We designed a scenario planning function in our clients Anaplan for FP&A model which reduced model size by two thirds. This is how we did it...

As a fast growing business their strategy is in a constant state of flux. Therefore, they wanted to be able to model multiple forecast within their current model design.

But as soon as they added new versions their workspace rejected the change as the additional modelling resources resulted in them them exceed their workspace allowance.

Initially we analysed and mapped out the existing model. We then reviewed the month end process with their FP&A lead to discover how the existing design aligned to their current ways of working.

What we discovered lead to the design of our signature archiving model.

Our design research revealed that once our client had completed one version of a forecast this was then locked down and secured not be amended once it had been signed off. This discovery contradicted the existing design which stored signed-off and locked down forecasts as parallel models to that which was current and live.

This meant that all stored forecasts consumed the same amount of modelling resources as the live model despite not being subject to any calculation changes or amendments.

As a result we devised a methodology which involved creating an archive; an isolated but fully coded storage area within the model where historical and signed off forecast data can be posted. This data could then be mapped downstream reporting as required.

We were then able to remove all reference to historical or other versions peppered throughout the calculation and consolidation modelling.

This resulted in a reduction in model size of at least 67% and a significant increase in model speed and performance.

These are the steps -

1 - Identify all forecast modelling outputs.

2 - Add an 'in use?' test for each output (removing time and native versions).

3 - Query each 'in use?' test. Create a combination list coded to include every possible 'in use' combination of forecast dimensions.

4 - Map forecast outputs into a live archive data module dimensioned by our new archive combination list.

5 - 'Copy and paste' this live archive data into a new identical archive module. This is our isolated archive.

6 - Create a archive list systems containing fully mapped properties. Use this to map our archive data into all downstream reports.

This was a game changer for how we now design scenario planning for our clients.

It also now takes centre stage in our new signature foundation model -LaunchPad, our introductory FP&A model designed to get our clients up and running in Anaplan in record time.