From Data to Information to Actionable Insight (Part 2 of 2)
Part 1 of this series showed that data needs the context of information to be useful. Here we explain why information alone is still insufficient.
- By Barry Devlin
- November 8, 2016
Having collected context-rich information, you need an understanding of how to use it to good effect. This is knowledge that originates in the human mind. Knowledge emerges in part from the information (gathered and processed in BI and analytics tools) directly related to the problem or decision at hand. Of at least equal importance is prior information garnered over a lifetime's experience. This extensive set of personally important information is unique to each user but completely ignored by current decision-making support tools.
The Importance of Real-World Knowledge
Indeed, the decision about which information to actually collect is (or should be) based on the knowledge of what could be done with it or gleaned from it. This tight feedback loop between information -- a physical representation of the world (today mostly in bits and bytes) -- and knowledge -- a set of mental constructs in someone's head -- has led to the confusion that surrounds knowledge management. Most knowledge management projects are actually about management of information, which may or may not depict human knowledge.
To solve the business problem of the new competitor posed in Part 1, the first step is to use information with context that describes the relationship of the underlying data to the real world rather than simple data.
Furthermore, this information has real value only via the business knowledge of the people who know its uses and possibilities -- as opposed to some external data scientist driven by statistical correlation rather than valid causation. One poster child for the limits of raw correlation can be seen in the rise and fall of Google Flu Trends.
Human knowledge is thus central to using information in decision making, but are information and knowledge sufficient for insight? Do they lead directly to taking effective action? One further level of qualification is needed: meaning.
Considering the Meaning of Meaning
Meaning is most easily understood as "the stories we tell ourselves" about the information we receive or the knowledge we have. Such stories are deeply influenced by the social milieu of the organization and the beliefs and intentions of its members.
Consider a common experience of BI analysts. Using a typical tool and her knowledge of the business, the analyst comes to some conclusions about the customers likely to churn in the case of the newly emergent competitor mentioned in Part 1. Her manager, however, responds: "Can you take another look at the data? I don't believe the situation is as bad as you think."
Information and knowledge together offer the insight to the analyst that the churn outlook is bad, but a particular offer to key customers might turn it around. That is at odds with the manager's story about the situation.
The meaning he attaches to it may reflect how he believes the CEO would react to such news. His insight is that he will be blamed for the problem. Alternatively, perhaps the manager has shares in the competitor, in which case his insight may be that he has a better job opportunity elsewhere. In each instance, it is through meaning that insight finally emerges.
How Meaning Leads to Action
It is meaning that also provides the incentive to action -- it makes the insight actionable: make that offer to key customers, redo the analysis, or submit resignation (analyst's or manager's).
Significant actions -- those that are not or cannot be operationalized and automated -- depend fully on a person's motivation to "pull the trigger." Motivation may or may not be well aligned with business goals or societal norms. Only by understanding the stories told and the meanings assigned by individual decision makers can we finally know what action will be taken.
This human and social dimension of assigned meaning is entirely missed by most discussions on the topic. For example, Brent Dykes, director of data strategy at Domo, describes actionable insights as the missing link between data and business value. He provides a useful set of six criteria -- alignment, context, relevance, specificity, novelty, and clarity -- for determining how actionable insights are ... in an ideal world, where business goals and human meaning align.
However, because his argument is based on a naïve data-information-insight triangle, it misses the reality of how many decisions are driven not directly by data or information but indirectly by the stories we make up about them.
Are Your Decisions Insightful?
In the modern, fast-moving, digitalized world, closing the gap between gathering data and taking action is paramount for successful business. Closing that gap depends on understanding the progression from information (and its computer-friendly subset, data) through knowledge and meaning to actionable insights.
Rather than being data driven or even information informed, we need to look toward becoming insight inspired. That demands deeper thinking about how we human beings actually make decisions.
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.