One of the nuanced skills that I’m noticing some people having difficulty with is the ability to identify the “key drivers” behind a particular business trend.

Mastering this skill is not only important in a case interview but is arguably far more important after the case interview — when you are doing real client work.

This skill is remarkably practical.

If you use a framework or issue tree to structure a problem, each branch of the issue tree or framework is a driver.

The key driver is the single most important driver out of the bunch.

This particular skill can be tested or presented in one of two different ways.

The first is a brainstorm type question… which initially seems like the interviewer wants you to list as many reasons as you can think of that explains why XYZ is happening.

Basically, this is another way of asking you to generate a bunch of hypotheses.

This style of inquiry is fairly common in McKinsey interviewer-led cases.

While in this context, they are asking you a little for a brainstorm (not just structure or an issue tree), there is still a way to do this in a more organized and structured way (while still demonstrating creativity or business judgment — being creative in a way that resembles real-life business).

So, they would ask you, “what are all the drivers that would impact this trend,” and then they might ask you, “Which one is the key driver, ” (a.k.a. the most important one impacting the largest portion of the trend you are seeing in sales, profits, market share, etc…).

If they don’t ask you for the key driver, there is a way to organize the list of drivers in a way that is more structured. If they do ask for the key driver, you need to think about which one is the most important.

This can be done either through “business judgment” or very thoughtful “common sense.”  The former is something you either have or don’t and is hard to develop in a short period of time.

One way to improve that skill somewhat is to read a lot of business magazines to get a feel for what decisions company CEOs have made, and whether or not they were the right decisions.

If they were the right decision, it means they focused on the right key driver and successfully executed a plan to improve that key driver.

If the decision was wrong, either they focused on the wrong driver, or they focused on the right driver but could not execute effectively in influencing that driver.

The second way this skill is assessed is in a case interview that’s candidate led.  They look at your issue tree structure and notice that, while they agree with the three key branches, you focused instinctively on the least likely one first… the total opposite of what they were looking for.

In other words, you focused on a “secondary” driver, not a primary one.

I am seeing emails from several people who have done every other aspect of the case well, except this specific one, and it cost them the interview.

I have included an email from a reader that asks about this specific issue. I suggest reading the email and then my comments that follow it.

***What do you think the main reasons are behind… ***


Thank you for all your emails and LOMS.

They are truly a great help and are helping me a lot preparing for the interviews coming this March.  While preparing, I have a few questions and if you had time to answer them, I would be very grateful.

My first question is on how to answer “what do you think the main reasons are behind..” questions. I have some friends who were asked questions such as…

1) Imagine that the number of [wolves] in your town fell suddenly, what do you think are the main reasons behind it?

2) The [bowling ball] industry is not doing so well these days.

What do you think are the main reasons behind it?

I imagine that this is the “making of a hypothesis,” but am unsure if there is an appropriate way to answer — if they are looking at:

1) Structure (how I answer), then I should categorize my thoughts.

2) How many items can I come up with, then I should list as many possible answers in each category.

3) How I prioritize my answer, then I should structure my items and list them in the order I think is important.

4) Drivers of the # of wolves, then I should list these.

I took a shot at #1, the number of wolves decreasing.

I believe that the environment the wolves are in and the food available are the main drivers behind the number of wolves, and therefore the main drivers behind the number of wolves falling in my town.

In terms of the change of environment, there are two major changes.

The first is drastically colder weather, where wolves could either have died due to the cold weather or left my town to the southern regions, where it is warmer.

The second is the probability of our town fumigating wolf homes since wolves tend to spread disease. In terms of availability of food, wolves mostly gain food from food dropped by people.

Due to the cold weather, people tend to eat less outside, which leads to less food being dropped for the wolves to eat, leading to a food shortage and reduction in pigeons.

Any feedback would be great.

Or I could answer it: I believe there are three main reasons behind the decline of wolves; change in weather, chance of fumigation, availability of food.

Is there a preference in structure? My answer is not MECE and should it be MECE?

My Reply:

In terms of what the interviewer is looking for, it depends on which firm is asking.

If it is McKinsey Round 1 (and increasingly I am believing subsequent rounds as well), they are looking for actual answers… not just a structure for how you would find the answer.

Other firms might ask the question in the context of a candidate-led interview and mean they want to see your structure and not just a brainstorm.

Given the ambiguity of the question and depending on who is asking, I would just ask them which one they are looking for.

I had one person who started doing the structure-only approach, and the interviewer (who I thought was good) said, “I’m sorry… I meant can you give me actual ideas…” and the candidate did.

Some candidates have just brainstormed and listed every reason they can think of and have passed the interview.

Personally, I would come up with the CATEGORIES of reasons, list those categories first, then the actual ideas next.

So, I would say the following:

“Do you want me to figure out a problem-solving structure to figure out why the wolf population has declined? Or, do you want me to just come up with as many reasonable hypotheses as I can think of?”

“Okay, you want the latter. Got it.”

“Well, in terms of what might cause all the wolf population to decline, I think there are three categories of reasons.

The first category is reasons that cause the BIRTH RATE to decline.

The second category is reasons that cause the DEATH RATE (or average lifespan) to increase.

The third category of reasons is the Net MIGRATION RATE — or the rate at which wolves leave the borders of the town.

Unless I am mistaken, these three categories should encompass all possible reasons that might cause the wolf population to decline.

Let me come up with a few specific ones in each category.

For the birth rate decline, perhaps there is some environmental factor causing males and females to not mate as often as they used to. Alternatively, perhaps they do mate, but the male “contribution” is not as fertile as it used to be. Or perhaps females are no longer able to carry the babies to term.

Any of these three could be caused by increase in pollution or toxins in the air, changes in quantity or quality of food supply, or changes in quantity or more likely quality of water supply.

In terms of the death rate increasing, there could be natural factors vs. man-made factors. For natural… list a bunch… for man-made… hunting has gone up, demolition of natural habitat for construction has increased, etc…”

Anyway, hopefully, you get the idea. I, personally, like grouping the ideas, rather than just listing a whole bunch of them — though I think for McKinsey Round 1 it seems like you can still pass without grouping in this way.

And, in answer to your question about being MECE, out of habit, I try to be MECE whenever possible. Again, for McKinsey Round 1, it may not be necessary. If you’re doing a candidate-led case, you want to be MECE or pretty close to it.

The main reason to be MECE is it ensures some level of completeness and gives your common sense and intuition some boundaries to work with.

When I think of why the wolf population has declined, that’s very amorphous to me. If I say, “Why has the birth rate declined?” — well, that’s easier for me to imagine the reasons.

So, I use the MECE approach to make it easier for me to come up with ideas and to not overlook one.