Question:
In a second round case interview next week, I’ll be giving a presentation of a case to a panel. Does this type of interview fit into the candidate-driven type, or is it the pre-structured type? Do you have any experience with or suggestions for second round case interview prep? Thanks.
My Reply:
This is very difficult to say: 1) I don’t have personal experience with being on either side of this type of interview, and 2) the interview will probably vary by firm.
But, let me pass along my best guess.
First, I think realistically you have to be prepared for both styles of cases — candidate-led and interviewer-led (with one variant being they give you a written document with some pre-structured data).
Second, because it’s a group case interview, you may want to be mentally prepared that you will be asked to use a flip chart or a white board.
So that pad of paper you bring to the interview (and you should bring one… I tend to prefer graph paper or all white paper, so the lines on regular ruled paper don’t obscure my drawings), may not actually be used. So if this happens, don’t be surprised.
Third, one thing you definitely want to be prepared for is potentially being interrupted a lot.
Some firms have a panel sit in to silently observe and get multiple perspectives on your performance, others will have all the members of the panel ask you questions.
It’s the latter that you need to worry about.
In particular, you have take really good notes as to 1) what your overall structure for your case is, 2) which parts you’ve already covered, and 3) which parts are left.
The way I like taking notes in a case is using an issue tree diagram.
If you take notes like you would in a lecture class, you capture all of the information about the case but none of the structure of the case— and it’s the structure of the case that you must remember if you’re being peppered by questions from three or four interviewers simultaneously.
By structure, I mean lets say there are 4 major parts of a case and its not entirely clear which part is the most important part, where the “answer” to the case lies. So you’re forced to analyze each part of the case, until you’ve determined it’s the right place to focus or you’ve eliminated it as a possibility.
So lets say you start on Section 1 out of 4. In the middle of Section 1, you realize that to analyze Section 1 correctly, you have to break down the problem into three separate parts. Lets call these 1A, 1B, and 1C.
In the middle of 1A, you end up doing an estimation question. In the middle of 1B, you get interrupted by a member of the panel who took say a particularly aggressive/skeptical tone (trust me, clients do this all the time).
Now you’re 18 minutes into the case. At this point, if you’re taking notes like you would in a lecture, there’s a really high possibility that you’ve forgotten Sections 2, 3 and 4.
But if you take notes using an issue tree approach, all you have to do is glance at the issue tree and you know exactly what’s been analyzed and more important, what is left.
In my Look Over My Shoulder® program, I have several sample issue tree diagrams. (See Cases 2, 4 and 5 — the final candidate in each.)
You’ll notice the issue tree diagrams are hand-drawn and if you follow along in the corresponding audio recording or transcript, hopefully you can see at which stages I drew each part of the issue tree.
They key thing to pay attention to is notice when I put a little X mark next to a particular branch of an issue tree — indicating that this branch is a dead end.
This is very important.
One of the bad habits / mistakes that 85%+ of candidates I’ve seen make is they start down a branch but never totally eliminate the branch. Most of a case interview is really about using a process of elimination to figure out what the problem is not.
Let me re-state that slightly differently, as it’s a bit counter-intuitive.
1) 80% of a case study interview is not about trying to solve a “client’s” problem, it’s about trying to isolate the client’s underlying problem (not just the symptoms).
If you isolate the underlying problem, often the solution is actually quite obvious.
2) The way you systematically isolate a problem in a case is not to guess at a solution, but to use a process of elimination to figure out what the problem is not.
This is a subtle, but important point.
If you listen to the recordings of the candidates in my Look Over My Shoulder® program who did not perform very well, you’ll see they did the exact opposite of what I suggest above.
In the middle of the case, rather than continuing to use a process of elimination to figure out what the problem is not… they start to guess what the problem is.
Examples:
1) Maybe the client could re-design the product so it could sell better… (proposing a solution before the problem has been isolated and defined.. basically assuming the problem is a design problem, and therefore incorrectly concluding that a re-design will fix the problem).
2) Maybe the client could partner with a Fortune 500 company with more resources and provide a joint offering of some type (again, this is an example of a candidate proposing a potential solution even though the actual problem has not yet been defined or isolated)
Remember this:
Never, ever, ever propose a solution to case until after you’ve isolated, and defined the problem. Then and only then do you propose a solution.
And if you do the isolation part exceedingly well, usually the “answer” is extremely obvious — stating the answer/recommendation is pretty much a formality, as both you and the interviewer know the “answer”.
If you ever do a case where the answer is not obvious, it basically means that you either:
a) set up an incorrect issue tree, or
b) you didn’t analyze a particular branch of the issue tree enough to definitively eliminate it as a possibility, or
c) you glossed over a key piece of information the interviewer gave you during one branch of the issue tree (and because you glossed over it, you didn’t analyze it further and failed to conclude this branch was a dead end… and you wasted all of your time analyzing a branch that should not have been analyzed any further)
You’ll hear and see examples of all three very common errors in the 13 interview recordings and transcripts in the Look Over My Shoulder® Program. For further materials, see my Case Interview Guide.
If you already have this program, you want to go through the recordings and transcripts very carefully. If you rush through the process, there’s a temptation to say that, “Oh, I was already familiar with that mistake. It’s not new information.”
Here’s the deal.
Most of the people who made those errors on the recordings also knew (intellectually) what they were supposed to do and what errors they were supposed to avoid. Most had studied all my free videos, in some cases multiple times.
The problem is: knowing what to do and actually doing it under pressure are two entirely different things.
So the right way to go through the Look Over My Shoulder® materials is to do so actively, not passively. Don’t sit there and listen passively or skim the transcripts passively.
The key is to put yourself in either the candidate’s shoes or in my shoes as the interviewer. Before you hear the little jingle that indicates that I’m going to provide my sidebar commentary on what the candidate did wrong, hit the “pause” button and try to anticipate what I’m about to say.
Or hit the “pause” button and re-state what the candidate should have said, instead of just passively listening to me do it.
In short, the Look Over My Shoulder® program is designed to be a participatory program, not an observational or lecture-type program.
Both of these “active” behaviors force your brain to take what you know and make it a habit… in essence taking knowledge into a skill.
If you’re not hitting the “pause” button on your MP3 player every five minutes, you are using the program in a sub-optimal way.You really need to use it actively to get the most value out of it so you can “ace the case”.
For additional materials, review my free video series on Case Interview Secrets and practice case interviews.