How do I handle a case study?

The first ten minutes decide the other fifty

The most common mistake is to start at page one. The material is deliberately built to contain more than you need — the in-tray, the sibling exercise, is described by the same source as containing irrelevant information on purpose. So read the question first, then every heading, caption and axis label in the appendices, and only then the running text. Do it the other way round and you will have read everything after twenty minutes and decided nothing.

Then spend five minutes not calculating. Write down the question in one sentence, the two or three quantities the answer depends on, and the order in which you will look at them. That outline is not scaffolding you throw away; it is part of what is being observed. The association's presentation exercise explicitly allows for a preparatory phase in which candidates analyse and structure the information material, and is meant to bring organisational-analytical skills to light.

Plan backwards from the clock. The last five minutes belong to the summary, not to the final calculation. A well-summarised recommendation with one loose end beats a complete calculation with no conclusion, because only the recommendation shows how you decide.

Put the recommendation first

Write a first sentence that can stand alone: what you would do, from when, and what you would look at in three months to know whether it worked. Everything after that is justification. This is not presentation gloss — it forces you to take a position early, and the position is the thing being observed.

Directly under it, list your assumptions, each with a number attached and each marked as replaceable: "I am assuming churn of around five percent a year; at ten percent the recommendation flips to option B." That builds the clarifying question nobody will answer for you into your own paper. It is the clean answer to the no-feedback rule.

Include the option you rejected — one paragraph is enough: what spoke for it, and why you still did not choose it. The definition of the exercise names decisions and alternative solutions in the same breath. Show one option only and you leave half the observation sheet blank.

Four types of case, and how the advert tells you which one is coming

The numbers case belongs to analyst and controlling roles. In analyst adverts on our site, analytical ability is named in 17.3 percent and key figures in 10.1 percent; in controlling and finance adverts, 13.4 and 6.0 percent — several times the index averages of 3.0 and 1.4 percent. Expect incomplete data with a gap you are meant to name, and check any figure by hand before it reaches your closing sentence.

The market or strategy case belongs to strategy and business development roles: analytical ability in 14.9 percent, stakeholders in 3.9 percent, strong communication in 7.8 percent. These cases almost never close arithmetically. What is wanted is a structure in which the answer could be found, plus a reasoned choice made inside it.

The stakeholder case belongs to product and project management. Those adverts name stakeholders in 3.9 percent — three and a half times the index average of 1.1 percent — and conceptual work in 4.1 percent. Such cases do not give you too little data; they give you too many interests. Two departments want incompatible things, and your job is to say so rather than average the conflict away.

The design case belongs to architecture and lead roles in IT: conceptual work appears in 6.0 percent of those adverts, three times the average, and problem-solving in 2.8 percent. What is expected is a sketch with variants and the cost of each variant, not the one elegant solution.

Consulting breaks the pattern. Its adverts name analytical ability less often than average (2.6 against 3.0 percent) but presenting more than twice as often (3.1 against 1.4 percent) and strong communication in 6.2 percent. If you prepare for a Swiss consulting case by drilling arithmetic alone, you are preparing the wrong half.

What to do when a number is missing

Missing numbers are usually not an oversight; they are the point. The workable response has three steps: name the missing quantity, fill it with a defensible order of magnitude, and show where that order of magnitude comes from. "I do not know the number of branches; I am working with twenty, because the case mentions five regions with three to five sites each."

Avoid two reflexes. The first is false precision: a figure with two decimal places resting on a guess. The second is refusal — "I cannot say anything without more data." Both end the conversation the exercise is actually looking for, which is the visible route from an uncertain basis to a decision you are still willing to defend.

Whenever you make an assumption, make it replaceable in the same line: note the value at which your recommendation would flip. That gives you, in one sentence, the sensitivity analysis there is otherwise no time for in forty-five minutes.

Which language the case arrives in

Across the case-heavy roles, 49.1 percent of adverts that state a language requirement ask for English, against 41.1 percent across the index; at professional level it is 43.5 against 36.1 percent. The advert text itself is more often written in English in this group too: 23.4 against 19.8 percent.

The spread inside the group matters more than the gap to the average. Strategy and business development adverts that state a language ask for English in 90.4 percent of cases and for German in only 42.6 percent; data and business intelligence roles ask for English in 85.4 percent. Consulting runs the other way: 76.7 percent German, 38.2 percent English, 24.3 percent French. A Swiss consulting case is more likely to arrive in German than a strategy case is.

So ask, in one line when the invitation lands, which language the material, the working time and the presentation will be in — they can be three different ones. This matters especially if you are job-hunting in Switzerland in English: a case handed to you in German with a discussion in English is a normal Swiss arrangement, not a trap, and it is far easier to handle when you knew about it a week earlier.

Then rehearse the twenty terms of your own field in that language — contribution margin, churn, utilisation, provisions. It is almost always the domain vocabulary rather than the grammar that stalls a presentation.

The take-home case, and where an exercise becomes work

When the case leaves the building with you, the situation changes. Under Art. 320 para. 2 of the Code of Obligations, an employment contract is deemed to have been concluded where the employer accepts the performance of work over a certain period in their service which in the circumstances could reasonably be expected only in exchange for salary. The article does not turn every take-home task into an employment relationship — but it names the boundary, and the circumstances are what decide.

Three questions are entirely reasonable to ask before you accept: how many hours is this meant to take, is the data anonymised or invented, and will the output be used internally? An invented starting position with a three to four hour frame is an exercise. A real, current problem with real data and no time limit is the point at which asking about compensation is appropriate — and how that question is received tells you something about the workplace.

If no limit is given, set your own and write it into the document: "time spent: four hours, as agreed." That protects you twice — from the evening work nobody actually asked for, and from being compared with someone who put in thirty hours and did not mention it.

Who owns what you produce

Under Art. 6 of the Copyright Act, the author is the natural person who created the work — not the company that set the task. And Art. 16 para. 3 states that the assignment of the ownership of a copy of a work does not include the right to exploit the copyright, even in the case of the original. Handing over the file is not the same as granting permission to use it.

The honest limit belongs with the rule: protection only exists insofar as what you produced is a work under Art. 2, meaning it has individual character. A filled-in company template will rarely qualify; an analysis you structured and argued yourself is far more likely to. The point of these articles is not to threaten anyone with the law, but to know that the apparently obvious principle — whoever sets the task may use the result — is not actually the legal position.

In practice three small habits are enough: your name and the date on every page, keep your own copy, and add one line on submission asking what happens to the document after the decision. That is not confrontation; it is the same care the role itself would expect from you.

Where these processes sit in Switzerland

Case-heavy roles are geographically more concentrated than the rest of the market — but only in two places: 22.0 percent of them are in Zurich against 16.4 percent of all adverts, and 8.3 percent in Geneva against 5.6 percent. Bern, at 7.3 against 7.7 percent, and Lausanne, at 2.3 against 2.8 percent, are slightly below their usual share; Basel is marginally above at 6.7 against 6.2 percent. If you are searching in Zurich or Geneva you will meet this format more often than average — elsewhere, rather less often than the method's reputation suggests.

Among the employers with the most such adverts in our index are advisory and audit firms such as EY, Deloitte and Eraneos, recruiters such as Michael Page, and insurers including Helvetia, die Mobiliar, Swiss Life and Visana. A tenth of these adverts come from a single intermediary platform alone — and where a recruiter is involved, the case is often the end client's rather than the recruiter's, which is worth asking about before you prepare.

17.6 percent of these adverts are one of several word-for-word identical postings from the same employer. In that constellation several people work the same case, often on the same day — one more reason to take a clear position rather than aim for the low-risk middle everyone else will also find.

The day after

Write down the same evening which question knocked you off balance and which number you did not have ready. Within an industry, cases repeat heavily in structure even when the content changes — the gap you noticed today will meet you again in the next process.

And if the answer is no: the written evaluation that belongs to this kind of procedure exists. Asking for it costs nothing, and it is the only way to take something out of a case that goes beyond your own impression of how it went.