We’ve all it seen the onslaught, if not experienced it first-hand. An inquisitive three-year old unleashing a salvo of, “Mommy, why is this,” “Daddy why is that,” and “What is that thing.” My recent favorite rendition came when I overheard a parent who dared admit, “I don’t know,” to which the immediate follow up was, “Why don’t you know?”
These types of questions are many things: adorable, persistent, but in my opinion, their most prominent characteristic is naiveté. That is by no means the child’s fault. They need guidance. Without experience, people need to learn the structure of asking effective questions, to develop awareness of the connection between the form of their question and the type of information they are seeking, to learn to narrow their subsequent queries, and, perhaps most importantly, to develop a sense of what types of questions the respondent is equipped to answer.
This dynamic is not dissimilar to what can occur when deep data science is suddenly made available within a company. Outside the data lab, no one can be reasonably expected to know what questions data science is prepared to answer. So how can smart business users ask the right questions in a wide open, unstructured landscape?
It is important that their natural curiosity be quickly honed to get beyond naïve questioning to a state that challenges the data science team to answer or pursue answers to questions that have significant meaning and value to the business.
When large numbers of people know about data, analytics and visualizations they become stakeholders in data science.Tweet This
So how can you quickly bring an organization along to that level of sophistication? Productized analytics, analytic solutions that address common business problems, are a very effective starting point because they provide a framework for putting analytics to work. In fact, while there is often an enthusiastic “sky’s the limit” attitude toward data science, productized analytics provide a means for quickly getting to work on a problem within an understandable construct.
With a productized analytic , a data consumer is presented with a solution that is well documented. This creates some useful boundaries within which the process of asking questions can be refined. A curious user can dig into a productized analytic to understand the data that is used, gain an appreciation of the specific analytics and models applied to the problem, develop an understanding of the depth within the visualizations used to convey the insights derived, and, significantly, use that awareness to contribute their business knowledge to improving that data product.
Suddenly the divide between the data scientists working in the vacuum of their lab and the rest of the business is eliminated. There is a feedback process through which people with business knowledge who will take action based on data insights can refine their questions and contribute to the improvement of the analytics. This is not only a meaningful improvement for executives and others who consume the output of data science, but it changes the script for the data scientists. As business users become more familiar, they can get beyond the basic questions of “what” and “why,” and start to ask questions that are more sophisticated and help the data science team better orient their efforts toward high value outcomes.
Curiosity, the drive to respond to each answer with a renewed cry of “why” is not limited to children. It is the engine that drives the most energetic start-ups as well as start-up like groups within big companies. Seeing a response, whether it is a product or a data insight, not as an end but an opportunity to ask questions that allow us all to dig deeper into the matter is something to be encouraged. Productized analytics, with their documented and explainable methodologies, provide a structure within which precocious, non data scientists can ask informed, sophisticated questions. When large numbers of people know about data, analytics and visualizations they become stakeholders in data science.
The question, “Why don’t you know,” can be quite meaningful when it comes from someone who understands the connections between previous questions, is aware of the data available, and has seen what analytics have previously delivered. Productized analytics are the fast-track to making business users good questioners.
Ryan Garrett is senior business development manager for Teradata Field Applications. His goal is to help organizations derive value from data by making advanced analytics more accessible, repeatable and consumable. He has a decade of experience in big data at companies large and small, an MBA from Boston University and a bachelor’s degree in journalism from the University of Kentucky. すべての投稿の表示