BMG Rights Management
Introduction
BMG Rights Management is an international music company combining the activities of a music publisher and record label to offer its artists a range of services that are often not provided by its competitors. This is unique in the industry and sets them apart, marking them out as a music company which is striving to do it differently.
Their portfolio of artists whose rights they manage range from Kylie Minogue, Bruno Mars, Robbie Williams and Kurt Cobain to 5 Seconds of Summer, Cat Stevens, and Mike Jagger.
With a vision to be a new kind of music company they push the boundaries of the industry, strive to be transparent and fair, to be a service to their artists and songwriters and to support them in their careers while also helping them to maximise their earnings.
Their values directly align with ours and it was a delight to onboard them as new Heathcote and Herran clients.
What is an Anaplan model health check?
Recognising the need to reengage with their existing Anaplan model after a period in which the business was reassessing its planning processes and technology requirements, BMG approached Heathcote and Herran for help in better understanding the health of its existing Anaplan platform.
An Anaplan model health check is a systematic analysis of the key fundamental building blocks and structures which make up 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 processes.
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 principal consultant accompanied by a detailed presentation of the associated recommendations completes the analysis.
The greatest opportunities
List design
Effective list management principles ensures that the defining structures of any Anaplan model can be updated, scaled, and expanded when needed with minimal disruption.
The existing model builds its lists via a series of actions which imports data from a central data hub. Where Anaplan models are extended across the business this approach is always recommended to retain continuity between key model structures. However, when done in isolation this approach can make the model inflexible. Updating how its array of lists and dimensions are managed offered BMG an opportunity to significantly improve the model’s overall design and ease of use.
We recommended in our report the use of flat master lists which exist independently of others and do not form any part of any structured hierarchy. These master lists are used to inform the construction of what we refer to as parallel hierarchies which may only exist in a single model and are used to better inform the specific planning activities the model is designed to facilitate. These parallel hierarchies contain identical, ‘leaf’ level or terminal lists but aggregate or sum up into different parent and grand parent structures. The result is a series of different hierarchy structures bound by identical leaf level items but contrasted by the varying routes up their own respective hierarchies.
As an example, a flat products list acts as an independent master list. Multiple parallel hierarchies are then created from this list. One may aggregate products into stores, regions and countries while another sums into warehouses and another into customers. The result is a series of shallow hierarchies all having a common leaf level list but progressing up their respective hierarchies into different and varying ancestors.
These parallel hierarchies are mapped across each other at this leaf level allowing modellers to query common data across a wider variety of dimensions and variables.
A common design flaw is to create over extended hierarchies over an excessive number of levels. This is typically an attempt to capture every iteration of every relationship in a single structure. Examples extend upwards of ten layers, getting more obscure and removed from the primary intent of the hierarchy.
The result is a structure which is extremely difficult to manage due to its rigidity and lack of scalability.
Data model
Efficient data management is focused on importing data into Anaplan in its flattest form, free of dimensionality and encoded via a sequence of columns used to house metadata properties. Data imported this way keeps sparsity, the allocation of model resources to empty intersections of dimensions to a minimum. Model builders can model the data across any chosen combination of dimensions using Anaplan aggregation functions such as SUM, ANY and ALL.
BMG’s current model imports data directly into modules having several dimensions. This creates unnecessary sparsity as most of the list combinations do not hold relevant data. It also presents a significant barrier to mapping the data into other dimensions. The dimensioned module locks the data into a defined structure requiring more modules and imports to be created if any other sets of dimensions are needed for new or expanded functionality.
We recommended importing data using a flat structure either directly into a module or into list properties. This method keeps sparsity to a minimum while also giving the modeller the ability to query the data across as many or as few dimensions as is required for the modelling challenge being solved. This approach will not only ensure the model is kept as efficient as possible it also ensures that the data model can scale appropriately with the organisation. The addition of extra data dimensions in the data only requires the addition of extra columns rather than the redesign and rebuild of new data structures as existing data has been locked into the existing model structure.
Data flows
A robust, logical, and unidirectional model design eases effective reporting. Data flows in a single direction from record to report while similar logic steps are grouped together and kept independent of each other. Inputs, calculations, and outputs are separate and sequenced in a logical design. The data flow can be audited effectively, and troubleshooting can be carried out efficiently.
While in part BMG’s existing model follows these model design principles there are key instances where better design would create significant model improvements. Some key data flows are bidirectional with outputs feeding both upstream and downstream calculations. These calculations which drive outputs for multiple variables are then found in the same modules with some also being used for both reporting and calculations.
We recommended a review and redesign of these key parts of the model. We suggested the breakup those structures which are used to calculate multiple variables and in turn ensure that data flows in a single direction in a logical sequence. These changes would culminate in a forecast consolidation holding all relevant dimensionality required to service the model’s reporting requirements.
The approach would ensure that the route from record to report is clearly delineated along the data flows defined by the key forecast drivers, that changes can be made independently and with minimal risk of disrupting other calculations. The model would be better prepared to scale and grow with the business, while also still being agile and responsive to small updates, changes, and inevitable tweaks in calculation logic.
We recommended a variety of enhancements that would significantly improve the durability, agility and responsiveness of BMG’s current Anaplan model.
Using flat master lists in addition to a series of shallow parallel hierarchies would allow for modelling across a wider range of dimensions while also ensuring the management of lists and hierarchies is more intuitive, reliable and scalable.
Data is the life blood of all models. Therefore, it should be an asset and allow the model to build insights that are locked up within it. By importing data into the model flat and removing the limitations placed on building the data model within a defined hierarchal structure the model is best able to access these.
Designing a more direct and logical route from record to report and delineating the calculations for all key variables will also allow the model to scale more reliably as the organisation grows and new requirements are discovered.