INTRODUCTION
Among the antecedents of good decision-making in any organization is Trust in Data and Trust in Technology. Failing to trust the data, analytics, KPIs, metrics, measures, or the technology itself will almost invariably channel the decision-maker down less optimal paths with unpredictable tactical and strategic ramifications for the decision-maker and the firm.[1] The concept of Trust has existed in the peer-reviewed literature for over a century and has expanded to include changes in the ecosystem. As the industrial age matured, researchers quickly realized that trust in the workplace played a key role in productivity, performance, innovation, and a person's longevity with the firm. As the Information Age matured, the concepts of Trust in Data and Technology quickly emerged, and it was discovered that trust in both appeared to develop in the brain in ways different than trust in people. Human Factors experts picked up this point, and significant leaps have been enabled by this body of knowledge in the design process of countless objects and systems.
Executive-level teams of countless organizations have advocated for data-informed decision-making and the desire to get away from “gut feel.” Yet this long-sought-after end-state has remained an elusive goal for scientific reasons surrounding the intractable functioning of the brain. This lack of understanding has not deterred executive teams from backing up their desire with sizeable investments in technological and methodological solutions designed to capture, model, and visualize data to inform decision-making at every level.
And yet, regardless of the genuine level of desire or dollar amount of technological investment, data that does not agree with the decision-makers 'mental model' (think 'gut feel') is almost always rejected. This is particularly true when Higher Order decisions (those requiring high degrees of complex reasoning) are being performed. These are also often the most important organizational decisions, as decision latitude and complex reasoning both tend to increase with the level of the decision-maker. Thus, a senior vice president determining strategy is far more likely to reject data that does not agree with their internal model or predictions made by the Neocortex than a low-level clerk ordering office paper. What is occurring here is a complex ‘dance’ of multiple organs and systems in the brain playing out. The visible manifestation is some permutation of “I don’t care what the data says, I trust my gut, and my gut is saying to do X.”
Complicating the issue of Trust in Data & Systems is that after data is rejected because it disagrees with the individual's internal model, the overarching trust in data within the domain and technological systems is often reduced. Again, this results from a complex dance of systems in the brain, and the visible manifestation is often a derogatory statement like “I just don’t trust the data from that system half the time.” This effect can have an immense impact on the enterprise's perception and weighing of benefits, risks, and actions as the lack of trust in data occurs precisely where it matters most – in the minds of most senior decision-makers. This concerning reality leads us to a critical path of scientifically grounded inquiry – what are the human and technical factors associated with influencing the trust of data in decision-makers' minds when making risky decisions?
FALLACIES AND FALSE ASSERTIONS
The impact of data trust on enterprise risk decisions is a delicate topic as everyone has an opinion, yet these opinions are rarely scientifically grounded. System and dashboard developers will say that we must begin with User Stories, Use Cases, or Requirements. These preferably come directly from the user/decision-maker, ensuring the system will be used – after all, they requested it. Use cases are combined with Design Thinking and creative visual techniques to create an intuitive interface that is easy to use and aesthetically beautiful. The result of these exhaustive efforts will undoubtedly be data-informed decision-making that is a far cry from “gut feel.” Unfortunately, with little understanding of the Cognitive Neuroscience of decision-making, this approach is almost always doomed to fall short of its goal. This outcome is well understood in the Business Intelligence community, and they have structured their success metrics on how many dashboards they build and not on how many times a user logs onto the system or even more difficult to determine how much decisions are improved through the use of data.
A SYSTEMIC ISSUE
Part of the problem with engraving “data-driven decision-making into the DNA of the firm” is systemic. If one asks a data scientist or business intelligence specialist what is needed to help data-driven decision-making become part of the DNA of the firm the answer is always some permutation of “we need more data, labeled data, better data models, algorithms, and more intuitive visual analytics” etc. This data-centric approach, which has become dogma, fails to consider precisely how the brain makes decisions with data. This is thus a widespread misbelief in what prevents data-driven decisions from being made; the contrarian but true answer is the brain is structured with systems that both cooperate and compete when making decisions. At the center of the difficulty of making decisions with data is the structure of the brain itself and that is it is not structured to easily enable decisions to be influenced by data.
Thus, if we want to improve data-informed decision-making especially where creative and high-risk decisions are made, we must place as much emphasis on understanding the human factor as we do on the external antecedents such as data quantity, quality, modeling, design, etc.
Next-generation Analytics will be far less about the models and methods than it will be about enabling human/computer complementarity through a Cognitive Neuroscience based understanding of how the human brain makes decisions with data.
Part II of this series will introduce a key to the next generation of Analytics in the form of the Cognitive Neuroscience of Decision-Making.
[1] Research of risk decisions made from qualitative models or consensus in the form of ordinal valuations (high, med, etc.) is often worse than the decision-makers use of System1 or Heuristics based on years of experience.
Comments