Case competitions compress a ten-week group project into 24 hours, and the compression changes which mistakes are fatal. Analytical weakness is survivable. Losing three hours to an unresolved scoping debate is not.
The teams that consistently place well are rarely the smartest in the room. They're the ones who decided their logistics before the case dropped.
Set the time budget before you read the case
Write the schedule down first, with clock times, and appoint someone to enforce it. A workable split for a 24-hour case:
- Framing (~20%). Read, agree the question, pick the structure.
- Analysis (~35%). The actual work.
- Deck build (~25%). Including the exhibits.
- Rehearsal and Q&A (~20%). Non-negotiable.
The failure mode is universal: analysis expands, everything downstream compresses, and the team presents a deck they finished twenty minutes ago and have never said out loud. The judges see a team that hasn't rehearsed. It's obvious from the first slide.
Treat the rehearsal block as immovable in the same way you'd treat the submission deadline. It effectively is one.
Commit to an answer early, then test it
The strongest discipline in case work is also the most uncomfortable: pick your recommendation in the first quarter of the time, before you have the evidence to be confident.
This feels backwards. It isn't. A provisional answer turns unbounded exploration into a testable claim—you now know which analyses matter, because they're the ones that would confirm or kill your recommendation. Teams that "follow the data" without a hypothesis produce four disconnected analyses and assemble a recommendation from them at hour twenty.
The rule that makes this safe: the answer is explicitly provisional, and anyone can challenge it with evidence up to a stated deadline. After that deadline it's locked, and the remaining time goes to supporting it rather than reopening it.
Assign roles for the specific hours
Generic role assignment doesn't survive contact with a deadline. Assign to the phase:
- The framer owns the question and the structure in the first hours, then becomes a reviewer.
- Two analysts own the quantitative work with a clean division—not "the model" jointly, but "revenue build" and "cost side."
- The deck owner starts building skeleton slides during analysis, not after. This is the role teams start too late, every time.
- The integrator keeps the story coherent and is the one person allowed to say the deck contradicts the model.
The deck owner starting early is the highest-leverage item on that list. Slides built from a finished analysis take twice as long as slides whose skeleton already exists and just needs numbers dropped in.
Freeze the deck before the rehearsal
Pick a time—typically 90 minutes before submission—after which nobody touches the file except one named person fixing typos.
Late edits are how teams break their own decks: a number changes on slide 8 and no longer matches slide 12, someone reformats an exhibit and the labels vanish, the file corrupts at hour 23. The marginal insight added in the last hour is almost never worth the risk of presenting something you haven't seen in its final form.
After the freeze: full run-through, out loud, standing up, with timing. Not a talk-through of what you'd say. The actual thing.
Prepare Q&A like it decides the outcome
Because it often does: when two teams present comparable analysis, the judges' questions are where they separate—and it's the segment teams prepare least.
Spend a dedicated block generating the twenty questions you least want to be asked. Assign each to an owner who'll take it live. The categories that come up almost every time:
- What's your key assumption, and what happens if it's wrong?
- Why not the obvious alternative you didn't choose?
- What does this cost, and where does the money come from?
- What's the first thing you'd do Monday morning?
- What would make you abandon this recommendation?
The last one separates teams. A team that can name its own kill criteria reads as having thought seriously; a team that defends its answer against every challenge reads as brittle.
One rule for the room: never contradict a teammate in Q&A. If you disagree with what was just said, build on it—"and to add to that"—and resolve it afterwards. Judges forgive a debatable recommendation. They do not forgive a team that clearly hasn't agreed on its own assumptions.
The logistics nobody plans
Small things that cost real time under pressure:
- One file, one editor at a time, or a tool with genuine simultaneous editing. Version conflicts at hour 20 are how teams lose work.
- Fixed formatting decided at the start. Font, colours, exhibit style. Reformatting five people's slides at hour 21 is a pure loss.
- Food and a real break scheduled. Twelve hours in, the marginal hour of tired work is worth less than the twenty minutes it costs to be functional afterwards.
- A submission dry run. Know the format, the file size limit, and the portal before you're uploading with four minutes left.
Afterwards
Do a twenty-minute retrospective the next day, while it's fresh—what worked, what you'd change, one thing to do differently next time. Teams that compete repeatedly get better mostly through this, not through getting smarter. The format that keeps it from becoming a blame session applies directly.
And write down two or three specific moments while you still remember them. Case competitions generate unusually good behavioural interview material—high pressure, real trade-offs, visible decisions—but only if you capture them before the details blur into "we did a case competition once."