Cost-Reduction in Business Intelligence- Governance & Standards

Establishing 'Making it Happen' as a 'Formal & Predictable' Discipline

Principles and Rules Listing Page
Field Tips Series- Streamlining & Cost-Reduction in Business Intelligence- Governance & Standards
Governance & Standards essentially points to the consistency and robustness of managing change in the BI environment. The key issue behind high cost of BI is varied IT platform, standards and processed, which have additional over-head and also make BI inefficient & defective. As an organization we can undergo different types of changes, which have to dealt by effectve governance and standards.

This page of 'Principles and Rules' is linked to: BI business intelligence end-to-end view, Data Analysis/OLAP, BI platform Tools Evaluation, Data Warehousing, Data Quality,

Different Classifications of changes in a BI environment

Different Level of BI Changes:

As new BI requirements emerge in an organization across functions and locations, it may lead to:

New BI platforms in terms of software tools.
New Data-Marts in the same platforms.
Changed dimensional models (dimensions, attributes and facts) in the same data-marts.
New Reports and analytics on the same data-mart structure.
New BI ETL scripts due to changes in the Dimensional Model (the way data is modeled in Data-Warehouse- it can also be relational model) in Data Warehouse and OLAP .
Changes in the formulae and business rules.
Etc.

Universe and context of BI changes:

BI platform structuring and re-structuring: This is more fundamental change, where BI platform foundation is being set. In this context, one is guided (and not driven) by the business requirements, as one establishes the holistic dimensional models. For example, if you are creating a data-mart on sales revenue analysis, you will include all or most aspects of sales revenue management, instead of limiting it to business requirements (which might be driven by some immediate burning issue).
BI platform guided by multi-functional- enterprise level changes : These kind of changes include business requirements management and delivery across multiple functions.
Functional level change in BI requirements.
A process level change (within the function or across the functions)

The above bullets are essentially to sensitize the readers on the variety of changes, which are possible. The question is on how to manage the change within the given change environment, which is quite varied.

Governance Methods- Less than Ideal but a good starting point.

Document the standards and circulate. Establish a central monitoring process post-facto.
Document the standards and circulate. Establish a central monitoring process pre-facto. This means that before implementing the designs, one has to take sign-off from the central monitoring group. This is a better solution, but its important for the central monitoring group to have needed sponsorship and teeth..

Governance methods- Highly Recommended

Make IT responsible for any BI platform acquisition, and for contractual agreements with the software vendor.
Post acquisition, make IT responsible for supporting the IT platform.
Create a BI council, which works together to develop the standards and governance methods. This BI council will have representations from different functions and therefore, will have a greater buy-in.
The most ideal method is to have the BI platform and design management done by a single and central group, which has end-to-end functional and skills representation. (please refer to Business Intelligence Competency Centre in 'latest trends')

The Business Intelligence standards

BI Data-Modeling standards:

The Foundation Dimensions: The common set of dimensions to be used by any dimensional model in any data-mart.
The way to store the BI data-model: De-normalized way (like dimensional model) or normalized way (Relational Model)
The naming and labeling conventions of different data-elements.
Data-Element definition- Formulae related to different terms and associated business rules

NOTE- Above falls in the gamut of the subject of Meta-Data repository. A meta-data repository is a comprehensive documentation of organizational universe. The above mentioned elements are important part from BI perspective.

BI Process Standards:

BI strategy formulation and change management process
BI Business requirements management process.
Once BI Business Requirements are defined, end-to-end Change management process.
BI security administration process.

BI Software tool acquisition and maintenance standards:

Set of tools to be used for BI. (Data Quality, Data integration, Data-Warehouse Database, OLAP Server, Enterprise reporting tools)
Software tool acquisition process.
Licensing regime.
Maintenance and Service contracts terms and conditions.
A list of software development vendors, who can work for the organization. For example, An organization would not like to have 20 different vendors working on their BI platforms.