My question is when should you say your hypothesis? .
Scenario 1 : there are two main drivers for profit which is revenues and costs. This is my main two buckets and before I dive deep into them my initial hypothesis is xx
Scenario 2: i read out my whole structure then state my hypothesis.
And how wide is the hypothesis in the beginning? What could it be for example for a profitability case?
I get this question a lot, and I think there’s a common misunderstanding here.
A hypothesis is not a guess. It should be based on some evidence or logic, not something you “throw in” at the start.
So for your two scenarios: I wouldn’t do either. Stating a hypothesis right after reading the prompt or right after your structure is, in most cases, just guessing; and that doesn’t add value. At the beginning of a case, your job is to:
Not to predict the answer without data.
The only exception is if you have very specific prior experience in that exact industry; then you can say something like: “In similar situations I’ve seen X being the main driver, so I’ll keep that in mind as we analyze.”
But even that should be light, not a strong claim.
The right moment to use hypotheses is during the analysis. For example, after looking at a chart or doing a calculation, you can say: “Based on what we’re seeing, I’d hypothesize that the issue is coming from X rather than Y.”
That’s much stronger because it’s grounded in evidence.
For a profitability case, a good hypothesis would be something like: “Given revenues seem stable and costs have increased, I’d focus on cost drivers as the likely root cause.”
Bottom line: don’t force a hypothesis at the start; build it as you learn more.
Hope this helps,
Franco
Hey Anonymous
Thanks for your question. I hear this daily from my coachees, and I honestly think that there's some confusion going on.
The basics
Hypothesis in the structure phase is not needed in standard MBB case interviewing. It's something that has been emerging, particularly in the US post-MBA market, where Firm coaches (often assigned by their firms to help candidates from specific universities) sometimes suggest stating a hypothesis as early as possible, including during the structure. I heard this from my coachees at Berkeley a few times, but even in the US this is not the norm.
The reason is simple: it's true taht consulting is a hypothesis-driven business and you want to start as early as possible to develop an hypothesis. However, you need solid data or solid evidence (or at minimum some qualitative context) before you can develop a meaningful assumption. Without that, you're just guessing.
Your scenarios
Based on that, neither scenario 1 nor scenario 2 works for me.
So when should you actually use it?
A. The first and most important principle is that you should always be hypothesis-driven from the very first piece of data you receive after your structure. The moment you get a number or a fact, you should be asking yourself: what does this tell me, and what hypothesis can I build from it? This is exactly where many candidates fall short: they read the data out loud and stop there, without developing any forward-thinking assumption from it.
B. The second moment is more of an edge case that happens in roughly 10-15% of cases: it is when your clarifying questions turn into a genuine conversation, one where you develop a real understanding of the industry and the situation before you even structure. This happens more often with senior partners. In that case, you might have enough context to frame an hypothesis, and importantly, it doesn't have to be a hypothesis on the solution. It can be a hypothesis on where to focus your analysis.
--> Here's a concrete example of case B. It's an M&A case, and during the clarifying questions, you learn that asset prices are very high, that the industry is very scale-dependent, and that you are client is an industry-leader in cost management. You could say something like:
"Before I lay out my structure, I want to flag an initial hypothesis: given current asset price levels, this M&A activity will only generate value if cost synergies are well-developed, relevant, and implementable within a one to two year window." That's not a full answer, nor a standard hypothesis, but it's a well-reasoned intuition that narrows the scope of where you'll focus most, and that's exactly what makes it useful. I would still caveat it with "I would focus first on this hypothesis" -- this way you show that you know that this intuition might also be wrong :)
A word of caution
Even in case B, watch out: if you force it, especially in the European market, most partners will question why you're building an hypothesis before you have any data. If it doesn't feel natural, don't do it.
Hope this helps!
Best,
Tom
Hi there,
OBJECTIVE-DRIVEN APPROACH
I personally am "against" the phrase "my hypothesis is". There should be many hypotheses, at many levels, throughout your framework and the case.
Rather, you need to have a hypothesES-driven or, rather, an objective-driven approach. Your entire framework is a set of hypotheses and views as to how to solve a problem.
In my view, the more natural the better. I tend to say things like "My thinking here is x" or "Based on what I know about x and y, I think this'll likely happen" or "My inclination is x".
Please get away from saying anything generic, i.e. hypothesis, framework, buckets, clarifying questions, etc.!
Example Using My Chinese Chess Case
Let's take my "Chinese Chess" case as an example: Chinese Chess Case
One "hypothesis" should be that Revenues are the main reason for our crashing profits. This makes sense since the profit fall is covid induced. As part of this, we would also venture to guess that volumes are the main culprit as demand has shifted left and gov. regulations have prevented us from operating (hypothesis). There could also be price cuts as we're trying to attract people to flying (Hypothesis). Ultimately though, we need to look at how we can bring volumes back (Hypothesis). Again, we can hypothesize that volumes can be brought back by 1) Making sure we're meeting gov. standards around covid (hypothesis) and 2) Stoking demand by running promotions, marketing, and ensuring people we're safe to fly with (hypothesis).
Then, we want to say that, while we want to get back existing revenues, we'll probably have to find alternative revenue streams (again, another hypothesis). We can do this by offering other services like cargo shipping (another hypothesis).
Our 3rd bucket would be, if we can't bring back our profits by recovering existing revenues AND we can't bring them fully back by finding new revenue streams, we need to cut costs to make up this difference (hypothesis). We could probably do well by cutting excess capacity in the form of labor and leases (hypothesis).
Do you see how "hypothesis-led approach" is a ridiculous statement and out of date? How crazy it is to say "My (singular) hypothesis is"?
Rather, you need a hypothesES-driven approach, or rather, an objective-driven approach.
Another way of viewing it:
Your framework is your structure for approaching the problem. It consists of a few main areas you'd like to look at. Inherent in your framework is a view that "If I answer A, B, and C, then we have an answer"
So, for a profitability case:
1) If revenues are the main driver, and volumes are down, then we still want to understand why
2) If #1 = yes, then let's see if it's external (market) or internal (our own doing)
3) If #1 and #2 = yes, then, what can we actually do about it?
Then, your summary becomes "I believe revenues are the issue, likely volume-driven, and I want to understand if this is market wide or specific to us."
^Now this is a hypothesis :)
Read these Q&As for some great context:
At what point in the case does the interviewee state hypothesis?
Interviewer-led case interview hypothesis and ideas
This is a tricky topic that's difficult to properly answer in writing... if you want a more thorough explanation and training in the mindset shift required here, don't hesitate to reach out: Book a session
This trips up a lot of candidates.
Scenario 2 is the right answer. Lay out your full structure first, then close with your hypothesis. Your hypothesis should come from your structure, not float before it. Stating it after shows the interviewer you have thought it through and know where to dig first.
Scenario 1 is not wrong, just rushed. You are giving your guess before showing how you got there.
On how wide, keep it directional and tied to a specific branch of your structure.
A weak hypothesis for a profitability case sounds like, "I think the issue is revenues or costs." That is just your structure repeated.
A stronger one sounds like, "My hypothesis is that the profit decline is driven by falling revenues rather than rising costs, most likely from a drop in volume in the core segment. I would like to start by testing that."
You have picked a branch, named the likely driver, and told the interviewer where you want to go first. That is what BCG wants.
Last thing. Your hypothesis does not need to be right. It needs to be specific, logical, and testable. You can update it as you go.
Good luck.
Good question — and something people often overcomplicate.
First, timing:
Best practice is to give your hypothesis after you lay out your high-level structure, not before.
So between your two options → Scenario 2 is better.
Reason is simple:
If you give it too early, it can feel a bit random.
On how “wide” the hypothesis should be
At the beginning, keep it simple and directional, not overly detailed.
For a profitability case, something like:
That’s enough.
You don’t need to be super precise. The goal is to:
A couple of tips
So overall:
That’s what interviewers are looking for.
This is a common sticking point, but Scenario 1 is much closer to how we actually work at the firms. Think of the hypothesis as your Day One view—it’s the mechanism that prevents you from going on a "fishing expedition" through every single bucket. You should state it immediately after laying out your high-level pillars so the interviewer knows why you’re about to prioritize one area over the others.
If you wait until you've listed every sub-bullet in your structure, the hypothesis feels like a tacked-on afterthought. For a profitability case, keep it grounded in the "where" rather than the "how." Instead of suggesting a solution, hypothesize about the root cause: "Given the market entry you mentioned, my hypothesis is that the profit hit is driven by price erosion in the core segment, so I’d like to start with the revenue side." It’s less about being right and more about showing you can prioritize under pressure.
Hope this helps!
hey there :)
good question, and this is exactly where many people overcomplicate things. at firms like BCG, the hypothesis is not a separate “block”, it’s just a quick directional guess to guide your structure.
so you don’t need to wait. best practice is: state it right after you outline your structure in one short sentence. scenario 2 is closer to what they want, but keep it very tight.
for a profitability case, your hypothesis should be simple and a bit specific, for example: “my initial hypothesis is that profits are declining due to rising costs rather than revenues, but I would like to test both.” that’s already enough. it shows direction without locking you in.
key point: keep it narrow enough to sound thoughtful, but flexible enough that you can pivot easily once you see data.
best,
Alessa :)
That's a great question.
A hypothesis is a belief supported by evidence. When you have overwhelming evidence, then it becomes a recommendation.
So you should communicate a hypothesis when you have data from within the case that allows you to support that belief.
The earlier you do it, the better.
However, if you don't have supporting evidence, then you don't have a hypothesis, so refrain from communicating one prematurely.
Best,
Cristian
