In this blog we will look at the components of a modern solution and why it matters.
At the heart of every Analytic application is a mathematically based business model. This model describes the organisation in terms of its relationships between:
- Income, which may come from goods or services for a commercial organisation or grants/donations for ‘not for profit’, vs external marketplace, i.e. who else is competing for those same resources
- Cost of producing those goods/services
- The cost and impact on income of how those goods/services are supported
The model will try and incorporate how those relationships have change with time and the impact of any external influences, e.g. inflation rates.
The aim of this model is two fold: First to challenge and then agree the internally held perceptions of how the organisation functions; and two, to use the model as a means of predicting likely future results given a number of anticipated scenarios. The model is never going to be perfect as the business world is far too complex and erratic to take into account every possible interaction.
For most organisations, this model will be ‘multi-dimensional’ as this provides the best way of setting up and maintaining the model. For those unfamiliar with the term, multi-dimensional means that each facet of the organisation such as its management structure, chart of accounts, product hierarchies and customer classifications are set up separately along with their relationships. The model then combines these facets using a set of rules which can be used to manipulate data to produce predicted results.
With the above in mind, there are 11 components that make up the ideal analytic solution:
Multi-user database: Data is stored in terms of the different business dimensions it represents, e.g. Sales revenue of £1500, in January 2017, for product X to customer y. Each item underlined is a separate business dimension – but all are required to fully define what the value 1500 represents.
Security: This provides for multiple people to access the data but whose actions are determined by their individual roles and in which periods of time.
Model builder: A simple user interface through which administrators can setup and make changes to the model dimensions.
Business rules: The ability to setup relationships between items in each business dimension using natural language so they make sense to those reading the model
Data acquisition: The ability to summarise and load data from a variety of sources into the model.
Application ‘intelligence: This allows the model to automatically calculate data without having to set up rules. E.g. converting any local current amount to a base currency, or the correct aggregation of data over time.
Data manipulation: Built in functions that provide the user with statistical/trend analysis, correlation, time-shift analysis, initiative / scenario planning
Presentation tools: That allow users to produce static / dynamic reports with charts, grids, drill-down, drill-through, and slice/dice/rotate capabilities
Data / User control: This provides workflow capabilities where users can be requested to enter data at set times and within defined parameters, e.g. to collect forecasts. Administrators are provided with an overview of the status and can chase up users who have not responded.
Collaboration: Provides a method for entering notes/comments to explain results, and a way of alerting users to exceptional items.
User access methods: Allows users to access the model from a variety of platforms including web, mobile, and spreadsheet access.