The functional database model is used to support analytics applications such as financial planning and performance management. The functional database model, or the functional model for short, is different from but complementary to the relational model. The functional model is also distinct from other similarly named concepts, including the DAPLEX functional database model and functional language databases.

The functional model is part of the online analytical processing (OLAP) category since it comprises multidimensional hierarchical consolidation. But it goes beyond OLAP by requiring a spreadsheet-like cell orientation, where cells can be input or calculated as functions of other cells. Also as in spreadsheets, it supports interactive calculations where the values of all dependent cells are automatically up to date whenever the value of a cell is changed.

Functional database model

Overview

Analytics, especially forward looking or prospective analytics requires interactive modeling, "what if", and experimentation of the kind that most business analysts do with spreadsheets. This interaction with the data is enabled by the spreadsheet’s cell orientation and its ability to let users define cells calculated as a function of other cells.

Functional database model

The relational database model has no such concepts and is thus very limited in the business performance modeling and interactivity it can support. Accordingly, relational-based analytics is almost exclusively restricted to historical data, which is static. This misses most of the strategic benefits of analytics, which come from interactively constructing views of the future.

Functional database model

The functional model is based on multidimensional arrays, or "cubes", of cells that, as in a spreadsheet, can be either externally input, or calculated in terms of other cells. Such cubes are constructed using dimensions which correspond to hierarchically organized sets of real entities such as products, geographies, time, etc. A cube can be seen as a function over the cartesian product of the dimensions. I.e., it assigns a value to each cell, which is identified by an n-tuple of dimension elements; thus the name "functional". The model retains the flexibility and potential for interactivity of spreadsheets, as well as the multidimensional hierarchical consolidations of relational-based OLAP tools. At the same time, the functional model overcomes the limitations of both the relational database model and classical spreadsheets.

Products that implement the principles of the functional model to varying degrees have been in existence for some time, including products such as Essbase, TM1, Alea, Microsoft Analysis Services, etc.

Analytics context

The management system of an enterprise generally consists of a series of interconnected control loops. Each loop starts by developing a plan, the plan is then executed, and the results are reviewed and compared against the plan. Based on those results, and a new assessment of what the future holds, a new plan is developed and the process is repeated. The three components of the control loop, planning, execution and assessment, have different time perspectives. Planning looks at the future, execution looks at the present and review looks at the past.

Information Technology (IT) plays now a central role in making management control loops more efficient and effective. Operational computer systems are concerned with execution while analytic computer systems, or simply Analytics, are used to improve planning and assessment. The information needs of each component are different. Operational systems are typically concerned with recording transactions and keeping track of the current state of the business – inventory, work in progress etc. Analytics has two principal components: forward-looking or prospective analytics, which applies to planning, and backward looking or retrospective analytics, which applies to assessment.

In retrospective analytics, transactions resulting from operations are boiled down and accumulated into arrays of cells. These cells are identified by as many dimensions as are relevant to the business: time, product, customer, account, region, etc. The cells are typically arrayed in cubes that form the basis for retrospective analyses such as comparing actual performance to plan. This is the main realm of OLAP systems. Prospective analytics develops similar cubes of data but for future time periods. The development of prospective data is typically the result of human input or mathematical models that are driven and controlled through user interaction.