According to D.J. Power‘s “A Brief History of Decision Support Systems“, the first use of the term Decision Support System (DSS) was in a 1971 Sloan Management Review article. Gorry and Scott-Morton argued that existing management information systems focused on structured decisions and recommended that systems for semi-structured and unstructured decisions should be called decision support systems. When I first got into the field twenty years later, the term DSS was already in decline, having been overtaken by business intelligence. Although Howard Dresner is often credited for coining the term Business Intelligence (BI) in 1989 while at Gartner, an 1958 IBM Journal article seems to have beaten him to the punch by thirty years.
Even though BI and DSS are frequently used interchangeably, my background in cognitive science and artificial intelligence suggests business intelligence tools are examples of data-driven decision support systems. Model and knowledge-driven decision support systems use optimization techniques, simulation models, and recommendation engines typically not associated with BI.
Most BI applications are targeted at senior managers and business analysts interested in understanding the past performance of the organization. This retrospective view is almost always based on information managed in data warehouses and updated infrequently (weekly or monthly). To allow increased business agility, sometimes data warehouses are updated daily or even multiple times per day; when close to real time decision making is required, BI might even bypass data warehouses and go directly against operational business processes.
This so-called operational BI masks an important change that takes BI back to its roots in decision support. While operational BI may be used by people, it also can be used without human intervention by the operational processes. This application-centric operational BI requires a fundamental re-thinking of BI tools. James Taylor of Decision Management Solutions argues:
Applications cannot read reports or look at visualizations to understand the patterns in data. They cannot apply their own experience or the organization’s policies to ensure they make a legal and appropriate decision.
Actually, they can. We just have to move from relying solely on data-driven decision support to include model and knowledge-driven decision support. This move brings BI full circle back to its roots. Claudia Imhoff and Colin White, two well-respected BI analysts, recognize this shift and in “Full Circle: Decision Intelligence (DSS 2.0)” suggest we adopt the new term Decision Intelligence because “decision support is considered old fashioned”.
Fashions come and go. Welcome back, decision support. I missed you.
Jonathan – spot on. Recent work with clients have shown that their is the need for a new nexus between traditional BI and more model and computer aided decision making due in part to the ever disparate communities of enterprise users spread throughout the world. Often staff don’t even know what data and analysis is available or needed, Without an army of support staff, inefficiencies or worse are the result.