Guest Post: The dirty dozen: twelve ways to fail at effective decision making
In the course of my decision analysis, analytics, and intelligence work for businesses and industry, I have identified a set of common points of failure in a typical decision engineering initiative. These characterize the “hidden traps”, where decision makers often struggle to preserve the integrity of the Decision Engineering life cycle.
Below is the chain of those failure points. I have listed them in a sequence in which I have found them to typically occur. Each one encompass the preceding two points in the list.
DE = decision engineering
DT = decision point(s)
DP = decision problem, including data and decision knowledge
DZ = decision processes (and workflows for operational decisions)
DM = decision modelling
DY = decision dynamics
DC = decision complexity
DA = decision analysis
DS = sequential decisions
DX = decision analytics
DO = decision optimization and automation
DI = decision intelligence
DQ = decision quality
Here is how to fail in each arena:
- Decision Points (DT): not identifying and mapping the decision points in one’s routine; this, at the technical/advanced level, is largely enabled with naturalistic decision-making assets and aids
- Decision Problem (DP): poor scoping and framing of decision problems
- Decision Process (DZ): ignoring process-awareness and subroutine-orientedness, which are requisite for most operational and tactical decisions (choices are often embodied in processes and workflows)
- Decision Modeling (DM): not accounting for context-embeddedness of virtually all decisions
- Decision Dynamics (DY): many decisions are “diffusion processes” (see decision field theory), so a decision-maker facing decisions with extended response deadlines may want to qualify it as such to engage the appropriate tool set
- Managing Decision Complexity (DC): objects of decisions can at times be complex (the core issue addressed by Decision Intelligence), including due to decisions/actions of the decision-maker. For this reason, dynamic decision-making important to consider. In particular, feedback and learning are to be taken seriously – either through experience (data mining/machine learning) or experiments (simulation and, less desirably, physical) or both
- Decision Analysis (DA): decision structuring, inclusion of dynamic uncertainties, multi-criteriality/pluralism and rational choice compromised, biases coming forward
- Sequential Decisions (DS): sequential decision-making is a relatively hard thing to model and use but in some cases it is a must
- Decision Analytics (DX): lack of competence or skills for implementing user-centered (let alone self-service) decision analytics to effectively enact co-agency in joint machine computation and human/expert judgment
- Decision Optimization and Automation (DO): inability to feasibly reduce or resolve convertible multi-criteriality/conflicting objectives into prescriptive resource allocation models to fully operationalize DX
- Decision Intelligence (DI): failure to translate DX/DO to closed-loop insight and, most importantly, foresight into the target function, process or task
- Decision Quality (DQ): failure to align, integrate and narratize DI with, to and ultimately into organizational decision-making processes in particular by differentiating among decision owners (those with decision rights), decision clients, decision users and other decision stakeholders, and mapping/assigning the first three (DTs, DPs, DZs) to them cohesively, supported with the previous three (DX, DO, DI).
With careful attention to these common traps, we can substantially improve decision making in our organizations!
Hayk Antonyan is a Thermo/Nano/Pico/Eco-Economist, Post-Normal Econophysicist, Decision Professional, Sustainopreneur, and Researcher and Developer of Nanoeconomics.