Leading with a hypothesis

Hi there,

You take a few things for granted here, so let's back up.

If your goal is to solve the case, and the structure is your analytical blueprint for doing so over the next 40 minutes, then every element in your structure must be tied to the overall objective. Moreover, it must be a prioritized list of things to analyze, neglecting those things with relatively minor impact.

• The question is: how do you know whether an analytical stream in your structure has impact and helps you solve the case?
• The answer is: you don't know but you establish it by hypothesizing and you do so for each element in your structure i.e., give your explanation as to why this element is there and what you think may be the outcome if we were to analyze it.

Leading with a hypothesis is merely prioritizing from this list according to the highest impact per time spent. Here, too, you should argue as to why you want to start with x and not y, before proceeding to analyse x.

Note: the above only applies to candidate-led cases and is really quite different from interviewer-led (McKinsey) cases.

Hope this helps. Best of luck!

Moritz

You don't. You structure around an hypothesis. Or even better: you structure around what a decision looks like (e.g. I will enter this market if it is “large enough”, a “relevant market share” is possible, I can “beat the competition" for the target segments, I have access to the relevant “channels” and I can deliver the product with “adequate margins”).

Buckets are relevant things to explore, but they are not a problem solving approach. You need to put yourself in the shoes of the decision maker, think about what would be the investment criteria/decision criteria, and then build an approach based on that.

Hi there, 

First of all, there's a common misconception about leading with a hypothesis. The V Cheng era of come up with hypothesis right from the beginning is long gone. It makes absolutely no sense to always try and offer a hypothesis and it becomes ridiculous when candidates generate one without having any supporting evidence.

Basically, only come with a hypothesis the moment you feel you can support it, i.e., if you were asked why you are suggesting that hypothesis, you could point at a couple of data points that you've already arrived at in the case. Don't do it earlier. 

Best,

Cristian

You shouldn't.

Imagine Exxon is paying your team $10M to figure out how to increase Revenues, and on the first day you say “so, I'm thinking we should build a ton of pipelines in Canada”.

That's ridiculous.

You need to create a framework. You need a hypothesES driven approach.

In some cases the prompt provides a lot of clues/hints as to what's going on. You need to incorporate these into your framework and “carry” your explanation with some hypotheses based on that, but you still need a hollistic, MECE approach to solving the problem.

Based on your Q&A I highly highly recommend coaching for you (this is a fundamental mindset shift required that is hard to get on your own)

Hello,

The short answer is no. If there is something obvious that you can point out and score some ‘business acumen’ points, then sure go ahead [E.g. say the cost of making a product improvement that will cost $100K for a company that's on the S&P500 is obviously negligible. So as long as there are no obvious red flags, your recommendation should be to give a go-ahead]

Otherwise, it would be premature to come up with a hypothesis when you have no idea what the answer is going to turn out to be.

Best,

Rushabh

Some great and practical responses from other coaches already. I'll add a bit more context on how it actually works on the job (which should explain why they are saying what they are saying).

On a real client case that requires problem solving:

• You need a ‘framework’ to guide how the team will get to the answer. Quite often, this ‘framework’ may also structure how the team is organized into modules/workstreams
• Very early and at the start of the case, the team comes up with either a hypotheses or issue tree
• Depending on office/case, often this hypotheses tree is created and prioritized with the input of a senior expert (e.g. Partner on the case, internal expert, external expert)
• Sometimes, the case leadership is already able to get a sense of where to prioritize based on interaction with the client and/or access to client context and existing data
• As a junior consultant, rarely will you have the experience and knowledge/data to create really strong hypotheses & be able to prioritize on Day 1

Solving the case interview is in some ways similar, hence the other coaches are saying it makes no sense to just blindly ‘lead’ with a hypothesis and there is no way even on the actual job we can go to clients and say "we think this is the hypothesis…just because".

Hope the above helps you to understand a bit more about consulting!

Hi there,

There is a common misconception about what hypothesis-driven thinking means. 

Whenever you work on a case, you are using multiple hypotheses at the same time, all the time (probably unconsciously). 

• For instance, when you are thinking about your initial structure you would only look into buckets that you think can help you understand the problem better. Every idea you list is based on a hypothesis. 
• Once you start drilling into your framework to get more data, you are focusing on the most promising areas first (again since you are hypothesizing that those areas will help you solve the problem in the quickest and most elegant way).
• When you get new data, you will automatically analyze it in the context of the case and your understanding of it.
• etc etc.

The underlying principle is always the same.

I have a whole chapter in my new book on hypothesis-driven thinking and how to use it at each stage of the case. Feel free to check it out.

Cheers,

Florian

Hi there,

I think this is an interesting question that may be relevant for many people. I would be happy to share my thoughts on it:

• First of all, structuring into "buckets" may not be the most meaningful approach, unless you are asked to consider specific dimensions to explore. Instead, consider the initial structure as a roadmap to guide your analysis. You should focus on listing analytical activities, rather than dimensions, in your structure.
• Additionally, Victor Cheng's advice to always formulate a hypothesis at the beginning of a case study may not be the best approach. It's important to understand that a hypothesis is simply a tentative assumption based on limited information - which you may not have at the very beginning of the case study. It's more effective to view a hypothesis as a starting point for discussion and further exploration, which you can share with your interviewer.

If you would like a more detailed discussion on how to address your specific situation, please don't hesitate to contact me directly.

Best,

Hagen