We build robust, reliable and scalable FP&A models in Anaplan because core to our model architectural designs is a concept we call sequential consolidation.
Process map describing how sequential consolidation creates a more robust core consolidation architecture
It is central to all our designs and relies on moving forecast outcome data through several transformational stages before mapping it into a final report.
An FP&A model will contain a core set of reporting dimensions ranging from general ledger account, cost centre, product type/category, geography/region etc. The list of options can be very unique to any given model, organisation or business unit.
However, forecast outcome data does not always align to these core modelling dimensions. Therefore, the data must be moved through a sequence of transformations to ensure it is ready to be reported on.
Forecast outcome data is grouped into cohorts based on their share of common dimensions. These are the modelling dimensions used in their calculation and present in the final version of any particular forecasting process. These cohorts move through each step from forecast outcome data to the penultimate stage just prior to being mapped into a report.
Each cohort is supported by a mapping module where mappings are assigned either directly or via other input modules.
Where core reporting dimensions are missing from a cohort they are mapped in using the dedicated mapping modules. Irrelevant dimensions are removed.
Forecast data is translated from local to reporting currency (LCON to RCON) if required and additional dimensions are added where these will enhance the user experience in the consumption of the final reporting (REP01).
This approach ensure that the consolidation remains modular. The supporting structures that move each cohort through the model are identical and can be easily reproduced when new cohorts are created. A new cohort is required whenever a new calculation engine is added to the model whose outcome data is dimensioned using a new yet unused set of dimensions.
As new cohorts can be easily integrated the architecture is also scalable and can expand as the business grows and new modelling is added to the consolidation.
By breaking down and extending the consolidation across multiple cohorts and over several stages the model remains not only effective as data is reliably moved from one stage to the next but also very efficient.