Taking an hypothesis driven approach to a problem means attempting to solve that problem by focussing on your best hypothesis as to the answer. This helps arrive at a solution quickly and efficiently.
The idea of being "hypothesis driven" is one of those concepts which is thrown around constantly, both in consulting circles and in the research world. With a quick Google search, you will find a slew of articles, books and videos about the hypothesis driven approach. Looking through these resources as ex-consultants left us quite puzzled, as the various authors all seemed to be trying to needlessly complicate a relatively simple concept.
To the uninitiated, the term "hypothesis driven" might either seem rather technical, complex or, alternatively, like a meaningless piece of management speak. However, the idea of taking an hypothesis driven approach to problems is a simple but genuinely powerful idea which should characterise your approach to all aspects of your interview.
Let's get a better idea of what exactly we mean by the term "hypothesis driven", before seeing how this way of working differs from more standard methods.
1. What is it?
When we work in an hypothesis driven way, if the cause of some problem is A, B, C or D, we start by assuming the most ostensibly likely option - say A - is the cause and then checking whether this is true or not. If it is not, then we iterate the same process with the other options - B, C and D - in order of their likelihood until we find the correct answer.
This seems pretty sensible, but it is actually very different to how most people will work in practice. Especially if you are coming from an academic or research background, you will likely try to analyse the entirety of a system before going on to focus on any of its parts. This is all very rigorous and appropriate in its own context, but consultants are employed to generate solutions. Analysis is only a means to that end.
The hypothesis driven approach is both dynamic and efficient and means that you will always be moving forward towards a solution to the main problem. You will maintain tight focus and avoiding getting side-tracked trying to do everything.
The best way to appreciate the difference between hypothesis driven and more typical approaches to problem solving is to have a look at an example of how they differ in tackling the same case. To this end, let's take a look at two ways in which a candidate might respond to the following case prompt:
Revenues for a leading supermarket chain have been declining over the last 10 years. How can we address this?
2.1. Non-hypothesis driven Approach
Given this prompt, a candidate taking a non-hypothesis driven approach would likely come up with a list of possible reasons for the observed declining revenue. This would then be delivered back to the interviewer as a list of questions seeking more information as to each possible cause. Thus, the candidate might respond:
We know revenues have declined. I want to find out why:
- Did customers change their preferences?
- Which segment has shown most decline in volume?
- Is there a price war in the industry?
- ...etc, etc...
2.2. Hypothesis Driven Approach
By contrast, a second candidate taking a more hypothesis driven approach might tackle the case as follows:
Candidate: We know revenues have declined. This could be due to price or volume. Do we know how these changed?
Interviewer: Prices have remained constant. Volume has declined in line with revenue.
Candidate: Since we know volume is the problem, this could be because the market size has contracted or that the client's market share has been reduced. Do we know how each of these have changed?
Interviewer: The market size has remained the same, but our market share has reduced.
3. How do they differ?
To understand how typical non-hypothesis driven and hypothesis driven approaches differ, we will compare them as to their impact on both analysis and communication, breaking down different factors within each of these areas.
Ostensibly, the approach taken by the first candidate might have seemed sensible enough. However, as we contrast the two approaches, we will see that this naive, non-hypothesis driven approach is inherently and critically flawed.
The first candidate basically generates a laundry list of questions for the interviewer. Of course, some of these queries might be relevant, but the candidate has done nothing to place them within a wider structure. This makes it unclear how their questions are relevant to one another and to the client's main issue.
As a result, even if the interviewer actually gives answers to this whole slew of questions (they might well not...), it will be far from clear as to how those answers relate to one another or feed into any single conclusion. For instance, say that the interviewer responds to the questions above by telling us that a larger decline in sales had been seen in the over 50s and that there was a price war with rival companies over certain product lines. These isolated facts do not actually tell us very much in and of themselves and certainly don't begin to suggest how we might begin to address the client's problem.
By contrast, the second candidate, taking an hypothesis driven approach, has been very effective in structuring the problem. They work from indentifying two initial factors (price and volume) to drill down and eliminate irrelevant elements as they go. The structure they impose on the problem means that it is clear how each of their questions is ultimately relevant to the client's problem.
The main underlying reason why the questions asked by the first candidate are so inefficient is that they are not addressing MECE subjects (if you are not familiar with the MECE concept yet, see our article here).
Their questions ask about overlapping subjects (are not mutually exclusive) and are also unlikely to cover all factors which might be relevant to the case (not collectively exhaustive). The overall result is that, even if the candidate is given answers to all their queries, the information they end up with will likely be both ambiguous and incomplete and they might out on some possibilities altogether.
For example, the first two questions - asking about changing customer preferences and the behaviours of certain segments - overlap in terms of what they are referring to. If, say, over 50s are looking elsewhere for healthier options, this will be reflected in the answers to both. The third question asks about price wars, but price is only one aspect of competition between firms. Customers might well be favouring competitors due to geographical convenience or better customer service rather than just due to price.
The more structured nature of the hypothesis driven approach taken by the second candidate means that their questions are indeed MECE. Thus, questions refer to separate subjects and, between them, cover all possibilities at a given level of analysis. This ultimately allows the candidate to gain more useful information from fewer questions whilst not missing any important possibilities.
3.1.3. Answering the question
The most crucial difference between the two approaches is in whether they actually answer the question. The first candidate does not really get to grips with the actual problem and we can expect that they will fail to provide a single, clear answer for the supermarket CEO. By contrast, the second candidate makes use of their hypothesis driven approach to quickly narrow down the problem area and we can expect that they will quickly be able to deliver a decisive client recommendation.
You might think it is unlikely that you would fail to actually answer the question in a real interview. Surely, you are smarter than that, right? In fact, not giving a single clear solution is a major stumbling block for a large number of perfectly clever candidates, who often tend to conduct a broad analysis of the scenario generally whilst failing to relate this back to the client's concern.
This is especially common for candidates using traditional "framework" based case systems, where the generic scheme applied might not fully capture the case or address the client issue. You can see an example of this in our article on Problem Driven Structures, where we see how Victor Cheng's Mergers and Acquisitions framework fails to actually generate a solution when applied to a case.
The hypothesis driven approach is characterised by a constant focus on the main question. As we have seen above, all of the candidate's questions and analysis are ultimately driven by, and feed back into, addressing the client's main concern. Thus, the way an hypothesis driven approach structures analysis means that it is almost impossible to avoid arriving at a solution.
Your interviewer will be looking for an hypothesis driven approach to problem solving as a sign that you have the correct mindset for consulting. By taking an hypothesis driven approach, the second candidate demonstrates their ability to structure problems logically, to ask precise and relevant questions and generally to work in an efficient, focussed manner.
By contrast, the first candidate might have demonstrated a little general business knowledge, but fails to display any of these other traits. Rather, they will come off as over-eager and lacking reflective thought. The interviewer will find the barrage of queries associated with the standard approach vexing and the candidate risks appearing as if they are simply trying fish for answers from the interviewer, rather than being willing to actually analyse the problem for themselves. The second candidate asks questions more sparingly and gives a strong rationale for each in turn. This is precisely what the interviewer wants to see.
The first candidate's disorganised approach to asking questions and lack of focus or clear direction risks losing the interviewer - who will not be able to follow their train of thought.
A candidate working on too many lines of reasoning at once will risk confusing not only the interviewer, but also themselves. Candidates who do not take a sufficiently hypothesis driven approach often become muddled or get bogged down in some tangential aspect of analysis. In terms of communication, this will manifest in candidates repeating themselves and jumping around between topics when talking to the interviewer, with a general lack of clear direction.
The second candidate's analysis is much more sophisticated. However, the more structured approach - particularly the fact that only one issue is being dealt with at a time - makes it much easier for the interviewer to follow. This also helps to prevent the candidate themselves getting muddled or sidetracked.
On a pragmatic note, an interviewer might give a promising candidate a nudge in the right direction if they see that their analysis has overlooked something important or risks getting stuck in a dead end.
The fact that the hypothesis driven approach of the second candidate makes their rationale so clear means that it will be easy for the interviewer to intervene in this manner. For the first candidate, the interviewer will probably not be able to be obliging. A helpful nudge on its own could make the difference between getting a job and being rejected - and it won't be possible if your interviewer has no idea what you are up to!
4. Next Steps
All of this has shown just how crucial it is to take an hypothesis driven approach to your case interview. The best way to ensure that you approach cases in an hypothesis driven manner is to use our priority driven structure case method. This highly structured case solving technique builds in an hypothesis driven approach at its core.
To be able to properly make use of an hypothesis driven approach to cases, you will need a good grasp of how to segment and the MECE principle that underlies valid segmentations and ultimately allows for problems to be structured in a hypothesis driven fashion.
We have focussed on the case interview here, but many of the same issues we have discussed around taking a structured, focussed approach will be just as relevant to your fit interview. Remember - the different portions of a consulting interview are really different ways for the interviewer to look for the same general capabilities and mindset. You can find out more about how to take a structured approach to your fit interview in our article here.
Find out more in our case interview course
Ditch outdated guides and misleading frameworks and join the MCC Academy, the first comprehensive case interview course that teaches you how consultants approach case studies.