Decision Engineering: Keeping the Decision Factory Humming

The subject of Decision Engineering has been gathering increasing attention, most recently with this article in IEEE Roundup.

Decision {insert any noun in the English language} ...

Among some of my decision, uh, interested friends and colleagues, there's a certain skepticism about new terms that start with the word "decision". From the very beginning, Decision Analysis has struggled with its name: the title of Ron Howard's 1966 paper, "Decision Analysis: Applied Decision Theory", doesn't exactly nail it to the wall, and Howard reportedly considered the phrase Decision Engineering before settling on DA. Is Decision Engineering yet another attempt to rename Decision Analysis in a dubious attempt to make it appeal to a broader audience? Don't be so quick to roll your eyes, old school decision analysts! This Decision Engineering is neither a vacuous rebranding of DA nor a hostile threat from outside our normative universe. So what is it, then? I prefer Tim van Gelder's definition:

"Decision engineering is applying relevant knowledge to design, build, maintain, and improve systems for making decisions."

This definition can be succinct because it takes a step back from the tools and processes of decision {_____}. The machine the decision engineer builds is not the decision, but the decision system. It's not about doing it once, but about having the organizational capability to process a stream of decisions, applying the appropriate treatment to each. In some such systems, the appropriate treatment will always be decision analysis, but in others, those that service a diverse stream of decisions, it might be some mix of decision analysis, data science and other methods. Is a civil engineer someone who builds bridges? Well, yes, but really the civil engineer's job is to figure out how to get things across the river, whether by bridge, tunnel or ferry.

DA, meet DE

A DA universalist would sweep all of decision engineering under the umbrella of decision quality. While that might make logical sense (to members of a certain tribe), it doesn't help us interface with and explain our domain to the outside world. Does the data scientist who's trying to figure out the probability of default for a few million credit accounts labor under a decision hierarchy? Yes. Does she look at it that way? Probably not.

The tools, processes and mindsets of decision analysis are an excellent way of looking at a broad range of decisions, and any decent decision engineer will be aware of them, but he'll also be aware of other techniques.

So, if you're a decision {analyst|coach|advisor|scientist|whisperer|ponderer|maker}, you'd be wise to pay some attention to the developing meta-field of Decision Engineering.