I have a case interview with a Top 3 firm in two days' time. Thank you for the excellent materials that you've put up on this site.
I have questions relating to the interview. I was told during the case interview workshop organized by the firm that their interview would be much more structured, i.e. they will ask a broad question initially, but then guide you along the way.
My friends who have attended the interview confirmed this point. An example is as follows:
1. The interviewer would first ask, "Why is profitability in Company X falling? What would be your approach for this question?"
2. Regardless of how Question 1 was answered, the interviewer would then ask the candidate to estimate the market size of the industry.
3. Then, the candidate would be asked to do some quantitative analysis on a particular area, e.g. the effect of raising prices on sales (data would be provided).
4. After that, the candidate would be asked to list down some broad suggestions as to the ways to increase revenue (pricing, marketing, geographical expansion, etc.).
It feels as if the interviewer actually has a checklist in front of him. I guess this is what you would call an "interviewer-led" case.
I have two questions here:
1. Do I need to adopt a different approach in such interviews (c/f to your examples in the LOMS program)?
I would probably not have too many chances to ask questions and isolate the issues. How and when should I synthesize? Only at the end of the case?
2. Questions 1-5 seem quite detached. Can I ask for some time (about 30 seconds) before answering each question? Or should I only ask for time at the very beginning?
In terms of your questions, yes, you need to adopt a slightly different approach, as you'll be doing the case one component at a time.
Consider each question as its own mini-case. So the first question is generally a problem structuring... framework or custom issue tree question. They want to know how you would set it up, then promptly do NOT ask you to solve it in the way you intend. It is an abrupt interruption to the case.
So this is most similar to the first five minutes of the cases in Look Over My Shoulder®... and then they abruptly stop you and say, "Okay, great answer"... and then they jump to a totally different part of the case.
(Many people find this format of case is easier, because you're doing the case in small "bite-size" chunks. Others who are not aware of this particular difference in dynamic can get distracted by the stop-and-go format of the case and the jumping from one part of the case to another.)
You have 5 - 8 minutes per "question", so it is useful to make sure your answer is concise (but it does not need to feel rushed). So don't spend eight minutes asking clarifying questions just for Question 1, thinking you are getting everything clear to answer the whole case.
Synthesis would only occur at the end for the whole case, and generally they would ask you, "So what should the client do?" (Read: synthesize the whole case.) It is useful to do a mini-synthesis (2 - 3 sentences at the end of each question), if it makes sense to do so.
In terms of asking for some time, I suppose do whatever you need to do to answer the question correctly. So if you need some time, ask for it or just take it. Thirty seconds feels like forever though.
Usually if I need 30 seconds, I'll say, "Oh, that's so very interesting..." then I'll ask a clarifying question or two (a question where I don't really care about the answer... just to get myself some time to think)... or I will restatethe question out loud or paraphrase it... (yet another way to get some more time, and also to hear the question again in my own voice... sometimes makes it easier to think)... And then after doing all of that which takes 30 seconds or so, I figure something will pop into my head and I'll proceed.
If you'd rather just take 30 seconds of silence, and that's what it takes for you to get the answer right, then by all means take the 30 seconds.
Some candidates find this style of interview much easier than the candidate-led cases. It tests the same skills, but because the format is a bit different, if you're not expecting it, it can be a bit disorienting.
But once you get used to the style or prep yourself to expect it, it's actually the same process as a candidate-led case, but done in small semi-independent pieces.
The only other thing to keep in mind is that in subsequent steps in the case, they will use their framework or hypothesis, not yours. So maybe they'll say, "I have a hypothesis that X is true. What data do you need to disprove this hypothesis?"
So it's the same skill as coming up with a hypothesis in a candidate-led case and then creating an issue tree to prove, only this time it's someone else's hypothesis.
So the questions asked generally fall into the following categories:
1) Given a particular problem, what framework or custom issue tree would you use (e.g., what are the branches of the issue tree you would use)?
2) A basic algebra problem.. so something along the lines of, "If a company was to cut prices by 10%, how much more units would they have to sell to maintain the same level of profits?"
Basically, is a solve for X type problem. (It's usually harder than this.) This is just to see if you can do math at the level they are expecting.
3) Given this situation, what are the range of possible hypotheses that could be going on in this situation? (trying to see if you have either business acumen or common sense or intuition to have a sense of what the possible issues are)
This type of question is usually phrased as, "X has happened to the company or industry, what are the possible causes for X?"
Example: "The XYZ market has shrunk by 15%, what are the possible causes for this?"
Or: "What different ways can XYZ Company attempt to grow market share?"
Now the wrong way to answer this question is to start listing a bunch of reasons in a random order.
The better way to answer this question is to come up with categories of possible answers, list the categories, and then within each category list your specific ideas.
This is a more structured way to answer this question.
(By the way, it is worth re-reading the last three paragraphs several times... yes, it is that important).
4) Here's a graph, tell me what you think (an analysis and mini-synthesis is good here).
5) Synthesize what you've learned in the case and what the client should do.
So overall, the case is much more stop/go than the candidate-led cases in Look Over My Shoulder®, but if you look carefully at what is asked, all the specific steps are the same as in a candidate-led case (and thus the skills tested are similar), but the part that you need to be prepared for is you are not controlling the tempo and direction of the case.
You are being told which parts of the case to do when -- and often it is not in the order that you would do it, even though the actual questions are ones you would naturally tend to answer yourself in a candidate-led case.
Or phrased differently, rather than being in the driver's seat during the case, you are now now in the passenger seat -- but despite where you sit, you are still using the same case interviewing skills, but using them in stop-and-go /bits-and-pieces.
For more case interview preparation resources, review my free video series on Case Interview Secrets .