From our experience bad Anaplan model design results from limited understanding of mutually exclusive and collectively exhaustive data structures...

What is mutually exclusive and collectively exhaustive?

Mutual exclusive is a term applied to the grouping of data items whereby there is zero overlap between groups. Items belong exclusively to a single group.

Collectively exhaustive is when all possible items are accounted for in our classification in the relevant combination, group or association.

Why is mutually exclusive and collectively exhaustive (MECE) import to model design?

When we fully apply MECE to organise, classify and characterise our data we design clarity, transparency and remove redundancy from our modelling.

We are clear on how our data can be effectively structured, confident that we have captured everything and that our data flows and model processes are logical.

When we fail to structure our modelling this way we find that we must create exceptions, work arounds and overrides. Our models are less intuitive, the end user experience is confusing and our administrative burden is too high.

For example, when 99% of a mapping requires a single dimension and the remaining 1% is split across a second we can only create mutual exclusivity by including the two dimensions. This is because using the single dimension will create a situation in which 1% of items will map to more than one target. They would not be mutually exclusive in their relationship.

This also applies to list properties and attributes. Where a dimension such as Country contains multiple options for a single characteristic such as currency for the purpose of reporting then it is not mutually exclusive.

We overcome these by creating additional sub-categories. We add the second dimension to our mapping table and a currency type sub-category to our Country attributes such as reporting currency and local currency.

As we interrogate our data, processes and data flows we will tease out those situations which do not sit neatly in their current structures.

This is almost always because there is a conflict in how the data is characterised and catalogued, the groupings are not mutually exclusive. It is ticking more than one box - therefore add more boxes.