There are known knowns…

Photo by Jose Castillo on Unsplash

“There are known knowns…” - This is from a quote made by Donald Rumsfeld in 2002. In 2002, I was part of the crowd that ridiculed this quote made to the press. The video clip of the press conference can be seen here. Another video of an interview of Donald Rumsfeld done by Stephen Colbert in 2016 presents a very interesting additional take on this concept (starting at about the 4:30 mark of the video).

However, over time, I have realized the nuances and the depth of this approach of holistic thinking. This technique lends itself very nicely to strategy building. For simplicity, I will refer to this concept of known/unknown as ‘KU’ or ‘KU Matrix’ in this post

The motivation for this post comes from a discussion with colleagues on Data & Analytics Strategy. Here are some of my thoughts in the applicability of KU concepts with regards to Data/Analytics work and strategy.

 

KU Matrix

Let’s begin with a generic form of the 2X2 KU matrix with questions and answers being the 2 axes/dimensions. The 4 quadrants are numbered starting with known-known and moving clockwise.

1. Known Question – Known Answer: This represents the class of already solved problems for potentially useful questions. I am taking a little bit of a chance with terming these as ‘potentially useful’ question based on the association with known questions because it has been asked previously and it was worthwhile enough to be answered/solved. 
2. Known Question – Unknown Answer: These are potentially harder problems to solve primarily because the answers are still unknown.
3. Unknown Question – Unknown Answer: This group is classified as mysterious. I am happy to take suggestions on the right words to describe this quadrant. In addition, this can also represent a gap in the knowledge/capability/scope of the person or entity involved.
4. Unknown Question – Known Answer: This is an interesting category that I have termed as ‘Useless’. The more apt categorization is probably ‘Useless at a given instant'. This connects back to the gap mentioned in #3.

A specific issue (question-answer combination) can move among the quadrants:

• Answers can move bidirectionally between the known and unknown states
• Questions can only move from Unknown to Known. Once a question is asked, it becomes known! 

 

Data & Analytics KU Matrix

Here is my representation/interpretation of the KU Matrix pertaining to a Data & Analytics setting.

1. Known Question – Known Answer: This is a solved problem. In the analytics world, this is a dashboard, or a report consumed by a known set of users on a regular basis i.e., it is part of a business process. Dashboards and reports can lie anywhere along the analytics maturity spectrum (descriptive, predictive, prescriptive, etc.) Dashboards and reports generally fall under the category of ‘Business Intelligence (BI)’ set of tools. Another key aspect of this quadrant is that the data needed to answer this question is available and of ‘good enough’ quality. This is generally thought of a low-level of sophistication/maturity. However, it is quite the opposite and I’ll make this case in more detail later in the article.
2. Known Question – Unknown Answer: The question/need exists and is not solved yet. The most common reasons for things in this state will be data availability/quality, bandwidth of teams to do this work and inability to employ techniques to solve for this question in THAT order. This is categorized as ‘Discovery’ as it is breaking new ground regardless of the magnitude of the incremental change. When the discovery is useful and valuable, ‘Innovation’ kicks in.
3. Unknown Question – Unknown Answer: This is the mysterious quadrant! In addition to the characteristic elements of quadrant #2, there is an additional element of ‘Danger Zone’ classification here. This sets up this quadrant with very divergent and conflicting characteristics.  Discovery/Innovation are generally positive. ‘Danger Zone’ is the opposite. Let me try to explain the ‘Danger Zone’ characterization. In many cases, executives and senior management will pose the following challenge to the analytics team – “Can Analytics/AI/ML help to find something useful and interesting from our data?” This question is an extreme example of the Danger zone area. It is very akin to trying to find the needle in a haystack scenario. It can potentially lead to a useful discovery/innovation outcome – however, the probability of this success is low. It has a higher potential to turn into a pit/drain of resources and effort over time. Additionally, this setup leads to ‘shifting goalposts/project creep’ scenario due to the broad scope ‘useful and interesting’.
4. Unknown Question – Known Answer: Quadrant 4 is a relative of quadrant #3. This represents an issue where the data/analytics has come to a logical milestone/conclusion i.e., some end of the ‘Discovery’ phase. It is a ripe starting point for the Innovation phase to find utility and value for this discovery. 

From a roles and responsibilities perspective at a team/organizational level, the KU Matrix looks like this and represents the teams having main and active responsibility for issues in each quadrant i.e., who is holding the ball?

What does this all mean? 😊

My overall opinion is that the intent of all the work from an Analytics focus should be to get the issue to the Known-Known quadrant (quadrant #1). This is where true value ($$$’s) to an organization is created with the analytics output becoming integrated into business processes that are repeated at scale.

 

Dashboards and Reports get a bad rep sometimes, much of the reason is that it is not a shiny new object anymore. Here, I am using dashboards and reports as a common means of disseminating valuable answers and information – it could be as simple as an alert on a mobile or an email. These create repetitive value. Additionally, things in Quadrant 1 have reached the ‘product’ stage and that is a more mature state than being in the ‘project’ stage.

There are potentially various other dimensions that can be used to get a new perspective on the KU Matrix related to the Data and Analytics scope of work. I’d be very interested to hear any feedback on additional dimensions to look at this.

There are known knowns… It is the place to be!