The case interview format required to get a management consulting job at a major firm can be daunting and brutal. I was lucky enough to successfully navigate Boston Consulting Group’s interview process after a fair amount of practice. My success involved a lot of practice on my part, but also a lot of help from some incredible people.
As I’ve been helping others get ready for their own case interviews, I’ve noticed a few tips that can help nearly all of the candidates I’ve talked to. I thought I would make it easy for them and anyone else to enjoy the knowledge I received from many better case-interviewers than I.
Here are what I consider to be the top 5 quick tips to improve your case interview performance. They won’t save a bad case, but they will greatly improve your communication, thinking process, and flow.
“Signposting” most typically refers to the act of numbering things, but includes other things that candidates can say to help an interviewer understand how the candidate is organizing his or her thoughts. Signposting helps a ton (1) when explaining the initial framework, and (2) when discussing insights from exhibits. These benefits are amplified when having an interview over the phone.
Signposting the Initial Framework
For the first point (explaining the framework,) the candidate should say upfront how many items or “buckets” that he or she wants to analyze. For example:
“there are two major things I want to look into to find the problem with profitability.”
I often suggest that candidates write numbers on each of their initial “buckets” to help remind them to say the number with it. Saying the numbers out loud is especially helpful on the phone when the interviewer can’t see the candidate’s paper.
Most initial frameworks will be logic trees with multiple levels. There are two major ways to explain through the different levels.
1. Level by level
This is the way I prefer. This involves giving an overview of everything at a certcain equivalent level before going deeper into any of them. It’s a bit like giving an agenda or preview for each level of analysis. For example, I might say:
“There are two things I want to investigate to identify why profitability is decreasing. First, I want to look for decreases in revenue, and second I want to look for increased costs. Within the first point, revenue, I want to look into three different aspects: first, the price of our products; second, the quantity of products sold; and third, our product mix.”
2. Go Deep, Then Rise
Although I’m not personally a fan of this process, I know that some interviewers prefer it. This means going from the first bucket all the way to the bottom of the logic tree before coming back up. For example, I might say
“There are two things I want to investigate to identify why profitability is decreasing. First, I want to look for decreases in revenue. Within revenue, I want to look for three things. The first thing to look for will be changes in the price of our products. [after completing all the revenue stuff] I now want to look into the second major thing, costs.”
Whichever method you prefer to use for your explanation, numbering the items and helping the interviewer keep track of the part of the analysis that you’re doing helps you be a clearer communicator, and can force you to be a clearer thinker as well.
This can be significantly more difficult for most candidates, but candidates who can do this are much more impressive. When the interviewer gives you a chart, graph, or something like it, you will usually take a few seconds to digest the information. After digesting, you identify the significant aspects of the exhibit to the interviewer. Whenever a candidate numbers the insights, gives an overview of what they are, and then explains them in a bit more depth, I’m always more impressed and I also better understand the candidate’s thinking.
Many times a candidate will say something like “I notice three things from this chart…” and then notice a fourth thing as they’re describing the first three. It’s not a bad thing to mention at the end of the third point that you just noticed a fourth significant point, especially if you’re clear that it’s a fourth insight. Don’t forget to number the additional point or you might sound like you’re giving two unrelated ideas as a single point.
2. The Synthesis Model
Frequently throughout a case the candidate should tie discovered information into the original question. There are good and better ways to do this.
The model is fairly simple, it involves saying three things in this order:
1. The answer/solution/implication
2. The supporting evidence for the answer
3. The next steps
For example, I might say
“The company is losing profitability due to cannibalization of its most profitable product.
We know this for three reasons: (1) 70% of the customers of the main product switched to the cheaper offering when it became available, (2) offering a cheaper product only increased our overall product sold by 5%, and (3) the most profitable product’s profitability itself decreased when we started producing fewer items due to the lower demand.
I think we should do two things next. (1) investigate what customers really want from our product to identify why the cheaper offering was adequate for more customers than we anticipated and (2) investigate the effects of a greater investment in recruiting new customers.” (Notice the signposting.)
When to Use it
The synthesis model should be used at two major times in a case: (1) at the end of every branch of analysis and (2) at the end of the entire case. In candidate led interviews (such as with BCG and Bain), the candidate should be using the original framework throughout the case as a guide. Whenever a candidate gets to the end of a branch of analysis (such as the revenue branch) the candidate should summarize what the analysis revealed about the problem, a brief recap of the supporting evidence, and the next step (which will be to look into the next branch of analysis, such as costs).
At the end of the case, the synthesis should give the answer to the original question, recap the most significant supporting evidence, and give next steps that cover one or both of two things: (1) analysis on things you’re still unsure about, and (2) possible future engagements with the client. Many cases will require you to make a decision when there’s enough of a balance of pros and cons to decide either way. Using the “next steps” in the final synthesis lets you identify the additional information you would like to have, without giving a weak decision on the main point.
For the second item, whenever you can identify additional work we could do for the client, you start to look a bit more like a coworker than a candidate, and that’s good. Being able to talk about how to sell the next piece of business is a nice icing on the cake that can make the difference between an adequate interview and a great one.
3. Look For, Not At
When explaining their framework, it’s common for candidates to say things like “I want to look at our product” or “I want to look at fixed costs.” It’s a start, but looking at something won’t tell you the answer. You need to know what you’re looking for as your eyes rest on the thing you asked for.
For example, a candidate trying to identify a profitability problem my say
“I want to look at revenues and costs.”
It sounds logical, but I’m not yet sure that the candidate will know what to do with the revenue and cost information. It sounds better, and forces you to think more clearly when you say something like
“I will look for decreases in revenues, or increases in costs.”
Here, I already know the criterion in the candidate’s head, and the candidate knows it too. Although profitability issues are a bit of an obvious example, this principle becomes even more important with non-financial things. Often times a candidate will say
“I would like to look at the product.”
To which I might thump a product in front of the candidate and say “here it is, what’s the answer?” At that moment, it will be clear that you need to look for certain things. Maybe you’re looking for defects in the product, or ways that the customer uses the product differently than what’s anticipated, or for increasing expenses in certain components of the product, etc. Looking at the product does nothing.
This also helps the candidate direct the case. It’s easier (though not guaranteed) for a candidate to get away with just identifying the “look at’s” with some McKinsey first-round cases, but to get the final offer, or to navigate the candidate-led interviews at other firms, the candidate needs to know exactly what to look for. Keeping this mindset has helped many candidates avoid the pause after explaining the framework as the candidate is forced to finally think about what we’ll actually start doing. When you know what you’re looking for, you know what questions to ask and how to ask them.
Using this slight change in language forces candidates to think about why they want the information and allows them to hone in on the important aspects of the dizzying swirls of data I might unload on the candidate.
4. The Calculations Method
Math will be a part of nearly every case interview. Besides just getting the numbers right, you need to get the process right. Here’s the best way I’ve seen to do the math:
1. Identify why you’re going to do some sort of calculation
2. Identify the equation (with words and variables) and write it at the top of the section you’re using
3. Do the calculations (out loud if you can)
4. Synthesize the meaning/implication of the number
What you should not do is simply say “give me a minute to run these numbers” after receiving a chart or numerical data, and then run all the calculations before talking to the interviewer again. You can’t just be a good calculator, you have to be a good thinker and communicator as well.
When you identify why you’re doing a calculation, you verify to yourself that you’re not doing unnecessary calculations and you know exactly the result you want. It’s best if you can give the implications of the result ahead of time too, such as
“If the the profit is greater than $3MM, then we’ll know that the promotion was better than no promotion. If it’s less, then we’ll know to recommend not doing a promotion.”
When you identify the equation, it ensures that you have all the pieces that you need, and that you understand what’s going on. It also gives the interviewer a chance to correct you up front. For some reason, interviewers are much more forgiving and willing to blame themselves for being unclear if you give them a slightly wrong equation than if you give them a wrong number. If you’re going to have a numerical error, it better not be from a wrong equation.
Saying the calculations as you do them is really difficult, but that’s what makes it impressive. If you have the equation written out, then often times you’ll be able to talk through each step of the equation and know the result through mental math pretty quickly. The math is usually designed to be relatively simple and fit together well, and if you can think through the next operation in your head while your mouth is explaining the operation you just did, then you avoid the awkward silence and look like you can do math at a lightning pace.
At the end, you must ALWAYS synthesize what the number means. If you can’t identify what a number means for the case, then you shouldn’t have done the calculation in the first place.
5. The Number is Not the Answer
One common mistake I see is that when a candidate finds that a certain product line is decreasing in revenues, they think they’ve found the problem with the profit. However, in most cases the number only tells you where to look for the story, but it doesn’t tell you the story itself.
For example, if a grocery store has declining profitability in its frozen foods, it’s not enough to identify that declining frozen pizza sales are causing the profit loss. The real story is that customers have decided that frozen pizza is destroying the planet and have switched to the less profitable but planet friendly frozen tacos. (The even MORE real story is what customers think the frozen pizza is doing to the environment that the company can change. Does the frozen pizza factory dump toxic waste into rivers? Do pizzas increase global warming? etc.)
When you know the real story, you can hopefully come up with better solutions as well. If you only know that frozen pizza sales are declining, you’re liable to make the typical, generic proposals such as advertising, adjusting prices, improving the quality in some vague way, etc. If you know that customers care about the planet, then you can have more specific suggestions about making the pizza planet friendly.
Hopefully you find these helpful. I’ll probably come out with another 5 tips later on if people find these tips useful. Let me know if they help, or if you disagree, either by emailing me or writing in the comments.
Keep seeking truth.
You may also be interested in: