This article is part of a series on Polaris best practices. Click here to learn more.
Natural dimensionality defined
Natural dimensionality can be defined as the inherent and most appropriate number of dimensions (aka “lists”) needed to accurately represent a dataset. Most business problems are inherently multi-dimensional, yet the resulting datasets are usually largely unpopulated. Polaris allows for modeling at the natural dimensionality, without regard to that unpopulated space because of how default values are handled.
Remember, Polaris only requires memory for populated cells in a model.
An example of natural dimensionality
The best decisions are made using data and models that accurately reflect the natural dimensionality of your business. Polaris excels at handling calculations over data with a high degree of dimensionality. Think about a manufacturer that uses 15,000 component parts and produces 1,500 products. Each of those products will likely use a small subset of those component parts, leaving the rest of the dataset largely empty.
Scoping your Polaris models to reflect the natural dimensions of your business and align with how your organization plans will reduce complexity, allowing end users to easily filter to exactly what they need to see and provide a more intuitive planning experience.
What does it look like to design to the natural dimensions of your business? Let's look at an employee planning example.
In this example, we have our employee dimension showing on our rows (1). We have some attribute information about their office and their title included in the module (2). And we have some time-based information that’s telling us what the compensation is for each of these employees (3).
Across the top of the page, we can see that we have an organizations list (whose top level is “Total Company”), a region list and an employees list. By breaking up our dimensions this way users can easily filter down to what they want to see. If they want to see Org 1, they can do so.
The grids will update, showing the employees that fit in Org 1 for all regions, while giving us a dynamic subtotal at the bottom.
Perhaps they also want to see just the US employees. They can select from the “All regions” dropdown and see only Org 1 employees in the US region.
Users can select any combination of the two lists in the example, see the dynamic total and input the plans for their employees. The natural dimensional structure of this solution allows for this type of user interaction and avoids having to do complex user filters to get to the selections that your user might want to see.
Conclusion
It’s important to think about and talk through the natural dimension of your business process with your team as it is possible to over dimensionalize a line item, potentially creating slow calculation times. Remember our definition of natural dimensionality: you’re looking for the inherent, but also the most appropriate number of dimensions for your business process. Just because you can add a large number of dimensions, doesn’t mean you should.
Video
This concept is also explained in the following video:
……………..
Authors:
Anaplan’s Theresa Reid (@TheresaR), Architecture and Performance Director; Stephen Rituper (@Stephen), Sr. Director, OEG; and Terry Archsmith (@TerryA), Sr. Director, Platform Training.
Authors:
Anaplan’s Theresa Reid (
@TheresaR
), Architecture and Performance Director; Stephen Rituper (
@Stephen
), Sr. Director, OEG; and Terry Archsmith (
@TerryA
), Sr. Director, Platform Training.
