YT YouTube2Text

YouTube2Text → Transcripts → Case Interview Fundamentals: Lesson 24 [Quantitative Analyses: The 4-step approach]

Case Interview Fundamentals: Lesson 24 [Quantitative Analyses: The 4-step approach] — Transcript

by CraftingCases · 9,780 words · 1,483 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:00Hello everyone, what's up? This is Julio
0:02speaking as I said in the previous
0:04video. Welcome to the quantitative
0:06analysis module of case interview
0:08fundamentals. What you're going to learn
0:10in this video is what an analysis is
0:13from a case interview question
0:15perspective and from the consulting job
0:18perspective. You're going to learn what
0:20you're being tested on, what your
0:22mistakes and your hits tell your
0:24interviewer about your day-to-day skills
0:26as a consultant, and what it tells them
0:28about what they can expect from you as a
0:30consultant. You're going to learn a
0:32four-step approach to make sure that you
0:34touch on everything that the interviewer
0:37wants to see during a quantitative
0:40analysis question regardless of what
0:43question they give you, what industry,
0:45what type, how hard it is. And you're
0:47going to learn a technique to create
0:49structures. It pisses me off that so
0:53much content out there, case books,
0:54videos, YouTube videos, everything just
0:58gives you an analysis question and then
0:59shows you the answer as if it's simple.
1:02And the truth is to most of us it isn't.
1:05There are techniques and there are
1:07skills that one can learn and it is
1:10possible to learn, it isn't a God-given
1:13skill. And I'm going to show you
1:15what these techniques are so that you
1:17can create structures by yourself. So
1:20let's jump right into what is a
1:22quantitative analysis from a case
1:24interview perspective. So a quantitative
1:27analysis question is when an interviewer
1:29gives you data and asks you to calculate
1:32a number or to calculate a conclusion
1:35for example, such as which of these
1:38options is better, does this investment
1:40make sense, etc. Let me show you some
1:43examples of what analysis questions
1:45could be.
1:47Take a look at these. Calculate revenue
1:49synergies from introducing your
1:50competitors products into Starbucks
1:53stores after an acquisition.
1:55Calculate cost savings in R&D, research
1:58and development after implementing this
2:00new process. Or given this table of
2:03data, how much will it cost to improve
2:05the quality of this university's dorm
2:07rooms?
2:08Now, you can expect to have
2:11one or two of these in every case that
2:14you do. Every interviewer is going to
2:15test your quant skills some way, and
2:18this is
2:19usually one of them.
2:21You might say that these questions are
2:23similar to estimation questions because
2:26well, you got to calculate stuff. And
2:28the difference from estimations is that
2:31in estimation questions,
2:33they won't give you any data. Or they'll
2:35give you a little bit of data, but
2:37they'll ask you to estimate stuff and
2:39you'll have to come up with a lot of the
2:40assumptions.
2:41In or maybe they will even give you the
2:44assumptions, but only after you
2:46structure. Now, in quantitative analysis
2:49questions, they
2:51give you some data to begin with.
2:53Just a heads up, it's hardly ever all
2:56the data you need. So, almost every
2:58single time you need to ask for more
3:00data. But yeah, that's the main
3:01difference that you're going to see. And
3:04a difference that's not conclusive, but
3:06it's helpful and it works most of the
3:09times is that estimations are often
3:11market sizings and they often start with
3:14can you estimate a number for something.
3:17And in this case, they ask you to
3:19usually calculate a number for something
3:21and it's usually not a market size.
3:23Although it could be. Could be They
3:25could ask you to calculate a market size
3:26and give you a few numbers and you have
3:28to come up with others. And that would
3:31really be a gray area between estimation
3:33questions and quantitative analysis
3:34questions. And
3:36honestly, there isn't that much of a
3:38difference between them if you look at
3:41them from a pure analytical perspective.
3:45All right?
3:46Now,
3:47I want to address a common misconception
3:50here. Many, many, many people call these
3:52questions math questions.
3:55And I want to break this myth down.
3:59These are not math questions, and here's
4:01why. So, a math question implies that
4:04the most important part of the question
4:07is the math, the actual calculation of
4:09the numerical answer and getting to the
4:12right number, like you would in a eighth
4:15grade math test. And this is not the
4:19answer at all. Calculating the number
4:21does not make you a consultant, it makes
4:23you a calculator.
4:25The problem-solving skills
4:28that you need as a consultant are all
4:30about how you get to the number and what
4:33you do with it. All right? So, this
4:35brings me to what is an analysis from a
4:39consultant's perspective. This is going
4:40to make you understand so much more why
4:43I'm saying that these are not math
4:45questions. I understand that that's an
4:48easy way to identify them when you're
4:50talking about them, but it also brings
4:52along a lot of baggage, and it makes it
4:54seem like the math is the most important
4:56part. So, I don't want you calling them
4:58that and getting yourself in the wrong
5:00mindset, okay? So, what's an analysis
5:02from a consultant's perspective?
5:05Conceptually, what you're doing in a
5:07consulting project is coming up with
5:10several hypotheses and then testing
5:13them. For example, you have a hypothesis
5:15that your client's addressable market is
5:18going to reach 100 million dollars in 5
5:21years, or that their customers are
5:24currently unsatisfied with their
5:25product, but the customers are buying
5:28anyway because they don't have any other
5:29option. Or that you if your client and a
5:33specific target merge, they can save
5:36half a billion dollars in procurement
5:38costs by buying in larger books. And
5:41[snorts]
5:42you're going to use data to test these
5:44hypotheses. And you're going to gather
5:47this data from research, from your
5:49client's databases, etc. But databases
5:52alone are of no use. You can't just look
5:55at green data falling like the Matrix
5:58guys and take conclusions from them. You
6:00need to model them into simpler
6:02information, which synthesizes that
6:04data, so that you can
6:07actually process it in your mind. But
6:09these numbers alone still don't tell you
6:11anything. If I tell you that your
6:13client's revenues are $20 billion, this
6:16will have completely opposite meanings
6:19if last year's revenues were $18 billion
6:22or if they were $22 billion.
6:24Or if I tell you that the market size is
6:2640 billion or if the market size is 400
6:29billion. It's the different The meaning
6:31is completely different. And so the
6:33lesson here is that you also need to
6:35compare that number to something else in
6:38order to make it meaningful and in order
6:40to prove or disprove any hypothesis.
6:43And the whole process of testing this
6:46hypothesis or all of these hypotheses
6:49using data, all the way from figuring
6:52out what comparison you're going to
6:54make, then gathering the data, then
6:56modeling it, then getting to a number,
6:58and then actually doing the comparison,
7:01this is what a quantitative analysis is
7:03from the consulting job perspective.
7:06Basically, from all this, what you can
7:09understand is that if you're doing a
7:12quantitative analysis in a consulting
7:14project, you need to have two main
7:16skills, the quantitative skills and the
7:19communication skills.
7:22I will tell you why in a second, but
7:24here's what they are. The quantitative
7:25skills is basically knowing what
7:28comparisons you need to make to get to
7:31that conclusion and where those bars are
7:34to know if something is true or is
7:35false, what numbers from the real world
7:38you need to be able to get to those
7:40comparisons, and how exactly you can
7:43model them to turn raw data into the
7:47information that you need. On the soft
7:49skill side, you need to be able to
7:51discuss on a conceptual level what your
7:55what your analysis is. You need to be
7:56able to discuss it before doing it, and
7:58you need to be able to discuss it on a
8:00conceptual level after doing it, so that
8:02you can talk and explain and get input
8:07from the team and from the client before
8:10and after running the analysis. Now,
8:13the reason you need the quantitative
8:15skills is obvious. If you don't know
8:17what are the comparisons you need to
8:19make, you can't make them. If you don't
8:21know what numbers you need from the real
8:23world, you won't get the numbers cuz you
8:25don't know what they are. If you don't
8:26know exactly what you need to do with
8:27them, you're going to do random stuff to
8:29them, and you're going to get to
8:30inaccurate numbers.
8:32Now, why do you need to be able to
8:34discuss it if you are already able to
8:36get to the right number?
8:38So,
8:40there are
8:41several reasons for that, right? Reason
8:44number one, if you discuss your analysis
8:46with your team before doing it, you
8:48might figure out that something in your
8:51thought process is wrong, and you need
8:53to fix it, you need to run an analysis
8:54differently. You might figure out that
8:57the analysis doesn't need to be done for
8:59some reason, or there's an easier way to
9:02do it, and so you end up saving time.
9:04And one of the worst things you can do
9:06as a consultant, who of course is not
9:09screwing up badly, is wasting client's
9:11time. All consultant projects are under
9:15time constraint. If you waste someone's
9:17time for a week running an analysis that
9:20didn't need to be run, you are all going
9:22to have a nightless week. It's going to
9:25suck suck for everyone. All right? And
9:27your client's going to be pissed, and
9:28you're not going to get to the results
9:29that you want. But there are other
9:31reasons for it. Another reason for being
9:34able to discuss the analysis on a
9:35conceptual level is that you can use
9:37more brainpower in the analysis and in
9:40the project because if you're not only
9:42your brain working on designing the
9:44analysis, it's everyone else's, you can
9:46be more precise, you can do better
9:48analysis, you can get to a better
9:50number, you can also do it more
9:51efficiently.
9:52Another reason yet for that is that so
9:55you can have everyone on board so they
9:57can vouch your analysis. Basically, an
9:59idea that I'm trying to give you here is
10:01that when you're doing a consulting
10:02project
10:04every single analysis that you do, this
10:06might be a bold statement, but every
10:07single analysis that you do
10:09someone will be against it. Some
10:11stakeholders are going to want one
10:13answer, other stakeholders are going to
10:14want another. And
10:16unless you are able to explain very
10:19logically why the number that you got to
10:22is true and it's undoubtedly true
10:26they're going to bash it and the
10:27discussion's going to turn political and
10:29not logical and you're not going to get
10:32your point across.
10:33The other person's just going to say, "I
10:35have experience in this industry. I know
10:36what you're talking about, you don't.
10:38Why would I trust your number?" Right?
10:40So, this is why you need to be able to
10:42communicate and discuss your assumption
10:45on a high conceptual level, not only you
10:48need to be able to perform it
10:50accurately. Okay? Now that you
10:53understand what skills you need to have
10:55and why you need to have them and how
10:57you're going to use them, I'm going to
10:59get exactly into the concrete situation
11:01we're simulating cuz that's going to
11:03make it a lot more clear to you why and
11:06how you should perform in those case
11:08interviews.
11:09We're simulating in a case interview a
11:12discussion about an analysis that you
11:15are having as an associate with a
11:17manager or partner before doing an
11:19analysis. Then we're simulating you
11:22doing the analysis and you're getting
11:25the data
11:26that you would get from the client or
11:29from the market or whatever, you're
11:31getting it from the interviewer. Instead
11:33of placing it all in Excel or FKL or
11:36whatever data processing tool you're
11:39going to use, you're going to use your
11:41pen and paper to do math. And then
11:43you're going to discuss the results with
11:45said manager or partner, which in the
11:47case is the interviewer, right?
11:49The other thing that we could be
11:51simulating is this very same thing with
11:52the client instead of a manager or
11:54partner. So you discuss analysis you're
11:56going to run with the client. They might
11:58like it, they might dislike it, but
12:01will need to agree on a conceptual level
12:03to what you're doing, right? And so
12:06given all of this,
12:08what I want you to understand is
12:11what are the skills that they're testing
12:14for in these situations? What skills are
12:16necessary in the situation that we're
12:18simulating and in all of the situations
12:21that
12:22will use quantitative analysis in the
12:25consulting job. Basically, number one,
12:29they're testing to see if you know what
12:31comparisons to do. They want to know
12:33that you can take conclusions from the
12:36data.
12:37Basically, that is seeing if you can
12:40have a discussion in the end of your
12:42analysis. They're seeing if you can
12:45interpret the number that you got and
12:47that
12:49you can lead from it. You can get to the
12:51next steps. This is step number four of
12:54the analytical process, which I'm going
12:57to show you the steps in a little bit.
12:59Because
13:00if you don't know what are the
13:02comparisons you're going to do and
13:05if you don't know how to interpret the
13:07data, when you need to design the
13:09analysis instead of the interviewer
13:11asking you to calculate something, you
13:13won't be able to because
13:15if you don't know what comparison you
13:16need to do, I want to compare, you know,
13:19the market size of this product to the
13:22market size of the other product that we
13:24have or whatever. If you don't know
13:26that, then how are you going to know
13:28that you need to calculate the market
13:29size? You won't. So, that's why they're
13:31testing this.
13:32Number two, they want to know that you
13:34know how to process the numbers. That is
13:37basically step number two, structuring
13:39your analysis. Because
13:42once you get all of that raw data, and
13:45trust me, it's a ton of data,
13:48how are you going to model it? What are
13:50you going to do to each of your numbers
13:52to get to the answer? If you don't know
13:54how to do this by yourself, then you
13:57can't be trusted alone with a computer,
13:59the data, and the Excel that you're
14:01going to get to an answer, right? Number
14:03three, they're testing to see if you
14:05know what data you need. Basically,
14:07what's going to happen here in so many
14:10cases is that the interviewer is going
14:12to give you a few pieces of data, but
14:14they're not going to give you all of the
14:16pieces of data that you need. And they
14:18want to see you asking for the data
14:22so that they can tell that you know what
14:24data you need.
14:26Right? They gave you some pieces of
14:27data, and that's mostly to guide you in
14:29what direction you should go. But they
14:31want to see that you can understand what
14:34data is missing, and you can look for it
14:36somehow. In this case, asking it. It's
14:39very similar to
14:41the GMAT data sufficiency questions,
14:44because you need to be able to tell
14:46whether you have enough data to get to a
14:48conclusion or not.
14:49Basically, this skill derives from you
14:52knowing the structure that you're going
14:53to need to use and you having the
14:55business sense to know what extra data
14:59you might need. So, what other factors
15:01might be affecting this number that are
15:03not obvious right now. And then you ask
15:05about those pieces of data to the
15:06interviewer.
15:08Next, number four, they want to know
15:10that you know how to communicate the
15:13structure in English rather than in
15:16Excel or in
15:18numbers or only drawings. So, that you
15:20can discuss it like we talked about
15:22before without wasting time and so that
15:25you can
15:26prove that what you've done makes sense
15:28rather than being a guess or being
15:31wrong, right? What you're going to find
15:33in consulting is that
15:36a lot of the work is convincing or
15:40maybe as much of the work in an analysis
15:42that is not uh what everyone wants it to
15:45be as much as convincing as it is
15:47analyzing, right?
15:49And what you might have noticed from
15:51these skills is that math isn't exactly
15:56in there, right? They do want to know
15:59that you are able to do quick and easy
16:01math and that you're not going to screw
16:03up with numbers. You're going to be
16:04dealing with numbers a lot. So, you need
16:05to be able to do some calculations in
16:07your mind in your head and quickly. But,
16:10given all of these situations, math is
16:13not a priority.
16:15That is why so many times you see people
16:18saying, "Oh, I made a tiny mistake. The
16:19interviewer didn't care and we moved
16:21on." If you make many mistakes, that's a
16:23red flag. If you don't catch your own
16:25mistakes, that's a red flag. But, math
16:28is essentially secondary. Excel's going
16:30to be doing most of the math in your
16:31consulting job. And yes, you need to be
16:33fluent with numbers, but that's not key
16:37to the analytical part of a consultant's
16:39work.
16:41And I'm going to show you
16:42a wrong answers only version
16:45of an analysis question and I'm going to
16:47show you what the wrong answers mean to
16:50an interviewer.
16:52So, imagine this is the question that we
16:56get in an analysis question in a case
16:58interview.
16:59So, imagine we are working on a
17:01potential merger and our client is the
17:03manufacturer of a product that is sold
17:06in Starbucks. And of course, the client
17:09wants to acquire a competitor. You're
17:11working on the revenue synergies front
17:13of the project and you get to a moment
17:16where you will calculate the potential
17:18increase in revenues from adding the
17:21competitors products to next to your own
17:24in Starbucks shelves. And we have a few
17:27pieces of data for you here. We know
17:28that we currently sell in 3,000
17:32Starbucks stores. Our products sell for
17:36four seven units per day. And we can
17:39expect per store, of course, and we can
17:41expect the competitors products to sell
17:43at the same rate.
17:44Okay.
17:45And now let's say that I ask my
17:47interviewer, "All right, so
17:50do we know the price at which the
17:53competitors products are going to be
17:54sold?" And let's say the interviewer
17:56says $2 per product. "And do we know how
17:59many days the Starbucks stores are open
18:02a year?" And let's say the interviewer
18:04tells me it's 300. Okay. So,
18:08I'm going to do some math here. 7 * 300
18:12is 2.1 thousand * 2, that's 4.2
18:16thousand. 4.2 thousand * 3,000, that
18:19gives me 12.6 million. So, 12.6 million.
18:23And let's say that I know what the
18:25target company revenues are and I'll
18:28say, "Oh, target revenues would grow 30%
18:31with that, so the acquisition seems like
18:33a great idea." And etc. etc. I say other
18:36things as well.
18:38Do you see the problem here? The
18:40interviewer has no idea what I just did.
18:42The interviewer just knows that I asked
18:44for some pieces of data. And I did some
18:47calculations, which he has no idea what
18:50they were,
18:51where they were going to, what I did.
18:53And the number might even be right, and
18:56he would only know that the number is
18:58right because he is the interviewer, he
19:00designed that question. But in a real
19:02case, in a real project situation, what
19:05happens is
19:06this guy did not create the question.
19:08He's overseeing you solve a real-life
19:11problem. And when you get to him with
19:15answer that is just a number
19:17and you don't at all explain what you
19:19did, you don't explain what you're going
19:21to do before,
19:22he can never know
19:24whether you were right or wrong. If he
19:28can't know if you're right or wrong, you
19:29can't convince him.
19:31And so he has to assume that you're
19:33wrong.
19:34And someone else that has to assume that
19:36you're wrong is the guy in the client's
19:38side that doesn't like your answer.
19:41They will say that you probably didn't
19:43consider a ton of stuff that they did
19:45consider. And they can only assume that
19:48you forgot and that you're wrong. And
19:50also from a case interview perspective,
19:52there's a problem with that because
19:54they might know that you were right in
19:58that one cuz you got the number right.
19:59But since they have no idea how you did
20:02that, they can't be sure
20:05that you would do everything right and
20:07get you the right number in a different
20:08case. They can't know if you were just
20:11lucky. You need to show your work. So
20:14this is why we create a structure
20:17when solving an analysis.
20:19Now, there's also another problem with
20:23the answer that I just gave you.
20:24So
20:25there's something you might not have
20:27noticed that I didn't do, but I didn't
20:29ask the interviewer if there was going
20:31to be any cannibalization
20:33on the units that we currently sell.
20:36So let's say we currently sell seven
20:38units per day. Our competitor that once
20:41we place them there is also going to
20:42sell seven units per day, but since the
20:45consumers are the same that are buying,
20:47some of them are not going to buy our
20:48product and buy only the competitors. So
20:50we can expect our sales to decrease to
20:53[snorts] four units per day or
20:54something.
20:55If that happens, then that changes the
20:58revenue synergies and I should have
20:59taken that into consideration.
21:01Another thing that I could have asked is
21:04are we going to be able to sell to
21:06Starbucks at a higher price since we are
21:08going to,
21:10you know,
21:11have more negotiation power with them
21:13cuz we're going to be selling more units
21:14to them.
21:15And
21:17in those two cases, the answer would be
21:19no. There would be no cannibalization
21:21and the price would not change. I only
21:24know this because I created this case. I
21:26created this case for this specific
21:28example. The case is in my mind. The
21:30answer is in my mind. It can be whatever
21:32I want and in this case it's those don't
21:34change. The fact that they don't affect
21:37the answer do not affect the fact that I
21:40need to ask about these things. Why?
21:44Because
21:45they show to my interviewer that if the
21:47case were different, then
21:50I would still have gotten it right. It
21:52doesn't matter that I got to the right
21:54number if I didn't consider a factor
21:57that does not play off in this case, but
21:59could have played off in this case.
22:02If I don't ask these questions, it shows
22:04to the interviewer that I can execute,
22:06but I can't think. The interviewer can't
22:08trust me with a problem that they don't
22:11fully understand and that they can't
22:14completely double-check my work.
22:17They will always know that I need
22:19someone who will save me, someone who
22:22will check my work and who will
22:25find the stuff that
22:28I need to be creative in order to
22:30consider,
22:31right? And
22:32so
22:34in the end, I won't be saving the
22:36managers and the partners time. I will
22:39be adding more work to their workload.
22:42And so they won't be able to manage as
22:44many projects as they wish they did and
22:46that will get them, you know, more
22:48money, more money to the firm and more
22:51happiness to the clients and whatever,
22:53right? And so
22:55they will be hiring if they hired me,
22:57they won't be hiring someone who helped
22:59them catch mistakes, who helped them
23:02solve problems. They're going to be
23:03hiring someone who will be making new
23:05mistakes and who will be creating
23:07problems for them.
23:09Now, I want you to consider this other
23:10potential answer. Let's say I tell the
23:12interviewer that I first want to find
23:15the revenues per store that we would
23:18have increased to the target's current
23:22revenues.
23:23And then I multiply that by the number
23:25of stores they would go in. And so that
23:27would give me
23:28the
23:29total revenue increase. And then I'd
23:31remove I'd account for a cannibalization
23:34of our own products. And then
23:36I would account for a potential price
23:38increase for both our products and the
23:40target's products given that we're now
23:42selling them more stuff. We might have
23:44more leverage in the negotiation.
23:47And then I ask them if they know what
23:49price our units and the competitors'
23:51units sell for.
23:53And if there might be a price difference
23:55if there is cannibalization expected and
23:58how many days the stores would be open
24:00in the year. Then I run the calculations
24:02like I did before and I tell them that
24:04the answer is 13 million or 12.6
24:07million. But I don't say anything
24:08afterwards.
24:10Right?
24:11The problem here is that
24:14I was not hired to give the interviewer
24:16a number.
24:17I was hired to solve the problem.
24:20And solving the problem doesn't mean
24:23getting to the number. It means
24:25>> [snorts]
24:26>> actually solving the problems that the
24:28project has. So what they would expect
24:31me to do in a real consulting project is
24:34get to this number, know exactly what to
24:36do next, and start doing it. And if that
24:40is
24:41finding the right people to communicate
24:43this number, then that would be finding
24:45the right people to communicate this
24:46number. If that would be finding if
24:49there would be cost savings given this
24:51volume increase or if there would be
24:53price increase given this volume
24:55increase, and figuring that out somehow,
24:57or if that would be finding out if there
24:59are capacity issues, or anything like
25:02that.
25:03That's what they want me to do.
25:05Or if it would even be just finding out
25:08what this 12.6 million means. That's
25:11what they're expecting me to do. They're
25:12not expecting me to give them a number.
25:15And if we're doing this project, and I
25:18get to this number on Wednesday, but I
25:20can only talk to the manager on Monday,
25:24they don't want me to waste 2.5 days.
25:26They want me to know exactly what to do.
25:29And hell, if I am the manager, I need to
25:32be able to tell them what to do, and I
25:34need to be able to do this now, because
25:35it's not such a hard skill.
25:37So, if you don't know how to do this
25:39when you're going in the consulting
25:41firm, they assume that you're never
25:43going to learn. So,
25:45conclusion, you want to show that you
25:48know how to interpret the number that
25:50you get to, and that you know what to do
25:53with it, and what other people should do
25:55with it as well. Now, let me show you
25:57what a right answer would look like.
25:59Let's say that the interviewer tells me,
26:01you know, the case question and all, and
26:03tells me the data, it's 3,000 stores.
26:06We currently sell seven units per day.
26:08We can expect the target to sell for the
26:10same amount of units. And so, I would
26:12tell them, "All right. So, first, what I
26:14would do is find our revenues per store,
26:17the revenues per store that the target
26:19would have. And that would be just
26:21finding the number of
26:23units per year that would be sold, and
26:24multiply that by the price per unit.
26:26Then, I would find the total revenue
26:30increase for the target by multiplying
26:32the revenues per store that I got by the
26:35number of stores. Then, I'd account for
26:38a cannibalization, so I would
26:40find what our revenue decrease would be,
26:42and remove that from the
26:44overall revenue increase.
26:46And then, I would account for a
26:48potential price increase, if there is
26:50any, given that we're now selling a lot
26:53more units to Starbucks, we might have
26:55more negotiation leverage. So, would it
26:58be okay if I asked you some questions
26:59right now? Let's say the interviewer
27:01says, "Yes." So, I'll tell them, "Okay,
27:03I'd like to know at what price do we
27:06expect our competitor to sell their
27:08products on average?" And let's say the
27:10interviewer tells me it's $2.
27:12And I ask them, "Is there a difference
27:14expected in price between what they sell
27:18and we sell right now, and what we would
27:20sell for given that we'd have more
27:23negotiation leverage?" And let's say
27:25they sell they tell me, "No, ignore
27:27that."
27:28And now I want to know, would there be
27:30cannibalization? Do we expect our
27:32products to sell less now that the
27:34target products are selling in the same
27:36shelf?" And let's say the interviewer
27:38tells me to ignore this as well. And
27:40then I asked about the number of days a
27:42Starbucks store is open in a year, and
27:43they tell me 300. Okay? Now, I'm going
27:46to multiply the number of units per day
27:50by the number of days in a year to get
27:52to the number of units in a store in a
27:54year. And so, that's 7 * 300, that's
27:572.1,000.
28:00Now, I want to multiply that by the
28:01price, so I can get to the revenues per
28:04store. And so, 2.1,000 * $2, that is
28:08$4.2,000
28:10per store per year.
28:12Now, I'm going to multiply that by the
28:13number of stores. So, that would give me
28:16the revenue increase in total. So,
28:184.2,000 * 3,000 is 12.6 million.
28:24And since there's no cannibalization
28:27and no price increase, uh that is the
28:30final answer, 12.6 million. And right
28:33now I don't know what the target
28:35revenues are, but I would like to know
28:37so that I can know
28:39like let's say that this is a 30%
28:41increase in the revenues. This seems
28:43like a really good idea of acquisition
28:46because we would acquire let's say we
28:48acquire them for a fair price
28:50but right off the bat we increase the
28:52revenues and therefore probably their
28:55value even even more than that by 30%.
28:58This is an amazing investment by all
29:00means.
29:01Now what I would do next is first find
29:05out what the target revenues are so that
29:07I can actually size the benefit.
29:09But then I try to figure out if there
29:11might be cost savings in the target's
29:15manufacturing process given that now
29:16they're going to manufacture more units.
29:18I would try to find out if there would
29:20be price increase given that now we're
29:23going to sell more to Starbucks. Even
29:26though you told me to ignore that I
29:28that's necessary to know. And I would
29:30also try to figure out if our competitor
29:32can actually increase their capacity in
29:36that much. I don't know how much that is
29:38but if it's significant I want to know
29:41if they can increase their capacity
29:42easily or if there might need to be
29:44investments cuz then I need to take that
29:46into account. That's not you know a free
29:49increase in revenues just from the
29:51acquisition. Now what you should have
29:53seen from this answer that I gave you is
29:57all of the
29:58mistakes that I showed before are not
30:00present. I showed the interviewer
30:02exactly what I was going to do before
30:04then I asked them the questions that
30:07could be the numbers that I needed and
30:10the numbers that weren't logically
30:11necessary but that could be affecting
30:14the case. And even though they were not
30:16used
30:17the fact that I asked those questions is
30:19important. Then
30:21I calculated everything and I guided the
30:24interviewer through as I calculated and
30:26to guide them through I used the
30:28structure that I showed previously. And
30:30then after that I told the interviewer
30:32exactly what that number that we got to
30:35meant and I told them what I would do
30:37next so that they know that I can
30:40not only work by myself but also need a
30:43team if needed.
30:45Now, enough with what an analysis is,
30:49what are the skills, blah, blah, blah.
30:51Let's get right to the techniques that
30:54you need to learn to
30:56ace your case interviews, never get
30:59stuck again, etc. So, basically,
31:02if you learn the step-by-step approach
31:05and the structuring that technique that
31:07I'm going to teach you right now, and if
31:09you practice those until they're
31:11internalized, you will be able to create
31:14a robust structure for every single
31:16analysis question that you'll get, and
31:18you'll be able to generate useful
31:20insights. And you're going to show
31:22Essentially, you're going to show your
31:23interviewer that one, you can work
31:26independently,
31:27which is something that they absolutely
31:29need as they're all the time staffed
31:32under that assumption that all of their
31:34consultants can work independently.
31:36Number two, you're going to show them
31:38that you can take a quantitative problem
31:40off their hands without them worrying if
31:43they're going to need to redo your work
31:45or check your work more than just a
31:48quick glance and not trust you, right?
31:52Imagine if you have six of these
31:53problems, you're overwhelmed, and you
31:55have
31:57in front of you someone who could do
31:59half of your work for you. You're in
32:01heaven. That's exactly where the
32:04interviewer is sitting on right now.
32:07And you'll also be light years ahead
32:10ahead of 99% of candidates who feel lost
32:13in the middle of the problem and need
32:15the interviewer's help to move forward
32:16because they
32:18don't exactly know what they are what
32:20they're doing, and they can't move
32:22forward or move backward.
32:24Candidates who annoy the interviewer by
32:26not knowing the meaning of the number
32:28they're calculating, who have to hope
32:31for an easy or average question because
32:33they know that they can't handle a tough
32:35one, and candidates who are still
32:38playing the wrong game and call this a
32:39math problem. Believe me, the
32:41interviewer can see right through their
32:44poker faces. Now, before I show you the
32:46techniques, let's remember something.
32:49What are our goals when answering an
32:51analysis question? You want to show the
32:53interviewer your thought process so that
32:56they always know where we are, they
32:58always know that you'll get to the
32:59answer, that you will consider all of
33:02the variables so that you and the
33:05interviewer can discuss the issues and
33:07the approaches and the assumptions in
33:09isolation.
33:10Your goal is also to get to the right
33:12number, and your goal is to create a
33:15conclusion that's biased towards action
33:18given the number that you've calculated
33:20and the case question. So, the
33:22techniques that I'm going to show you
33:24are going to help you do exactly this.
33:26This is the four-step approach that you
33:28will follow. First, I'm going to give
33:30you just the steps, and then I'm going
33:32to give you an example of an analytical
33:34question, and then I'm going to go
33:36through it while I explain the steps to
33:38you, and then right away show you an
33:40example of how it's done. The steps are:
33:43Number one, structure
33:45your problem. Number two, ask for the
33:47data you need. Number three, calculate.
33:50And number four, once you get to the
33:52answer, lead from the insights.
33:55Now, let's get to the case question so
33:58that I can explain what each of those
34:01steps mean and how to do it
34:03from a real example and show you right
34:06away how it's done. This case question
34:09right here is comes straight from a
34:12business idea that Bruno and I had way
34:14before we started uh crafting cases or
34:18even case interview coaching. It didn't
34:21go forward. It wasn't really good idea.
34:23There might be ways out, but anyway,
34:25this is where this business case
34:28question comes from. So, your Your has a
34:30business idea, and he asked what you
34:32think of it.
34:34They want to create a chain of high-end
34:37coffee micro kiosks in airports.
34:40The each unit would only occupy around 2
34:44square meters, thus saving a lot of rent
34:47costs, and it would deliver super
34:49high-quality coffee, which wouldn't cost
34:52significantly more per cup if well
34:55purchased, which your friend believes
34:57they can do.
34:58This way, they could easily take
35:00significant market share from the
35:02low-quality overpriced coffee shops
35:04around. How would you analyze whether
35:06this is a good idea or not?
35:08And then let's say you answer with your
35:10framework for this question, which I'm
35:12not going to get into cuz I don't want
35:14to go into the framework question. But
35:16then, you get the following data.
35:19Let's say that the price per coffee cup
35:22is $3.75.
35:24The
35:26the total costs per cup of coffee for
35:30the micro kiosks is $1.
35:33And then there are the costs to open and
35:36create the micro kiosk stand
35:39for only one unit, which is $150,000.
35:43And the yearly operating costs,
35:45including rent and labor and etc., and
35:48that is $200,000 per year.
35:51And the question is
35:53for one single unit, how many cups of
35:56coffee would they need to sell in order
35:58to break even their investment in 2
36:00years? Now, you can start and right now
36:04run through all of the steps if you
36:06already feel safe to do so. What I'm
36:08going to do right now is pause before
36:11each step, explain you more details
36:14about the step, and then perform the
36:15step myself
36:17before moving on. Okay? So, I highly
36:20recommend you
36:22watch my description of the step before
36:24pausing the video and doing the step by
36:26yourself. All right, let's go. Now, I'm
36:28going to talk to you about the step
36:29number one, structure. Structuring the
36:32problem is basically showing to the
36:35interviewer
36:36what exactly you're going to do from
36:39beginning to end to get to the number
36:43that they've asked you to calculate.
36:45So, you'll say, "I'm going to do this
36:47and this and this and that and then I'm
36:49going to get to the final number." It
36:50should take you all the way
36:53throughout the calculations that you're
36:55going to do and it should have no
36:58numbers. It should only be conceptual so
37:01that the interviewer can agree or
37:03disagree with you
37:05in terms of logic and you don't have to
37:08dive and you even can't dive into the
37:11specific of each numbers. All right?
37:15I highly recommend you use step-by-step
37:17structures. There are several ways to do
37:19this. You could use formulas, you could
37:20use trees, but I highly recommend you
37:23use step-by-step structures because
37:25they're intuitive, they work for every
37:27single case, and they also give you a
37:30little bit of space to figure some stuff
37:32out on the go, but already have
37:35a
37:36proven path that will get you to your
37:39final answer. Now, I'm going to show you
37:41a technique. I'm going to explain to you
37:44a technique right now that will teach
37:46you how to create step-by-step
37:47structures for any problem. You're of
37:50course going to need a little bit of
37:51practice and that's what you're going to
37:53get in the next videos and also in this
37:55example right now, I'm going to show you
37:56myself using it, but it's a structure
37:58that if you practice, you will never
38:01have trouble
38:03getting to the structure for a problem
38:06again unless it's a really complicated
38:08problem you've never seen before and
38:09etc. etc. if it's super hard. Then, you
38:11might need a little bit more of mental
38:14power and perhaps the right kind of
38:16practice, but this technique is already
38:19a super strong tool even for those
38:21cases. So, basically
38:23here's what you're going to do.
38:25You come up with a number. You You look
38:27at the numbers that you have. You look
38:29at the number that you need to get to.
38:31And you come up with a number
38:33that you know that you can get to and
38:35that you know that will get you there.
38:38Right? So, intuitively you know that you
38:40can get to that number and that it will
38:41get you there.
38:43Maybe you need to ask for data from the
38:45interviewer, that is fine. But, you know
38:47that you can get to it even if you don't
38:49know anything else. Now you have that
38:51number.
38:51Okay? And that is step one.
38:54And then you look at everything that you
38:56have again and you come up with one
38:59number that you can get to, that you
39:01know that will help you get there.
39:03That's all you know.
39:05That is step two. And then you repeat
39:08that again and again and again until
39:11the last step gets you to the final
39:13number.
39:14Okay? I'm going to show this to you
39:16right now as I solve this question by
39:19myself.
39:20In a real case, I would have asked the
39:22interviewer for some time and then I'd
39:24create the structure in my mind. But, in
39:26this case, I want to think out loud so
39:28that you can follow me. Okay? So, I'm
39:31looking at these numbers and one thing
39:33that I know that I can get to and that I
39:34know will help me somehow is
39:37the gross profits per cup of coffee,
39:41which is basically the price minus the
39:43variable costs. I know that's going to
39:45help me cuz it's the contribution that
39:47each cup of coffee gives me
39:50to cover those fixed costs and
39:53investments that we need to cover to
39:54reach break even. Okay?
39:57Now what I have is fixed costs,
39:59investments, and profits per cup of
40:01coffee. And so, the next thing that I
40:04know that I
40:05can get to and that I know that will
40:07help me is the total fixed costs and
40:10investments that I'm going to have over
40:13these 2 years. So, that's basically 2
40:15years worth of fixed costs plus the
40:17investments. And this is my step two.
40:20Getting to total fixed costs and
40:22investments in 2 years. And then my step
40:24three is going to be finding how many of
40:28these tiny bits of profit I need in
40:30order to cover for those costs and
40:32investment. So, that's step three,
40:34actually getting to the number of cup of
40:37coffee that I need to sell. Okay? So,
40:40this is my structure. After thinking
40:42through this way, I present it to the
40:44interviewer as step number one, get to
40:47the growth profits per cup of coffee.
40:50Step number two, getting to the total
40:52fixed costs and investment that these
40:55growth profits will need to cover over 2
40:57years. And step number three, dividing
41:00the total fixed costs and investment by
41:02the profit per cup of coffee to get to
41:05the break-even number. Okay? Now, let's
41:08talk about step number two, asking for
41:11data.
41:12Many numbers that you will need are
41:15going to be clear to you that you're
41:16going to need and that you're going to
41:18need to figure it out somehow from your
41:20structure because
41:22you are going to need to divide the
41:24number that you have by a number that
41:26well, the interviewer didn't tell you,
41:27so you know that that's a piece of data
41:29that you need. But there's also another
41:32type of potentially missing data that
41:35you should look out for. And that is
41:37things that might affect your answer,
41:40but that don't necessarily need to be
41:43there. Kind of like the cannibalization
41:45effect that I was talking about in the
41:47previous example, right? And in this
41:49case, it could be something like, let's
41:51say we are also selling something else
41:53other than coffee. The interviewer just
41:56didn't bother telling us, but we're also
41:58selling muffins or whatever or mugs, I
42:01don't know. And so, that would change
42:04because that would change the structure
42:06because we would need we would have
42:08something else that's also covering for
42:11costs and investments, And so, that
42:13would change the break-even.
42:15Now, that's something you should always
42:17look for when doing analysis questions
42:20because
42:22even when they're not there, the
42:24interviewer expects you to ask for some
42:26stuff because
42:28that shows them that you have the
42:30creativity to think about other factors
42:33that weren't mentioned that but that
42:35might change the answer, which happens
42:37all the time in the consulting job. And
42:40they want to know that they can trust
42:41you to solve the problem by yourself and
42:44not have them need to ask you in the
42:47meeting when you present your results,
42:49but did you consider this? Did you
42:51consider that? Did you consider that?
42:53Okay?
42:54Another thing that you should know about
42:56asking for data. This is a mistake that
42:59a lot of people do. They see that they
43:01need a piece of data and so, they right
43:03away start estimating your guessing this
43:06that piece of data or that number or
43:07whatever. And you should always request
43:11the missing numbers before estimating
43:14your guessing the numbers.
43:16Think about the interviewer like a
43:18skeptical client. They know the
43:20industry.
43:22They know a lot of numbers, but they
43:24don't know what you want until you ask.
43:27What this means is if you ask, they know
43:29it and they're going to give it to you.
43:31And if they don't know it, then you can
43:32estimate. But if you don't ask and start
43:35estimating and they know that number,
43:36you're going to look silly cuz you're
43:38going to be estimating and you're going
43:39to be Of course, you're going to be
43:41wrong compared to someone who has the
43:42real number in their minds and that's
43:44just going to look silly. That just
43:45makes you look bad. Now, if they don't
43:48know, then you can estimate because both
43:50of you don't know and you just look
43:51smart cuz you're
43:53calculating a number from scratch.
43:54You're inventing a number from scratch.
43:56Okay?
43:57Now,
43:58one question that people ask us a lot is
44:01when do you ask? Do you necessarily need
44:04to ask right now after the structure?
44:07And this is preferable. If you want to
44:09ask before the structure, you might get
44:12the interviewer a little bit confused,
44:13but that is generally fine. If you want
44:16to ask during the calculations
44:18afterwards, it's a little bit worse
44:21because you you kind of show to the
44:23interviewer that you'll start working
44:25without knowing all of the numbers and
44:27having all of the ingredients, which is
44:29a little bit weird, but it's also fine.
44:31The timing of the questions is not that
44:33important. Okay? Just make sure that if
44:37you get a new number during your
44:38calculations and you need to change your
44:40structure, you will stop and you will do
44:42this explicitly. So, "Oh, so this means
44:45that I need to change my structure in
44:46this and this and this way." And then
44:48you do it. All right? Now, in this case,
44:51the only question that I would ask, the
44:53only thing that's missing here is I want
44:55to know if we're selling something else
44:57other than only cups of coffee. And the
45:00interviewer will tell me, "No, we're
45:01only selling cups of coffee." Okay? Now,
45:04let's get to step number three,
45:06calculating.
45:07As I told you before, mental math and
45:11being really quick and being a genius
45:13calculator is something that is
45:16secondary in case interviews.
45:18Interviewers don't really care that much
45:21if you can, you know, divide 3,000 by 17
45:26in a split second. They really care that
45:29you are precise and they really care
45:31that you're efficient. So, you've got to
45:34be aware of that trade-off between
45:35precision and efficiency. All right? And
45:38you've got to find the soft spot for
45:40you. You've got to make sure that you're
45:41precise and then you add efficiency to
45:44your process. If there's a faster way,
45:46you use the faster way. If rounding will
45:49be a lot better, it will not change the
45:51result a lot, then you will round or you
45:53will request to round. Okay? This is how
45:57it goes. Don't worry about it too much,
46:00but of course make sure that you're not
46:01making a ton of mistakes. And if you can
46:04add efficiency, go ahead and add
46:06efficiency. Now, what a lot of people
46:09worry about, and you might be worrying
46:11about right now, is what if I make a
46:12mistake? And many people think that this
46:15means that they've lost the job and
46:17whatever. And most of the times it's not
46:19a problem. If you do only one mistake in
46:22the interview, the interviewer is just
46:23going to brush it off and say it's all
46:25right, everyone makes mistakes. Happens
46:27all the time. But, if you do it often,
46:31that is a problem. We want to avoid any
46:33mistakes, okay? But, if you do them
46:35often, it's a really bad problem cuz the
46:37interviewer is going to be annoyed, and
46:39they're going to think that you're
46:40careless and reckless, and that you
46:43might make those mistakes in a important
46:47conference, and not like important
46:50conference room, and important meeting.
46:52Might not check your numbers well, and
46:53you might not check other people's
46:55numbers well, and so that's not really
46:56good. So, if you make a mistake, just
47:00follow on, and do a great job in the
47:02rest of the case, the interviewer won't
47:04care. If you are practicing, and you
47:06notice that you're making a lot of
47:07mistakes, then you've got to solve that
47:09problem, okay? And in my opinion, I'm
47:12going to be quick here cuz this is not
47:13key, but in my opinion, the best way to
47:16fix that problem is before you do any
47:19math, any math that is a little bit more
47:22complicated, you do a quick reality
47:25check before you run the numbers. So,
47:27like 28 /
47:306. All right, you know that 28 is 7 * 4,
47:33so you know it's close to 4. But, then
47:36you do the right math afterwards. Now,
47:39you know that should be close to 4, but
47:41it should be a little bit more, and then
47:42you know, do the calculation, you get to
47:44the right number. All right?
47:46Now, let's move on to the actual
47:48calculations in this case. So, step
47:52number one, I'm going to find the profit
47:54per cup of coffee. And that is 3.75 - 1,
47:58that's 2.75.
48:00Step number two, I'm going to find what
48:02is two years worth of fixed costs plus
48:05the investments. So, that's 2 * 200,000
48:09+ 150,000.
48:112 * 200,000 is 400,000
48:14+ 150 is 550,000.
48:18Now, step number three, 550,000 / 2.75.
48:23So,
48:24first I'm going to
48:25quickly see this as 600,000 / 3. Both
48:29numbers are close
48:30to that, so it should be around 200,000.
48:33Now, I'm going to actually divide 550 by
48:36275
48:39to get to this number and
48:41it seems like 550 / 275 is two,
48:45actually. So, the number really is
48:48200,000.
48:49And so, this is the number of cups of
48:51coffee that we need to break even in two
48:54years.
48:55200,000, yes.
48:57And now,
48:59we're moving on to step number four,
49:01which is getting insights and leading
49:04from them, finding next steps.
49:06Keep in mind that this is the most
49:09important step in an analysis question.
49:12And not only from a case interview
49:14perspective, but also from a consulting
49:17job perspective. Because this is what
49:19makes consultants difference. It's what
49:22makes their job so valuable to the
49:24client. It's what makes McKinsey's
49:26projects so expensive. It's because they
49:30give guidance to the client. They tell
49:32them what to do to get the most value
49:34out of their company, right? And
49:36basically, what you're doing in this
49:37step is
49:39telling the interviewer what that number
49:43you got to means to the case and what
49:46should be done based on it. And you're
49:49also telling them any other interesting
49:51insight that you see from everything
49:52that you did that might help get to an
49:54answer to the case question.
49:56What you're showing is that you're not
49:58the one that's going to sit and wait for
50:00guidance once you're done. You are the
50:02one who knows how to use that number,
50:04and you know how to guide yourself and
50:06the project using it. And you, because
50:09of that, are the person that people will
50:10look for when they need guidance.
50:13This is a skill, as I said before, that
50:15everyone in a consulting engagement must
50:17have. It's expected that you're able to
50:20do this, and this is why it's tested in
50:24the case interview. Now, you're going to
50:26do this in three steps. I know this is
50:28step four, but I'm going to divide it up
50:30in three more steps.
50:31Step number one, you're going to compare
50:33the number that you got to to some other
50:36number. And
50:38you're going to do that to have a basis
50:41to say if it's too high or too low, too
50:43much or too little, or whatever, right?
50:45And you're going to have to find what
50:46that comparison is. Number two,
50:49you're going to say what this means from
50:53the case question point of view. So,
50:55what does this number mean to the entire
50:57engagement that we're doing, to the case
50:59question that the interviewer told us
51:01right in the beginning of the case? And
51:03number three, you're going to say what
51:05you and the client and the team should
51:08do next based on that. Okay?
51:11Now, let's get to what I would say in
51:14this case.
51:15Well, basically, I would look at this
51:17200k and I would say, "Oh, okay, so
51:20200,000 cups of coffee in 2 years."
51:22If we assume 350 days in a year, that's
51:26around 300 cups per day.
51:29And if we assume that the coffee shop is
51:32open
51:3315 hours a day,
51:35uh that would mean 20 cups an hour, or
51:38one cup every 3 minutes. And that's not
51:41only in peak times, but also like 10:00
51:44p.m. when not that many people drink
51:46coffee.
51:47This looks way too optimistic in my
51:49opinion. Like I'm thinking of looking at
51:51a coffee shop in a in the airport. And
51:53of course, there's a lot of movement,
51:56but not necessarily all of them are
51:58buying coffee. And honestly, very little
52:01of them are buying only coffee. And
52:04since this is only one location, we
52:05can't really assume that we can take
52:07market share from all of the airport,
52:09only from the surroundings.
52:11So, it makes sense to think of one or
52:14two specific coffee shops that would be
52:16close to this one.
52:17And we can only
52:19get the customers who care about the
52:21quality of coffee, which are obviously
52:23not all of them.
52:25And customers who don't want anything
52:26else. That are not going into the coffee
52:29shop and asking for, you know, a cup of
52:31coffee and something to eat or whatever.
52:33Cuz
52:34they want to buy everything at the same
52:36place. So, that's not a very large
52:38fraction of the customers. I would not
52:40trust that we can easily get to those
52:43200,000 cups of coffee in 2 years. Now,
52:47if I were to do something next, it would
52:49probably be see if we can sell something
52:52else in order to increase the average
52:54ticket of each customer. So, for
52:56example, we could sell coffee beans,
52:59bags of those coffee beans, so that
53:01people can take home to to brew their
53:04own coffee or so that they can give away
53:06as gifts with a very high value added.
53:09Like if it's if these are really high
53:11quality gifts that are well sourced
53:12straight from the farm or whatever, we
53:14can probably package those as gifts and
53:17sell them at a high price. And this
53:19might make the business worthy, but it
53:21doesn't look too promising to me from
53:24the perspective of only selling cups of
53:27coffee.
53:28So, now you might be asking yourself,
53:30what happens next? Like we're doing a
53:32real case here. What happens next?
53:35If this were a candidate-led case, then
53:38the interviewer would just allow you to
53:40follow on with your next step, and maybe
53:43give you some data for that one next
53:45step. They might question your
53:48assumption, your conclusion, and then
53:51you can discuss that conclusion with
53:52them.
53:53Or the interviewer might just say,
53:54"Time's up. I want a recommendation."
53:57And you might have at least a better
53:59recommendation to give than before
54:02before doing the analysis, since you
54:04already know what that number means to
54:06your case question.
54:07If this were an interviewer-led case,
54:10then
54:11the interviewer might be satisfied and
54:14just ask you a new question, cuz you
54:16fulfilled everything they wanted to
54:17know. The interviewer might be curious
54:19and impressed with your follow-up, and
54:22ask you for a follow-up question. Like,
54:24for example, "What other kinds of things
54:26can we sell other than coffee bean bags,
54:29so that we can include increase the
54:30average ticket and make this work?" Or
54:32maybe they could ask you, "What are
54:34other places that we could place this
54:35kiosk if we're able to make it
54:38profitable? Give me some ideas."
54:41The most common mix is the most common
54:44type would be a mix between the
54:48interviewer-led and the candidate-led
54:49interview, where the interviewer might
54:52be satisfied, like it's a candidate-led
54:55case generally, but the interviewer is
54:57satisfied with what you show them, and
54:59they need to see other types of things.
55:00They need to test other stuff in your
55:03skills, and so they'll just ask you a
55:05new question, or tell you to go a
55:07different way, or whatever. And that's
55:09that's the most common path, honestly,
55:11unless it's a McKinsey first-round case,
55:13where they have the questions lined up.
55:15This is probably how it would go.
55:17Now, if this were a real project, then
55:20after the analysis, you have tested the
55:23hypothesis you wanted, you go back to
55:25the beginning of your project, and see
55:28what this means to the big picture,
55:29create those next steps, and go do them.
55:32You'd probably present this to someone,
55:34to the manager, to the partner, discuss
55:36those next steps that you came up with,
55:38and then go do them. And in this case,
55:41it would probably be, as you saw,
55:43finding ways to improve the average
55:45ticket so that we can make this business
55:47work.
55:48All right, this is it for right now.
55:51Now, I'd highly recommend you go check
55:54out the drill videos and work through
55:57the step-by-step process and internalize
56:00it. Internalize the technique to create
56:03structures by yourself, also, so that
56:06you can be able to create structures for
56:08any case and never again get stuck in an
56:11analytical question. Cuz I know that
56:13there's no worse feeling than blanking
56:16out in the middle of an analytical
56:18question case. Okay?
56:20Right, I'll see you in the next video.

About this transcript

This page contains the full transcript of Case Interview Fundamentals: Lesson 24 [Quantitative Analyses: The 4-step approach] by CraftingCases, generated from the public captions YouTube serves with the video. The transcript has 9,780 words across 1,483 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to choose a caption track available on the video.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, caption language selection, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.