On Insights I: ‘Aha’ Moments
On one side scientists talk about ‘Insight’ with a sign of reverence when referring to the processes, patterns, models, metaphors, stories and paradigms used to generate and communicate insight. Conversely, data professionals seem to regard ‘Insight’ as something trivial, achievable just by picking and combining the right visualizations and storytelling. Are the scientists exaggerating when talking about insight, or do the data professionals downplay the meaning and role of insight? Or maybe the scientific and business contexts have incomparable complexity, even if the same knowledge toolset are used?
One probably can’t deny the potentiality of tools or toolsets like data visualization or data storytelling in providing new information or knowledge that leads to insights, though between potential usefulness and harnessing that potential on a general basis there’s a huge difference, no matter how much people tend to idealize the process (and there’s lot of idealization going on). Moreover, sometimes the whole process seems to look like a black box in which some magic happens and insight happens.
It’s challenging to explain the gap as long as there’s no generally accepted scientific definition of insights, respectively an explanation of how insights come into being. Probably, the easiest way to recognize their occurrence is when an ‘Aha’ moment appears, though that’s the outcome of a process and gives almost no information about the process itself. Thus, insight occurs when knowledge about the business is acquired, knowledge that allows new or better understanding of the data, facts, processes or models involved.
So, there must be new associations that are formed, either derived directly from data or brought to surface by the storytelling process. The latter aspect implies that the storyteller is already in possession of the respective insight(s) or facilitates their discovery without being aware of them. It further implies that the storyteller has a broader understanding of the business than the audience, which is seldom the case, or that the storyteller has a broader understanding of the data and the information extracted from the data, and that’s a reasonable expectation.
There’re two important restrictions. First, the insight moments must be associated with the business context rather than with the mere use of tools! Secondly, there should be genuine knowledge, not knowledge that the average person should know, respectively the mere confirmation of expectations or bias.
Understanding can be put in the context of decision making, respectively in the broader context of problem solving. In the latter, insight involves the transition from not knowing how to solve a problem to the state of knowing how to solve it. So, this could apply in the context of data visualization as well, though there might exist intermediary steps in between. For example, in a first step insights enable us to understand and define the right problem. A further step might involve the recognition of the fact the problem belongs to a broader set of problems that have certain characteristics. Thus, the process might involve a succession of ‘Aha’ moments. Given the complexity of the problems we deal with in business or social contexts, that’s more likely to happen. So, the average person might need several ‘Aha’ moments — leaps in understanding — before the data can make a difference!
Conversely, new knowledge and understanding obtained over successive steps might not lead to an ‘Aha’ moment at all. Whether such moments converge or not to ‘Aha’ moments may rely on the importance of the overall leap, though other factors might be involved as well. In the end, the emergence of new understanding is enough to explain what insights mean. Whether that’s enough is a different discussion!
Originally published at sql-troubles.blogspot.com.