $600,000 is four years. His rule of thumb is that roughly three years is fine and four is getting long. (He misspoke once, calling the switching cost annual; the interviewer had defined it as one-time.)

Takeaway. Caesars would gain about $600,000 a year but wait four years to recover the switching cost. He was not compelled for a show already this profitable and stable, and said he was leaning toward staying with Penn and Teller. Whenever a case mentions switching or upfront costs, go straight to break-even.

Brainstorm

When the interviewer asked how to make the current show more profitable, he asked for 30 to 60 seconds, grouped ideas into two buckets (the show itself, and around the show), and admitted he was bad at brainstorms when he started.

01 · Start with the objective

How could we improve the current show’s profit?

Use the evidence: premium seats sell out; the lowest-price section is half full.

Improve the show

  1. Change the seat mix: move seats from category C, which only half fills, into category A, which sells out at the highest price. He tied this back to the observation he had flagged before doing the revenue math.
  2. Test price elasticity: can prices be raised in any segment or for any shows.
  3. Add shows: the act only performs 40 weeks a year, so ask whether the duo would do more.

Test next → Test willingness to pay and demand before changing seats, prices or show frequency.

Build around the show

  1. Around the show: bring in new concessions.
  2. Around the show: add an opening act beforehand to bring more people in.
  3. Around the show: bundle tickets with the hotel.

Test next → Test whether concessions, an opening act or hotel bundles add profit after their extra costs.

03 · Prioritize and test

Choose the idea with the strongest profit potential

Compare incremental revenue, extra cost, feasibility and risk. State what evidence would change your mind.

Continue to the recommendation →

Recommend

Stay with Penn and Teller. Reasons: the show is highly profitable today, about $18 million in profit on $22.8 million of revenue, and it has been very stable. The new act is only a hair more profitable and would take four years to pay back its $2.4 million switching cost, a long and risky change built on assumptions about how it would perform. Risk one: customer preferences may be evolving; next step, a customer survey on entertainment preferences. Risk two: Caesars cares about profitability and the current show is technically a little less profitable than the alternative; next step, pursue the brainstorm ideas such as improving the seat mix. He closed by thanking the interviewer.

How to sound like that in the room

Talk the interviewer through your head like a colleague you respect: unhurried, candid about what you do not know, naming the next move before you make it, and never apologizing for the dust.

Demeanor · 7 examples
  • Checks the room before anything else and shrugs off friction: he asked whether people could hear him, took a thumbs-up, said the mic felt weird and that they would do it anyway; when the projector cut off part of his page he said that was most of it and good enough.
  • Self-deprecating about everything except the work: not used to the mic, less experience than the organizers, horrid handwriting, sucked at brainstorms when he started, took three minutes on frameworks early on, and puts a customer survey in 75% of his recommendations.
  • Opens the case with a light personal aside instead of jumping to questions: he had visited the property years ago, great property, never saw the show, maybe next time. About ten seconds, then straight into the replay.
  • Calm confidence that never tips into certainty: he states leans, not verdicts (I am leaning towards staying) and says I will be honest before changing his mind in front of the interviewer.
  • Treats the interviewer as a colleague on the same team: thanks her after every data drop, says makes a lot of sense, asks permission before taking time or switching topics, and says love the case, had a great time at the close.
  • Blunt and warm at the same time: thank your caser even if the case went badly, be a good person, it matters. Closes the hour with thanks for listening to me talk at you and an offer to help on Slack.
  • Comfortable saying I do not know in public: on page spacing he said he had no mind-blowing answer and had not thought much about it; on pacing he opened with the consultant answer, it depends.
How he talks · 15 examples
  • Replay before questions: what I heard from you is, client, situation, the question in their mind. He says the replay also buys him time to decide what to ask.
  • Clarifying questions open softly and name their purpose: the first question I always ask is getting more detail on the exact objective; we would love to hear a little bit more; the last thing I would love to hear.
  • Picks on the odd word in the prompt and says why: one thing you said was stale, that was a bit weird, so I just want to hear a little more about why.
  • Labels the case type out loud after the objective lands: I have my objective here, this is going to be a profitability case, so that will inform a lot of my framework.
  • Asks for thinking time with the same two-part line every time: I think I have what I need, I am going to take a few moments and jot down some thoughts, then I will come back to you.
  • Framework readout is hypothesis first, then a count, then the list, then depth: my hypothesis is X; to answer that there are three things we want to look into; most importantly I will start with A, then dig into B, and lastly touch on C; within A, things like. Ends with but we will come back to that if we need.
  • Drives the case with an explicit next step and a request: I would love to start by understanding the financials; maybe you can start on the revenue side if you have information there.
  • Announces the calculation plan before touching a number: I will calculate revenue per show, then gross it up to the annual figure; I will start with segment A, work across, then come down.
  • Narrates arithmetic in parts: 300 times 100 is 30,000, 20 times 300 is 6,000, add those and that is 36,000; 640 times 100 is 64,000 and 75% of that is 48,000.
  • Closes every number with a check-in: does that sound good to you; how does that number sound to you; does that sound a little better to you.
  • Puts each number back against the objective before moving: off the bat that seems like a good number, but I know profitability has plateaued, so I would love to understand the cost side to put it in context; if it is okay, let us jump to costs.
  • Business-sense sentence after the big number, with the margin stated: 4.8 million in costs on 22.8 million in revenue, that is 18 million in profit, nearly 80% margin.
  • Reflex phrase for investments: whenever you hear switching costs your mind should go to when do we break even, followed by a rule of thumb, three years is okay, four is getting long.
  • Recommendation follows one script: our recommendation is X, and the reasons why; that said, it is not without risks, the first is, so the next step is; the second is, so the next step is; love the case, happy to answer more questions.
  • Uses we for the team and I for his own thinking: we would love to hear, our recommendation; I am leaning, I will be honest. Transition words are cool, awesome, sweet, love that, and they do a lot of the pacing work.
Signposts · 22 examples
  • What I heard from you is...
  • The first question I always ask is the objective.
  • One thing you said was a bit weird: stale.
  • I think I have what I need; a few moments, please.
  • My hypothesis is...
  • To answer that, there are three things I want to look into.
  • Most importantly, I want to start with the financials.
  • We will come back to that after the financials.
  • I would love to start on the revenue side.
  • One thing that quickly stands out...
  • Per show first, then gross it up to the year.
  • Does that sound good to you?
  • If it is okay, let us jump to the cost side.
  • I think I missed one of the buckets.
  • The one I want to clarify quickly is...
  • Thanks for the clarification. Let me go back through each.
  • I will be honest, my hypothesis was to switch, but...
  • Switching costs: when do we break even?
  • As of now, I am leaning towards staying.
  • Our recommendation is... and the reasons why...
  • That said, it is not without risks.
  • Love the case. Happy to answer more questions.
Page and hand · 9 examples
  • First page before the prompt: notes area, four reminders (business, geography, objective, time frame), framework space on the right, hypothesis line at the top.
  • Writes the hypothesis last, after the three buckets, as the final thing before turning back to the interviewer. He also adds a forgotten sub-point, customer preferences, at that moment rather than pretending it was there.
  • Writes a bucket label and moves on when short on time: if you have time jot variable costs down, if not just write variable and elaborate later.
  • Treats data pages like a spreadsheet: seats, occupancy, filled seats, price and revenue in a row per segment, so the per-show total is a column sum.
  • Does longhand multiplication off to the side of the page when no shortcut exists, then carries the result back into the table.
  • Rewrites a number when it comes out illegible rather than leaving it (the 240 shows).
  • Keeps earlier pages within reach for the recommendation and reads the headline numbers off them, 18 million profit, four-year payback.
  • Brainstorm page is two labeled groups, the show itself and outside the show, with three ideas under each.
  • Recommendation page is a fixed shape: recommendation on top, two reasons with a number, then risk and next step pairs.
With numbers · 9 examples
  • Every exhibit in this case was verbal. On the first data drop (seat counts) he acknowledged it, then asked for the exact complementary inputs needed before any math: how often the seats fill, at what price, how many shows per year.
  • States a so-what before any calculation: A is at 100% and those are the most expensive seats, C is only 50%, that mix is a possible opportunity for later. Only then: but first, let me calculate.
  • Announces the structure of the calculation, per show then annual, left to right then down, before starting.
  • Looks for a shortcut before brute force: 75% of 64,000 rather than 640 times 75; props as a tenth of crew because 200 is a tenth of 2,000; deltas against the already-computed 22.8 million rather than a fresh build.
  • Checks the number with the interviewer and immediately interprets it against the objective before asking for the next piece.
  • When six items arrive at once, repeats the list back and names the one he missed.
  • Checks the time basis of any rate: housekeeping per week, is that 52 weeks or the 40 live weeks.
  • For the alternative act, asks what else changes before computing (same number of shows?) so the delta method is valid.
  • On rounding, asks first whether rounding is okay; some casers want precision.
When stuck · 8 examples
  • Missed an input: said so plainly, I think I missed one of the buckets, and asked for it rather than guessing.
  • Told a number was off: no apology. Okay, thanks for the clarification, let me go back and look through each of these cost items. He re-read every line until he found the ambiguous one, asked the clarifying question, corrected the total, and checked in again.
  • Hypothesis stopped holding: said it out loud and reset the lean while keeping the test alive, I will be honest, my hypothesis was to switch, but this seems pretty good; I still want to see the alternatives.
  • No mental shortcut available: named that too, I do not have a great shortcut off the top of my head, so I am just going to multiply this out, and did it.
  • Interviewer said there was no more information: accepted it and moved straight into the recommendation instead of pushing.
  • Uncertain mid-case: his advice is to say so, I am not sure, it could go either way, I want to keep investigating.
  • Hitting a no on every bucket: step back, say it sounds like something is missing, return to the framework.
  • Pre-emptive framing for the room: we will all make math mistakes, what matters is how you respond.
With the interviewer · 9 examples
  • Asks permission before taking time, three times, and says when he will be back.
  • Asks permission before changing topic: if it is okay, we would love to jump to the cost side.
  • Acknowledges every answer with a short affirmation and a thank you before building on it.
  • Treats hints as help: okay, no, thanks for the clarification, then acts on it immediately.
  • Reads the interviewer's cues as instructions: when she asked for options to make the show more profitable he named it as an invitation to brainstorm; when she asked for a formal recommendation he said that means time to wrap up.
  • Offers depth without forcing it: happy to dig into any of those more if you want.
  • Keeps the fourth-wall narration separate from the interview: by the way, I would not be voicing this over to the interviewer, this is just for you all.
  • Thanks the interviewer explicitly at the end and tells the room to do the same no matter how the case went.
  • Eye contact: the facilitator flagged it as awkward in a demo; the caser kept turning back to the interviewer at each check-in and invited her to repeat anything, which she also offered.
Pacing · compare the stages

After the prompt was read there were about seven seconds of silence, then the replay, roughly half a minute. Three clarifying questions took about 95 seconds. He asked for time and wrote the framework in about three minutes, slower than his usual minute thirty so the room could read it, then read it out in two minutes. The revenue math was narrated at a steady pace over about three minutes, slowing only for the longhand 95,000 times 240. The cost math took about three minutes including the correction. Brainstorm: 55 seconds writing, 80 seconds delivering. The alternative act, worked in deltas, took under two minutes. For the recommendation he asked for time, prepared for a little under two minutes, and delivered in just under two. About 32 minutes end to end.

Drills

Each drill isolates one habit from the case above. Run them with a partner and a timer.

  1. Replay drill: a partner reads any prompt once; restate client, situation and question in under 30 seconds, then ask one objective question and one time-frame question, in that order.
  2. Odd-word drill: for ten casebook prompts, underline the one loaded word (stale, struggling, considering) and write the clarifying question it earns and why.
  3. Framework timer: write a three-bucket framework with the hypothesis on top in 90 seconds, read it across then down in under two minutes, then rewrite the framework after every practice case as he did when it was his weakest step.
  4. Narrated math: take a seats, occupancy, price table and compute per-show then annual revenue aloud, naming every shortcut you use, ending with does that sound right to you.
  5. Planted mistake: a partner builds one ambiguous input into the data (per person versus aggregate, per week for 40 versus 52 weeks) and you practice the recovery, thanks for the clarification, let me go back through each line, without saying sorry.
  6. Break-even reflex: every time a prompt mentions an upfront or switching cost, compute payback before anything else and state your rule of thumb for an acceptable horizon.
  7. Sixty-second recommendation: answer, two reasons with one number, two risks each paired with a next step, then thank the interviewer; run it on the last three cases you did.

What the facilitator corrected

Points raised in the debrief and in the breakout rooms, kept as they were given.

  • A hypothesis is a hypothesis, not an answer. Keep referring back to it as information arrives, and say out loud when you are leaning one way or when it could go either way.
  • It is not held against you if the case ends somewhere other than your hypothesis. Changing it as evidence comes in can read as a good sign, as long as the starting hypothesis rested on a reasonable assumption. Always start with one.
  • Lock in on the objective first and keep asking until you know exactly what you are solving for. Here the objective was purely bottom-line profitability, so he did not ask why Caesars liked the alternative act; a secondary objective would have changed that.
  • Clarify the time frame up front. The case was bounded to one year, but any upfront investment should still trigger a break-even question: how long until we make this back.
  • On rounding, ask the interviewer whether rounding is acceptable; sometimes they want precision. The interviewer added that casers work from an answer sheet, so rounding can produce a number they cannot confirm.
  • Page spacing comes with practice. Keep one sheet for final answers so values are easy to find even when the working pages are messy. Use as much paper as needed, write large, do not cram.
  • There are many ways to case well. Do not copy anyone's setup if it does not work for you; find your own style.
  • Buckets you never reach are fine. The interviewer ends the case when the answer sheet is covered, and some cases make you walk through several buckets getting no before one opens up. If you get a no on everything, step back, say you may be missing something, and revisit the framework. Casers want you to succeed.
  • Business, geography, objective and time frame are reminders, not required questions. Do not ask about geography when it is obvious; a facilitator called that a filler question that does not land well.
  • Pacing: a minute thirty to two minutes on the framework, 30 to 60 seconds on brainstorms and on the recommendation, 30 to 40 minutes for the whole case. Find your weakest step and drill it; he rewrote a framework after every practice case, and others drilled mental math on their phones.
  • Keep clarifying questions to three or four. Burning time up front shortchanges the math and brainstorms at the back, but still make sure you know the product and the objective before moving on.
  • Casing is pattern recognition. A tangible product behaves like other tangible products; find comparables, and the flow of cases becomes familiar with repetition.
  • Math mistakes: do not apologize. Thank them for the clarification, fix it, and keep moving. Talk through your math so the interviewer can follow and correct you, and use shortcuts: 75% of 64,000, props as a tenth of crew, computing the delta instead of redoing the whole calculation.
  • Always thank your interviewer at the end, however the case went.
  • Breakout facilitators added: say areas or segments, never buckets, because you interview the way you practice; read a framework across and then down, in bullets rather than sentences; a recommendation is the answer, one quantitative and one qualitative piece of evidence, one or two risks, one or two next steps, in about 45 seconds; and build the recommendation as the case goes so the 15-second prep is not a scramble.
  • Program advice: cast a wide net of 10 to 15 firms rather than three or four, and keep every connection for full-time recruiting.

What to watch

The facilitator framed three things to watch during the demonstration. Each card pairs that framing with what the caser did.

Flow

The facilitator asked the room to watch the natural progression from the interviewer's introduction through clarification and questioning, the framework, math or exhibits, a brainstorm, and the recommendation. The caser followed that arc exactly: he replayed the prompt, asked three clarifying questions, took time to build the framework, drove to the financials himself (revenue first, then costs), brainstormed when invited, tested the alternative act against his hypothesis, and delivered a recommendation when the interviewer asked for one. He stressed pushing the case forward at every step until the interviewer calls for the recommendation.

Structure

The facilitator asked the room to watch how the caser used scratch paper, how things were segmented, how information stayed easy to reach, how the framework was built and each element approached, and whether exhibits followed a pathway. The caser's first page had a notes box with business, geography, objective and time frame, the hypothesis on top and the framework on the right; revenue inputs went into a table; the framework was read across its three buckets and then down into each; the brainstorm was split into two groups; the recommendation followed a fixed shape of answer, reasons with a number, then risks each paired with a next step.

Presence

The facilitator asked the room to watch how the caser brought his personality into the case and how he communicated with the interviewer, noting that eye contact is awkward in a demo. The caser opened with a personal aside about having visited the property, called out a business observation on the seat mix before touching the math, talked through every calculation out loud, handled the deliberate math mistake without apologizing, said honestly when his hypothesis stopped holding, and thanked the interviewer at the end. He described the goal as showing business sense alongside personality.

Log your own

The same record shape is available in My Harper as a private case log: date, source, partner, your role, then the seven flow steps, what went well, what to fix next time, the feedback you received, and a rating. Logged cases stay private to your account. The download there is an .xlsx workbook with one row per case.

Open my case log →