How Do You Make Decisions? (Part 1 of 4)
Is decision making really a purely rational process? If not, what is the importance of business intelligence?
- By Barry Devlin
- May 31, 2016
Data warehousing and business intelligence emerged from an earlier discipline called decision support systems (DSS), the evolution of which is described by Professor Dan Power in "A Brief History of Decision Support Systems". While coming more from the academic rather than commercial view of DSS, the paper provides an interesting insight into the foundations of decision making in business. Power describes DSS in five categories: data-, document-, communications-, model-, and knowledge-driven DSS.
The category names explicitly refer to a series of drivers of decision support -- from information through collaboration to analytics and artificial intelligence -- each driver being some type of technology (including electronically stored information as technology). However, something is missing. Where are the people who make these decisions? How do people reach decisions and how do these technologies actually help?
The most widely used theory of how people make professional decisions is rational choice theory. Beloved of economists, academics and business schools, the idea of rational choice dates back to the Age of Enlightenment in the 17th and 18th centuries and was extended by a focus on "facts" that emerged in the following century.
This theory suggests that we start solving a problem by defining objectives, and then we list pros and cons of the various choices. We then construct measures of utility that describe numerically how desirable or valuable each option is and multiply these measures by the estimated probability of their occurrence. This may sound somewhat tedious and indeed, it is. The complexity and urgency of business decisions demand more practical approaches.
This has led to the concepts of bounded rationality (making a choice from a limited set of alternatives or consequences considered), satisficing (choosing an alternative that exceeds some criterion), and rule following. As discussed by James G. March in his 1994 book A Primer on Decision Making: How Decisions Happen, such approaches have long been used to describe the mechanisms underlying business decision making at an individual level.
Lack of information is typically considered an important constraint on decision making, together with decision makers' limitations in attention, memory, comprehension and communication. Within these constraints, the basic assumption is still that decision makers strive to be rational.
I have never, in over thirty years of business, encountered anybody fully and formally using these approaches a priori to address any business decision. Nonetheless, it is the foundation -- albeit largely unspoken -- of BI.
The assumption is that making information available to decision makers contributes to their ability to understand pros and cons, to measure utility, and to estimate outcome probability, and therefore providing information increases their ability to make good decisions. Of course, this is true, at least to some limited extent, or we would see no correlation between BI use and better decisions.
Certain types of decisions, such as near real-time, high-volume operational decisions, definitely benefit from being data driven. Indeed, given their volumes, the only way such decisions can be acted on is through automation -- the decision is algorithmic and human decision making occurs only at design time or in the event of exceptional circumstances.
However, the limited success of BI in significantly improving business decision making in a general sense over the past thirty years must surely suggest that we are, at the very least, missing something. Could this contribute to the limited adoption of BI tools frequently commented upon by analysts and vendors alike?
Noting the apparent lack of rationality in human decision making, various experts -- notably including psychologist and Nobel prizewinner in Economic Sciences, Daniel Kahneman in Thinking, Fast and Slow and Professor of Psychology at Duke University, Dan Ariely in Predictably Irrational: The Hidden Forces That Shape Our Decisions -- have proposed that the human mind is somehow unfit for purpose in the highly complex decision making of the modern world.
They contend that various brain constructs dating back to our evolution through hominid and hunter-gatherer history prevent us from exercising proper rational thought. Predictably enough, these thoughts have been seized upon by economists and politicians as reasons why "ordinary" people need to be nudged in their thinking toward the "right" decisions in everyday life.
In a 1998 Harvard Business Review article "The Hidden Traps in Decision Making", John S. Hammond, et al., describe a range of snares for the unwary decision maker, such as anchoring, confirming evidence, and framing. They go so far as to suggest that "sometimes the fault lies not in the decision-making process but rather in the mind of the decision maker. The way the human brain works can sabotage our decisions". Their advice can be summarized as: we should gather even more information to offset this risk. They offer no proof that this might help.
These conclusions may be somewhat disturbing. Decision making in business is, at best, only a partially rational process, and information contributes only lightly. Is business intelligence a misnomer? Have we been barking up the wrong cognitive tree? Despair not! Part 2 of this series offers a new and broader perspective that positions current BI more realistically and points to areas where new approaches are needed.
Dr. Barry Devlin defined the first data warehouse architecture in 1985 and is among the world’s foremost authorities on BI, big data, and beyond. His 2013 book, Business unIntelligence, offers a new architecture for modern information use and management.