Model Allocation Capabilities

Question

Hey everyone,

I'm doing some strategy work around optimal allocation strategies and and was wondering what your experiences are.

What is an acceptable row count to have pushed through the model (Pre - Post) that still offers high performing? Other than reducing granularity, what have you been able to do to imrpove model performance?

#ApptioforAll

Accepted Answer

Hello Mac - @Jenny Franklin is spot on!

I do have some additional information for optimal allocations.

Apptio best practice is to keep allocation rows under 500k. There are many reasons for this.
- Allocation lines without data relationships and/or over 500k are problematic. Calculations based on source row count X destination row count. Even if you have a filter on the to and from it does not affect the calculations across rows. The 2 ways to affect the calculation on rows is granularity or data relationships. Granularity can play a huge factor here but depending on the object, other allocation lines and reporting needs objects are not always as simple as reducing the granularity. In situations like where you cannot change the granularity, I would would introduces keys to help consolidate the counts wherever possible.
- Reports and drill reports can cause huge calculation spikes because Stage and Production are pre calculated. If you have an allocation of over 500K and a drill through that specific allocation line it will drastically increase your calculation time and cause the report to load incredibility slow in dev or to changes made in the model or report. Especially depending on the amount of time a project has open.
  - If you have 3 years open and 1 drill path is taking 1 minute to calculate per month. It is taking 36 minutes for 1 report. Not including any other reports or models.
- Instead of having 1 really large allocation that is many to many. Try to break down into multiple allocation lines.
If Statements with nested lookups can cause huge performance issues.
Modeled Metrics - It is easy to only think of Cost when we are troubleshooting model performance. Depending on a client and the components installed there could be several modeled metrics. There could be an OOTB allocation line lighting up causing a crazy allocation. Ensuring all allocations for different modeled metrics is key for model performance. Attached is a picture of the reference project and OOTB Metrics.
Data Refresh Cycles - This one might be obvious, if key data sets are set to Ah Hoc but are used monthly in allocations Apptio will see it as data and still run the data. Even if it is for months in the future for the current time period open.

Hope this helps! Curious to what other folks have experienced!

Renee
#ApptioforAll

Apptio for All