How to ensure MECE?

Hi Levia, happy to answer your question again! 

This is a great question IMO and I think one of the biggest challenges to implement especially in a pressure interview situation where no common frameworks seem to fit. I would advise you to keep the following in mind:

The first step to becoming MECE is being aware. Stay conscious that your structure must be MECE, and double-check it in your head before you start laying it down for the interviewer. Of course everyone knows it in theory, but it’s important that you deploy this check in practice, as well.

IMO, the CE (Collectively Exhaustive) part is slightly more important for analyzing the case and getting to the solution. On the other hand, ME (Mutually Exhaustive) is super important for you to check your own structured-ness. It also helps to ensure that you don’t get confused between different branches of the issue tree.

The major issue with getting an effective MECE issue tree is meaningful classification – i.e. separating the problems into parts that are relevant to the case. For example, consider a situation where you need to break down the costs of a factory to suggest whether they need HR cutbacks. Now the traditional way of breaking down costs is “Fixed vs Variable” which is reliable for most cases. And while you can still use that classification – it seems like a better idea to classify costs by cost heads in this case - Material, Manpower, Maintenance, Miscellaneous, (4Ms? ;) ). The second classification is more useful in the context of the case. Even though the first is also MECE and might work ultimately, it might suggest that you’re sticking to a set formula.

Here are some handy ways of breaking down a problem in a MECE way:

1. Product method: This method is useful when you need to reduce a particular high-level factor into its smaller components. The idea is to start at the top and break it down into constituent units that can be multiplied (hence the name) with each other to arrive at the high-level factor. For example, say you have to calculate the annual fuel consumption of an airplane, in dollars, this is how you could break it down

Fuel consumption (in $) = Price of fuel * Fuel consumption (in kg)

Fuel consumption (in kg) = Number of km per year * fuel (in kg) needed per km

• Fuel in kg/km = Some assumption (or ask the interviewer for a value)
• Number of km per year = Number of km travelled per trip * Number of trips per year
  • This is followed by more calculations / assumptions for each factor

While this method is very useful for guesstimation-type problems, it can also help when you’re trying to analyze the root-cause of an issue, such has “The fuel costs of our airline client have risen 15% in the last two years, can you help us figure out why?”

2. Value chain method: I find this method quite useful for analyzing vague situations, when I don’t know much about the industry or there’s no clear framework to solve the business situation. The idea is to break down the situation into a time-based series of activities, from start to finish, that lead to the larger situation. A couple of examples:

- Analyzing the sales process of an insurance policy company to find inefficiencies: Take the point of view of the salesman and go through each of the steps he executes – starting with selecting the target audience and ending in collecting payment from a customer

- Identifying inefficiencies in a retail store – In this rather open question kind of scenario, you could put on the customer’s hat and track her journey from driving into the store, to checkout and driving out. Each step along the way might hold some hints about the processes used by the store that might cause inefficiencies)

3. The 'other' option: When you are not sure about other possibilities, tell the interviewer that you’ve clearly delineated the major possibilities but there are some more, which can be clubbed together in a “miscellaneous” / "other" bucket. However, use it only if you think you’ve covered the major factors already, or as a last resort. It should not look like an excuse or covering up on your part.

Hope these ways helped! Excited to hear what the experts on PrepLounge have to say on this interesting topic.

Levia, 

I see there are great answers below, but I wanted to add a couple of points.

1. MECE structuring is also a sense that comes with time. The more structures you do and the more feedback you receive, the better you get at it. Don't imagine that it's like an on-off button where you either have the skill or you don't. Even current consultants still receive feedback on how to improve their structures. 

2. There is no perfect MECE in reality. Especially in real client work, the situation is too complicated to be reduced to MECE. So MECE is more of an aspiration to ensure clarity of communication. In the interview, what matters more is coming up with a structure that is original and relevant, rather than one that is perfectly MECE.

Best,
Cristian 

Achyut's answer is already MECE :)

Just two brief hacks:

1. Generally, really important to order your points by importance (unless structure demands otherwise). Most important first.
2. Never ever say 1, 2, 3. Always say 1, 2, 3, and others (Achyut's point 3). I want to re-mention it, because it is really important to a) stay MECE, and b) it's the wild card if you're a bit off-track.

Quick examples: If you list regions for international expansion for a US company, it could be "Asia, Europe, ..., other". Or "Canada, Japan, China, ..., other". The ranking already tells a story which shows you're hypothesis-driven, and if it's not where you should go, you're still covered with the "other" category.

In one of my case interviews I had to structure the cost base for a steel manufacturer. Somehow, the "solution" was to cut IT costs; it took a while to get there, but thanks to the "other" category it was part of the original structure already.

Thanks for your detailed reply! I think both these methods (product method and value chain) can be helpful in an unstructured case as well in other regular cases. 

Waiting for others' views on this! 

Thanks,

Levia 

Dear Levia,

in general a good structure can be evaluated by a certain depth and breadth. The “depth” should be at least 3-4 levels while the “breadth” should cover the entire solution space. You can cross-check this with the MECE principles (For details see respective article on Preplounge), but the CE (collectively exhaustive) part is basically defining your breadth.

Finally, make sure to check for inter-linkages in your structure and point them out.

Simon