Chapter 11 – 11.2 – The Business Intelligence Perspective – Part 2/3

11.2.3 Methodologies and Approaches

.1 Methodologies

There are no formalized business intelligence methodologies that impact the responsibilities and activities of the business analyst. However, a business intelligence initiative can operate within or alongside methodologies applicable to other disciplines or perspectives which themselves might impact the business analysis role.

.2 Approaches

Within the business intelligence framework there are a number of less formal and potentially overlapping approaches that map to particular business and technical contexts.

Types of Analytics

There are three types of data analytics that represent incremental solutions, with increasing levels of systems complexity, cost, and value:

• Descriptive analytics: uses historical data to understand and analyze past business performance. Business information can be categorized and consolidated to best suit the stakeholder’s view including executive management dashboards, middle level management key performance indicator (KPI) scorecards, and operational level management charts. No assumptions are made as to which situations are of interest to the stakeholders, what decisions need to be made, or what actions might be carried out. The business analysis focus is on the information and communication requirements for standard reporting and dashboards, ad hoc reporting, and query functionality.
• Predictive analytics: applies statistical analysis methods to historical data to identify patterns, and then uses that understanding of relationships and trends to make predictions about future events. The particular situations that are of interest to the stakeholders are specified, and their business rules are defined. The business analysis focus is on the information requirements for pattern recognition through data mining, predictive modelling, forecasting, and condition-driven alerts.
• Prescriptive analytics: expands on predictive analytics to identify decisions to be made and to initiate appropriate action to improve business performance. Statistical optimization and simulation techniques can be used to determine the best solution or outcome among various choices. For situations of interest to stakeholders, full specification of the associated decisions and potential actions are required. The business analysis focus is on the business objectives, constraints criteria, and the business rules that underpin the decision-making process.

Supply and Demand Driven

The objectives and priorities of a business intelligence initiative can be based on the technical goals of improving existing information delivery systems (supply-driven) or on the business goals of providing the appropriate information to improve decision-making processes (demand-driven):

• Supply-driven: assumes the view of “for a given cost, what value can we deliver?”. This approach maps existing systems data to define what data is available. A common implementation strategy would be to:
  1. phase the inclusion of existing databases into the business intelligence solution architecture,
  2. progressively replace or repair existing outputs, and
  3. explore new insights that might be gained from the consolidated data.
• Demand-driven: assumes the view of “for a given value, what cost do we incur?”. This approach starts with identifying the information output needed to support business decisions, and then tracing that information back to the underlying data sources to determine feasibility and cost. It provides for incremental implementation strategies that are not determined by existing database structures, and allows for early exploratory usage of business intelligence beyond existing reporting requirements.

Structured and Unstructured Data

There are two types of data that business intelligence approaches consider:

• Structured data: traditional data warehouse solutions have been based on consolidating the structured data (numerical and categorical) recorded in operational systems where business information sets are identified by predefined structures (referred to as ‘schema on write’) and where a rules-driven template ensures data integrity. The business analysis focus is on data models, data dictionaries, and business rules to define information requirements and capabilities.
• Unstructured data: business intelligence solutions can include semi-structured or unstructured data which includes text, images, audio, and video. This data frequently comes from external sources. For this type of data, the structure and relationships are not predefined and no specific organization rules have been applied to ensure data integrity. Information sets are derived from the raw data (referred to as “schema on read”). The business analysis focus is on metadata definitions and data matching algorithms to define information requirements and capabilities.

11.2.4 Underlying Competencies

As in any business analysis discipline, the business analyst requires the fundamental communication and analytical competencies to be effective in liaising with both business stakeholders and technical solution providers.

In the business intelligence discipline, this coordination of business information requirements with business intelligence systems outcomes can be further enhanced by the business analyst’s specific competencies in:

• business data and functional usage, including terminology and rules,
• the analysis of complex data structures and their translation into standardized format,
• business processes affected including KPIs and metrics,
• decision modelling,
• data analysis techniques including basic statistics, data profiling, and pivoting,
• data warehouse and business intelligence concepts and architecture,
• logical and physical data models,
• ETL (Extract, Transform, Load) best practices including historical data track and reference data management, and
• business intelligence reporting tools.