
One of the most fundamental concepts in analytics is known as "Mutually Exclusive and Collectively Exhaustive (MECE).
MECE is a property of categorizations, classifications and hierarchies. We say, "this categorization is MECE", or "that hieararchy is not MECE" to indicate whether summing all values at a given level will sum to 100%.
For instance, if I had three pets, a dog, a budge, and a budgee, the following categorization of my pets is non-MECE:
| Mammal | Lays Eggs |
Dog | 1 | 0 |
Budgie | 0 | 1 |
Platypus | 1 | 1 |
Total | 2 | 2 |
In contrast, the following alternative categorization is MECE:
| Mammal | Bird |
Dog | 1 | 0 |
Budgie | 0 | 1 |
Platypus | 1 | 0 |
Total | 2 | 1 |

Non-MECE categorizations can lead to inaccurate decision making in business contexts, because one cannot compare metrics at different levels of categorization. This issue can become harmful when organizations use non-MECE categorizations for goal setting.
In practice, non-MECE categorization appear more often that one might think. Specifically, when using multiple, separate flag-attributes as in the tables above. In contrast, when we use a single column or attribute to categorize items, the categorization will typically be MECE, because each field in that column can only have one specific value.

For instance, consider an organization setting goals to increase consumer sales and business sales respectively for the consumer and business sales units. If the organization does not enforce MECE categorization both units may meet their goals without increasing overall sales, simply by categorizing more customers as both "consumer" and "business". This non-MECE categorizations may appear "defensible" when we consider that consumers may sell items so that they could be considered a "business".