Struggling with the suitability of an hypothesis driven approach in some particular cases

Question

Dear follow case enthousiast,

I have difficulties with the hypothesis approach in some perticular cases, I do understand how to have an hypothesis approach on a profitability framework (i.e. my hypothesis is that revenue a driving a decline in profit, that you then try to disprove).

But in some case where the task is more open (e.g. case with "how would you"), I find it difficult to lay out an hypothesis. For example, what would be you approach in the following case: "How would advise a wealthy philanthropist in setting up a NGO"? and assuming the question is more centered about 2 areas 1) where and for what action to invest money (e.g. eduaciton in SE-Asia, Access to basic health in sub-saharan countries) 2) how to get financial and non financial support. For me a natural approach/issue tree would me more to have a list of qualitative factors and somehow weight them to get to a recommendation. However, I'm curioius to hear how you would solve this question using an hypothesis approach.

Thanks,
Ben

Answer 1

Dear Ben,

As you mentioned, in some cases you can easily communicate your hypothesis and which ones are you prioritizing in your analysis, after you set your initial structure. However, there are cases (like the NGO example you wrote) in which the hypothesis are indirectly mentioned in the structure.

For instance, following your example, imagine that you already ask the relevant question to understand about the objective, the philanthropist profile, etc. and the first criterium in your structure is "Selection of the country/city". You would probably have branches like investment/budget, impact on the population, impact on the economy, expertise, etc.

So you start by guiding the interviewer "To help the philanthropist I would analyze X criteria 1) Selection of the country/city 2) .. 3) ... 4) ... 5).... In the first one, I would analyze where the NGO can have the biggest impact considering 4 factors: 1) Amount of investment required 2) Impact on the population (e.g. literacy, death rate, nutrition in children, etc.) 3) Impact on the economy (jobs generated, corruption, delinquency, etc.) 4) How much expertise does the philanthropist have to solve this problem. (AND HERE COMES YOUR HYPOTHESIS). I think that a potential location for the highest impact is Poverty in South America, because X,Y,Z. However, I would like to probe this hypothesis with hard data. In the second one ......"

To sum up, in this type of cases the hypothesis are embedded in the structure and throught the case and is not as straightforward as profitability cases, turnaround cases, or any type of cases in which you are pinpointing a problem.

Regards,

Hugo

Answer 2

Hi Ben,

some cases may indeed appear difficult to structure in ways different from a qualitative list. Here are my suggestions to identify a structure that could potentially work for cases of this type where you could formulate a hypothesis, using the NGO example:

Clarify the objective. What is our specific goal in terms of setting the NGO? Do we want to reach a certain average income for a certain number of individuals? Or maybe literacy level?
Identify available options. Do we have a short list of countries/areas to cover and/or of fundraising options? If not, based on the objective, you will need to brainstorm available options.
Identify the variables the objective would depend on and you can have an impact on. In case you are looking for average income, you could look at employment level; in case you look at literacy, number of teachers; etc.
Formulate a hypothesis on the best option.
Analyse the expected outcome of each option in terms of the variables you could have an impact on, in order to verify your hypothesis. Given a certain amount of investment, how much would employment, and therefore average income, grow in South Asia compared to Africa?

As usual, the hypothesis step could be useful, but not strictly necessary to reach the right conclusion, as you would always have to double check the actual data at the end.

Hope this helps,

Francesco