The case study is the round that decides hedge fund hiring, and it is the one candidates prepare for least specifically. It is not a modelling test with a longer deadline. It is a simulation of the job, and it is marked on the quality of the judgement rather than on the completeness of the spreadsheet.
The formats you will meet
| Format | Time given | The real test |
|---|---|---|
| Bring your own pitch | Prepared in advance | Whether you have an original view and can defend it |
| Assigned company, take-home | 2 to 7 days | Research process, judgement, and written communication |
| On-site timed case | 3 to 5 hours | Prioritisation under pressure; what you do first |
| Model build from filings | 2 to 4 hours | Technical accuracy and speed |
| Earnings reaction case | 1 to 2 hours | Reading new information and changing your mind coherently |
| Read and react | 30 to 60 minutes | Whether you can form a view from a document quickly |
What they are marking
A fund does not need another model. It has models. What it is buying is the judgement of the person who built it, and the case is the only part of the process that shows judgement at work rather than described.
- Did you reach a recommendation, or hedge? A case without a clear long or short and a price target has failed the first test whatever its analysis.
- Did you find the two or three things that matter, or model everything equally? Discrimination is the skill.
- Is the work honest about what you do not know? A stated assumption beats a hidden one every time.
- Can you defend it under pressure, including changing your mind when given a genuinely new fact?
- Is the writing clear? Funds are small and communication is a real constraint on a team of ten.
How to spend a three-day take-home
The default failure is spending all of it modelling. A rough allocation that works better:
- First few hours: read the last annual report, the most recent two results releases and the most recent transcript. Take a provisional view before touching a spreadsheet.
- Next: understand the industry well enough to know where the profit pool sits and who else competes for it. Competitor filings are the fastest route.
- Then: build a model that is deliberately simple. Revenue drivers, margin, capital intensity, and a valuation. Three sensitivities, not thirty.
- Then: attack your own thesis. Find the strongest argument against it and decide whether it changes your mind.
- Last portion: write. Reserve a genuine share of the time for the memo, because it is the deliverable and the model is only the working.
The model: simple, transparent, and checkable
Nobody is awarding marks for a forty-tab workbook. They are looking for a model somebody else could pick up, understand in five minutes and disagree with specifically.
Colour inputs consistently, keep assumptions on one sheet, never hard-code a number inside a formula, and put a short note beside any assumption a reader would question. If your model contains a number you cannot source or justify, either source it or replace it with an explicit assumption. A case study that gets caught with an invented figure is over, regardless of everything else in it.
A company you have never heard of
This happens deliberately. The fund wants to see how you approach unfamiliarity, because that is the job.
Work in a fixed order: what does it sell and to whom, how does it get paid, what does the profit and loss look like against peers, what does the balance sheet allow, who competes, what does the market currently expect, and where could that expectation be wrong. Say out loud, in the memo, what you would want to know and cannot find out in the time. That sentence is read as maturity rather than as an excuse, provided the rest of the work is done.
The presentation
Assume fifteen minutes and expect to be interrupted in the first two. Lead with the recommendation, the target and the return, then the thesis in three sentences, then the evidence. Do not narrate the company history.
When challenged, distinguish between a question and a correction. If they give you a fact you did not have, take it and say what it changes: an analyst who updates cleanly is showing exactly the behaviour a portfolio manager needs. If they are testing your conviction on something you have actually worked through, hold the position and show the work. Folding under the first push is read as having no view, and arguing past a genuine correction is read as worse.
After the case
Send the model with the memo unless asked not to, keep the file names professional, and do not send a revised version the next day. Reopening it signals that you do not know when work is finished, which is itself a judgement being assessed.
Whatever the outcome, keep the case. It is a completed piece of original research, and the next fund will ask you to talk about your process.
Frequently asked questions
What is a hedge fund case study?
A research exercise on a company, given either in advance or in the room, ending in a recommendation with a price target. Formats run from a three-day take-home to a three-hour on-site build, and all of them are marking judgement rather than the completeness of the spreadsheet.
How complex should the model be in a case study?
Simple enough that someone else can open it, understand it in five minutes and disagree with a specific assumption. Revenue drivers, margin, capital intensity and a valuation, with three sensitivities. Forty tabs signal that you could not decide what mattered.
What if I am given a company I know nothing about?
That is usually deliberate. Work in a fixed order: what it sells and to whom, how it gets paid, the accounts against peers, the balance sheet, the competition, what the market expects and where that could be wrong. State in the memo what you would want to know and could not find in the time.
Should I change my view if the interviewer pushes back?
If they give you a fact you did not have, yes, and say explicitly what it changes. If they are testing conviction on something you worked through, hold it and show the work. Folding at the first push reads as having no view; arguing past a genuine correction reads as worse.
Now try it
Market Sizing Drill
Nobody in the room knows the answer. Build the chain of factors under a clock, then put it next to a reference chain and find the term that carried your error.
Related guides
Want this applied to your recruiting?
Reading is the easy part. For practitioner feedback tailored to your situation, work 1:1 with Suro.