What is decision analysis?

Decision analysis is a subcategory of decision science. Ultimately, all data science could be classified as decision science, because the entire point of data analytics and informatics is to enable better decision making processes to yield better outcomes from decisions. While decision science and decision analysis cannot eliminate the possibility of sub-optimal outcomes, they are in themselves useful if not mandatory processes. They can, for example, enable teams or decision makers to move forward with confidence in their assessments of decision variable behaviour across decision parameters, and thus with more reasonable expectation of probable outcomes. Decision analysis includes the rigorous evaluation of alternatives against criteria in a variety of environmental states. For example, a policy decision can be evaluated under a variety of budget levels, a macroeconomic cycle, or a parliamentary mix.

What kind of decisions can be analyzed?

Decision analysis is relevant across all kinds of organizational decisions. Anywhere you can find decision variables, independent choice variables that combine to impact dependent outcomes, decision analysis can be applied. For example, recruitment agencies can apply decision analysis to determine which segments of the population should be targeted to increase the probability of finding the best fitting candidates. Schools can use decision science to determine which scheduling selection is likely to produce the optimal class attendance. Organizational boards can use decision science to determine which markets are the first best choices for business unit expansion. Decision science can even help organizations decide if they should train, hire, and retain staff, or outsource complex operations to specialized firms.

How should decisions be analyzed?

Before proceeding with decision analysis, management needs, of course, to decide which decisions should be rigorously analysed. When key decisions are identified, analysis feasibility should begin.

To what extent should the decision be analysed? What options are available given the data?

After feasibility, management decides in what way to proceed with analysis and this triggers the data science research pipeline flow. A project plan is prepared. Data are gathered, then prepared. Models are selected, then constructed. Finally, results are analysed. At many points during the decision analysis pipeline data are revisited and plans amended, if necessary.