Creating simpler Anaplan solutions for end users can mean adding more...this is what happened for one of our recent clients.

The model design combined two separate modelling processes into a single core data flow in an attempt to keep the model simple. It didn't work. This is how we simplified by adding more.

Where modelling processes share common characteristics but differ in dimensionality it can be tempting to fudge a solution which tries to capture everything in a single modelling process. This is bad model design as the data flow needs to be forced in order for it to work.

The first process required dimensions A and B while the second required an additional dimension C. By combining the two we must force the first process using some kind default from dimension C as this dimension is not required for this process.

Users needed to select this default from dimension C every time they entered data against process one.

Neither intuitive nor simple.

Therefore, we decided to separate them and create two parallel processes where users could enter data naturally against all required dimensions without having to preselect a default when accounting for dimension C.

We then consolidated the two processes, mapping data entered in process one against the default dimension C.

This resulted in us having to add more modelling.

Why did adding more simplify?

Dimension C was simply not relevant to process one and therefore should not have been present.

The initial design attempted to build a process using a single modelling cohort where it required two. As a result errors were common across process one as users entered data against dimension C items which were incompatible just because they forgot to select the default item prior to data entry.

Recognising there were two distinct modelling cohorts; A + B and A + B + C we simplified the data flow and created a more intuitive end user experience.

We simplified by adding more.

But what we added broke a very clunky and poorly designed modelling structure into two shorter and far more direct data flows. By shortening the route from record to report we not only improved the end user experience but also the accuracy, precision and value of the forecast information the processes created.