I’ve been giving more practice cases and a few more things popped up that most people can do to improve their performance in case interviews. Here are 5 more quick tips for case interviews (see the first 5 tips here).
Keep A Case Log
Your performance in cases will improve much more quickly if you track your performance. There are a few things you should keep track of.
First, record some basic info. This should include a name for the case, the person who gave you the case, and the date you took the case. It can also be helpful to record something about the premise or issue of the case, as well as your own conclusion for the case.
Second, record the feedback. You should note whether you got the correct answer, and possibly take note of any background information in the case that you didn’t obtain through your questioning. You should also record feedback from the interviewer. I chose to break my feedback into a section for things I did well, and another section for things I should improve on.
Third, take some general measurements. It may be enough for you to just give yourself (or ask the interviewer) for an overall score out of 10, with an 8 being “pass first rounds” and a 9 being “you get the job.” You may decide that it would also be valuable to score different components of your case, such as the framework or your performance on the quant section. If you’re struggling with any specific aspect over many cases, consider scoring that section.
Finally, record some “Next Steps.” You should identify one to three skills that you want to focus on in the next case. You should also take note on how you perform on these skills on the next case. As an interviewer, I’ll usually ask the interviewee if there’s anything specific that he or she is trying to work on for our case, and if I have a case that fits his or her needs I can make adjustments. Try to get the most out of your interviewer by communicating the skills you’re trying to improve.
You may decide that you don’t need to track all of the things that I listed, but you should certainly be tracking something.
Take One Question at a Time
Many case prompts have two or more questions in them. This is especially common in profitability issues. The two questions may be something like, (1) why is profitability declining? and (2) how do we improve our profitability? Or, it may be something like (1) how “good” is this company? (2) Should our private equity client purchase the company?
When encountering multiple questions from the start, my strategy was to take them one at a time. Instead of trying to address all of the issues in one grand framework, I clearly told the interviewer that I would only approach one aspect of the prompt, and then address the second aspect with a separate framework after we finish the first part.
In most cases, the crux of the issue is isolating the problem in an organized way. The interviewer doesn’t want you to ask a few questions about a problem then start spouting off specific solutions. Most of your job will be breaking down the vague issue into something tangible and identifying the specific trouble-maker. You often won’t get to the solution stage until nearly the end of the case, and sometimes this stage involves little more than a short brainstorm.
If you do delay addressing a part of the prompt, make sure you clearly communicate to the interviewer that you’re intentionally not addressing part of the prompt at first. Without communicating, the interview might think that you missed something important.
Don’t Assume Anything
Imagine a scenario: you’re a consultant talking to a client. You assume something about the economy and its affect on your client. You base a recommendation on your assumption. That assumption turns out to be incorrect. How do you think the client will like you? How do you think your case manager will like you?
Scrub the word “assume” from your vocabulary in cases. There may be appropriate times to use the word “assume,” but there are so few times that you can probably just never use it at all.
If you think that you need to make an assumption, see if you can verify it. For example, lots of people will assume that since the economy sagged at time X, this must be hurting our business. Instead of assuming, you can sometimes directly ask “is the sagging economy hurting our business?” If that doesn’t work, you can identify the information you would need to find out if the economy really is the cause behind our hurting business. If you still don’t get anything to verify the assumption, then you probably don’t need to make the assumption to solve the case, or the solution will be highly tentative and your language should reflect that.
In consulting, your word is your product. Unfortunately, if you release a faulty or dangerous product, you don’t get much of an opportunity for a recall. That product will be out and causing all sorts of injury if you don’t get it right. So don’t assume anything in the case any more than you would assume that chemical X would be a great pill to solve problem Y.
That said, here is when you can make assumptions: when the client/interviewer tells you to. The interviewer doesn’t need to scrub the word “assume” from his or her vocabulary. But notice that if the interviewer says “assume our product is the same,” that you have just received actionable data for yourself. The interviewer’s word is law, so the interviewer’s “assumptions” are merely conditionally packaged facts.
Identify the First Place To Go
After you explain your framework, you should not pause and wait for the interviewer to direct you. I also don’t recommend asking the interviewer where to go from there. Even in an interviewer led case (like in McKinsey’s first round) you should identify where you would start if it was up to you. This is extra important in a candidate led case where most of the time you will actually start where you want to start.
There may be room to solicit a little input from the interviewer. Instead of asking what to do, I recommend using a phrase more along the lines of “unless you have objections, I would like to start here.” This nods toward the authority and greater insight of the interviewer while showing that you can take control of the case and do the analysis without too much prodding. In the consulting job you might not have someone looking over your shoulder to tell you where to analyze first, so demonstrate that you can handle that.
You will often need to compare numbers to other numbers. This is the only way to tell if a number is “good” or “bad.” When you do benchmark numbers, make sure that you always try to benchmark (1) against the competition and (2) against history.
Benchmarking against the competition may be in the form of comparing different promotions at different stores within a single company, or it may involve looking at the 3 biggest competitors. Whatever you’re comparing against, the basic principle is that you compare the number at issue to comparable numbers today.
Benchmarking against history basically means looking at how your number now compares with numbers of the same entity in the past. Make sure you use relevant comparisons. If you’re dealing with a company that changed its price 3 years ago, the relevant period may be to compare its performance 4 years ago before the price change. Try to get to where the only difference in history is the issue you’re trying to solve.
Always try to benchmark against at least these two things. You won’t always be able to do both, but unless the case logically prevents you from making a comparison, you should always try to compare against both.
These issues have been common mistakes recently, but they are fairly easily fixable. Some of these (like benchmarking or assuming) really can make or break your case performance. If you need to work on any of them, stick it in your case log and let your next interviewer know to watch for them during your next case.
Keep seeking truth.
You may also be interested in: