The recommendation goes first
Not in the conclusion, where you have been trained to put it. In the opening paragraph, stated as a decision somebody could act on this quarter.
This feels wrong to anybody who learned to write essays, because it gives away the ending. That is exactly the intention. A case is a memo to a person with limited time, and the structure that scores is the structure that would actually be used in the room: here is what I recommend, here is what it costs, here is why the alternatives are worse. Everything after the first paragraph exists to support a position the reader already knows.
Write for somebody who will act on it
The framing that fixes most case writing is imagining a real reader with a real decision and about four minutes. Not a professor assessing your understanding, but somebody who has to choose on Thursday and would like to know what you think.
That reader wants the answer, the cost, the main objection and what happens next. They do not want a summary of the situation they are already in, which is why the background section is usually the longest part of a weak submission and the shortest part of a strong one. If you find yourself explaining the case back to the case, cut it and spend the words on implementation instead.
Kill the alternatives explicitly
A recommendation with no rejected options reads as the only thing you thought of. Name two or three genuine alternatives and dispatch each in a sentence or two, with a reason attached that comes from the case rather than from general principle.
This is also where most of the marks for analysis actually sit. Saying that a joint venture was rejected because the case states the partner's distribution network does not reach the target segment demonstrates that you read the material and used it. Saying a joint venture was rejected because it introduces complexity demonstrates nothing at all, and could have been written without opening the case.
- Recommendation, in the first paragraph, specific enough to act on
- Two or three alternatives, each rejected with a reason drawn from the case
- The numbers the case supplied, actually run rather than referenced
- A framework only where it earns its place, never as decoration up front
- Implementation: who does what, in what order, starting when
- Risks, with what you would watch to know if it is going wrong
Use the numbers you were given
Cases include financial exhibits for a reason, and a strategic argument that never touches them reads as opinion. You do not need sophisticated modelling. You need to show that the recommendation survives contact with the arithmetic the case put in front of you.
Contribution margin, payback period, a simple break-even, a sensitivity on the one assumption everything rests on. Any of those, done plainly and interpreted in a sentence, separates a submission from the majority. Where the case gives you an exhibit you did not use, ask why it was included, because case writers do not add tables for atmosphere.
Implementation and risk are where marks go unclaimed
Almost everybody stops at the recommendation. The sections describing how it would actually happen, and what would tell you it was failing, are frequently worth as much as the analysis and are frequently absent entirely.
They also do not take long. Implementation is a short sequence with owners and rough timing. Risk is two or three things that could go wrong, each with a leading indicator you would watch for. Writing those two sections properly is the cheapest available improvement to a case submission, and it is the part that most resembles what the degree is supposedly preparing you to do. If simulations rather than cases are the difficulty this term, the decision log is the thing that fixes those.
Questions people actually ask.
Should the recommendation really go first?
Yes, and it is the change that most reliably moves a case grade. A case is a memo rather than an essay, and the structure that scores mirrors how the decision would be presented in practice: recommendation, cost, why the alternatives are worse. Withholding it until the conclusion reads as indecision even when the analysis is strong.
How many alternatives should I reject?
Two or three, each dismissed in a sentence or two with a reason taken from the case itself rather than from general business principle. That specificity is where much of the analysis credit sits. A rejection that could have been written without reading the case demonstrates nothing and markers recognise it instantly.
Do I need complex financial modelling?
No, but you do need to touch the numbers you were given. A contribution margin, a break-even, a payback period or a sensitivity on the key assumption, each interpreted in a sentence, is enough. Case writers include exhibits deliberately, so an unused table is usually a signal you have missed something the case wanted examined.
What gets left out most often?
Implementation and risk. Most submissions stop once the recommendation is argued, yet those closing sections often carry comparable weight and take fifteen minutes. Implementation is a sequence with owners and timing; risk is a few failure modes each paired with the indicator you would watch. It is the cheapest improvement available.
