why is PCM limited to only 5 Cost Objects, How can we increase it?

[Former Member]

PCM seems a bit limited in the number of input dimensions on 3 main input dimensions and 5 main output dimensions. I've have cost inputs with 10 dimensions and outputs with 20 dimensions but I want to model all in PCM. How can I do this?

[Former Member · Accepted solution]

Hi Martha,

What do you have as the 10 Cost Object inputs and 20 outputs? Understanding these would help position them in the application?

Regards,

Tony Teixeira

[former_member783461]

Hi Martha,

We've build models that simulated up to 27 cost object dimensions, both for reporting and for actual cost allocation modeling.

There's some nuanced rule writing that needs to happen to allow cost assignments to work within the extra dimensions.  The core concept at the crux of making this work is to create leaf-level elements of the Cost Object dimensions that are very low-level (think SKU-level, for a product dimension).  These can thereby be rolled up via many different hierarchies, using native PCM alternate hierarchy functionality.

The key to making everything work from a math standpoint it to create a set of rules that allow drivers to be input at the rollup level in the alternate hierarchies.  There's a chance that I might have some old rules that could be modified to hit the mark for you on this front.  It's a pretty-involved overall solution, but we've made it work across a number of different models.  There are 2 different types of higher-level driver inputs that can be entered as higher-level driver data: Weight Factors and Cascades.

Weight Factors:

Imagine a bunch of SKUs, and and a need to assign data based on a Weighting applied to SKU-level driver data.  The weighting can be entered at the level at which you know it, and then applied to the SKUs beneath that level.

Internal Cascades:

Input driver data at a rollup point in a cost object hierarchy and have the driver data cascade down the leaf members beneath using a secondary driver via a rule.

[Former Member]

Thank you Michael.

We are using up to 35 dimensions in our source system we want all of these but selected 20 that we must have. These are output dimensions and we have activity drivers that 20 dimensions.

Your solution sounds good but does it work with activity drivers that are 20 dimensions? And need to allocate cost on all dimensions in one step, will it work?

We looked at cross models rules, Bill of Material models, using up to 4 models each sharing at least one output dimension with another. None of these were exactly what we needed and each had their own problems.

Then we come to input dimensions we require 10 dimensions minimum. How to get these input dimensions in the model also?

We are thinking to use BPC for these allocation now as there are no issues with dimensions there.

Thanks

Martha

[former_member783461]

Hi Martha,

I'd like to have a quick conversation on this matter, if it's of interest to you.  I have a working approach to allocating data starting with 10-20 dimensions into results that span more than that.  I'm totally with you on the challenges of doing this in PCM directly.

Ping me on email or Linked In, perhaps?

Thanks,

Mike