3.8.2 Business Data Analytics Perspective

Decision modelling helps demonstrate how data and knowledge are combined to make decisions, and decision analysis is used to examine the possible consequences of different decisions about a given problem. Both techniques are used extensively in business data analytics. Benefits of using decision modelling and analysis in business data analytics initiatives include:

• assessing the business decisions that need research and verifying if they align with the overall business strategy of the organization.
• verifying whether the research questions stem from the right business objectives the organization wants to assess.
• studying and modelling organizational decision-making and the architecture established so that repeatable decisions can be made.
• identifying root causes and appropriate actions based on understanding the business decision-making framework in an organization.
• analyzing complex organizational challenges by using advanced decision analysis models such as multi-criteria decision making (MCDM), game design, goal programming, simulations, and case base reasoning.

.1    Identifying the Research Question

Organizations rarely make decisions in isolation, and many decisions are part of the enterprise architecture model that drives a sequence of decisions or consequences. Identifying and analyzing potential impact of specific decisions can lead to forming higher-quality research questions. Decision modelling and analysis activities include:

• discovering decisions that require an analytical inference and entities affected by such decisions,
• prioritizing decisions that would most contribute to the business objective,
• enumerating decision alternatives in MCDM and similar approaches, and
• identifying the appropriate research question for each of these alternatives.

The necessary analysis tends to be more complex than decision trees, tables, or empirical cause and analysis. Selecting the appropriate decision modelling technique and the most effective decision analysis approach plays an important role in forming higher quality research questions.

.2    Source Data

Decision Modelling and Analysis do not have a significant role in the Source Data domain. However, the outcomes of decision analysis can help identify or validate source data.

.3    Analyze Data

Decision Modelling and Analysis do not have a significant role in the Analyze Data domain. However, the outcomes of decision analysis can support the overall data analysis work.

.4    Interpret and Report Results

Decision Modelling and Analysis do not have a significant role in the Interpret and Report Results domain. However, the results being reported may impact future decision models and decision analysis frameworks being used in the organization.

.5    Use Results to Influence Business Decision-Making

Decision Modelling and Analysis plays a significant role, after studying the analytics result, in the team's assessment and recommendations. The baseline decision model established during the identification of the research question phase is utilized to assess how analytical insights affect business decisions. Decision Modelling and Analysis can be used to influence business decisions by:

• evaluating conflicting insights and outcomes through the decision analysis process (for example, a decision tree can be used to follow a business decision to its logical impact).
• prioritizing business decisions that are most aligned to the organization’s strategy.
• formulating relevant recommendations based on understanding the impact of business decisions and their dependencies.
• allowing decisions to be integrated with the relevant business processes by an assessment of benefits.

.6    Guide Organization-Level Strategy for Business Analytics

Decision Modelling and Analysis do not have a significant role in the Guide Organization-Level Strategy for Business Analytics domain. However, Decision Modelling and Analysis are used to address challenges determining the optimal organizational level practices for business analytics.