The Economist Group
The Economist Group is an established global media brand with a portfolio of product and services whose mission is to provide expertise, insight and perspective to press the world forward. They take pride in believing strongly in free trade, openness and free thinking. It is their belief that these values drive progress and it is this philosophy that inspires them as they continue to expand into new markets, products and services.
Introduction
The organisation is complex. Multiple business units conduct their planning in a variety of different ways, across a varied portfolio of products and services and at varying levels of detail. It is the role of the ‘Group Reporting Model’ to bring all this data together to facilitate group level reporting and analysis.
The existing model has served them well. However, the model’s current design requires extensive maintenance and detailed validation after even the most minor of changes. Adding and updating dimensions such as new divisions, products or accounts requires drawn out reconciliation checks extending across multiple mapping tables, related hierarchies and report line items.
Recognising the challenges of maintaining and developing the model, combined with an ambition to expand Anaplan across the organisation the group financial, planning and analysis (FP&A) team partnered to Heathcote and Herran to carry out a model health check. The objective; to help them understand fully their current pain points and to identify the primary opportunities for development and optimisation.
What is an Anaplan model health check?
An Anaplan model health check is a systematic analysis of the key fundamental building blocks and structures which comprise of all Anaplan models. The analysis covers use and management of time settings, time frames and versions along with lists, functions and users. In addition to these structural elements it looks at the management of data imports, the role actions and the use of process. An assessment of model architecture, model design and use of established best practice, key functions and formulae rounds off the health check. A two hour de-brief with our principle consultant accompanied by a detailed presentation of the associated recommendations completes the analysis.
The primary challenge
The model health check identified the primary challenge was the absence of a strong overall design philosophy and efficient architecture. The Economist Group is structured across a variety of business units or divisions. While individual divisional processes operated well in isolation, allowing users to capture actuals data, input forecast and report divisional outputs there was no overall design solution that elegantly brought this all together into a group level reporting layer.
Actuals data mapped successfully into all key dimensions required for divisional reporting and divisional leaders were able to accurately update their forecasts and report their performance. However, subsequent downstream aggregation of this data relied heavily on mapping reporting outputs across P&L line items, acting here as a proxy for general ledger account, from one report into those of downstream group FP&A reports. This was a classic example of ‘daisy chaining’.
This design will operate successfully in the first instance. However, its design flaw will be exposed as soon as P&L line items are changed, removed or added; when new general ledger accounts, divisions or departments are created. Adding new products and services or updating the forecast logic of existing products and services requires extensive updates to mapping tables and the data flows.
Simple, intuitive and obvious updates become a burden to manage, requiring far too long to integrate and consumed an unreasonable degree of valuable resources to implement.
Change at in any stage within this chain is transmitted throughout the linked data flow resulting in often unintended and unwanted results. The outcome is a time consuming process of validation, reconciliation and remapping in orders to correct and re-establish the data flow.
The model becomes very fragile, unreliable and rigid as any changes results in a catalogue of checks and measure in order to rectify.
The flaw in this design comes primarily from the systematic removal of the general ledger account dimension once data has been mapped into divisional reporting. Line items are used to map into group reporting resulting in the loss of account level detail as P&L lines often aggregate data across a variety of general ledger accounts. Not only are group reports dependent on the outputs of reporting at the lower divisional layer the granularity of account level information is lost making any changes to how that data is displayed at the group level significantly limited.
The primary opportunity
The health check recognised this an ideal opportunity to optimise the ‘group reporting’ model and recommended the model be redesigned to include a consolidated data hub structure. This would involve creating a data hub for each division which in turn would feed into an overall group consolidation hub. By creating divisional data hubs where actuals and forecast data can be consolidated prior to feeding into divisional reporting each division will have a dedicated data source to serve all their divisional reporting needs. These divisional hubs will contain all the dimensionality required to serve the division leaders which is often more detailed than at the group level.
These consolidated hubs will retain all levels of dimensionality, including general ledger accounts and will feed directly into a group consolidation hub which would be used to populate group level reporting. The data feeds will be independent of any output and will map directly into a single group FP&A consolidated hub. This will in turn independently service all group level reporting irrespective of how each individual division chooses to map their data into their own reports.
This design separates the reporting requirements of each division with those of group FP&A. It removes the conditional nature of the data flows where the outputs of each divisional report, which has removed the general ledger account detail, feeds into group reporting. It is the conditional nature of the data flows which generates the vast majority of the current ongoing maintenance. This design will generate significant time benefits in addition to simplifying and optimising the model architecture by removing these dependent data flows.
The primary benefits
This design will allow The Economist group FP&A team to make changes to the structure of reports at all levels, add or remove list items all while being able to remain confident that unrelated data sets or reports will remain unchanged. Relationships between dimensions and line items will be simplified as the design fully leverages the platforms multidimensionality with mappings defined in a much reduced number of locations rather than across multiple tables. This design reduces the stages involved in going from record to report by eliminating dependent data flows, creating a more direct route from data to P&L by eliminating multiple mapping stages and generating a more reliable, robust and agile model.