Insurance & Technology is part of the Informa Tech Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.

News & Commentary

09:16 AM
Rachel Alt-Simmons, <a href=SAS" />
Rachel Alt-Simmons, SAS

Train Your Data Scientists to Be Storytellers

Analytics aren’t all science – data scientists need to be able to communicate complex results to a broad community of stakeholders.

As organizations are building out their analytic capabilities, people everywhere are asking the question “What skills do I need to hire?” The October 2012 issue of the Harvard Business Review touts the emergence of the data scientist. It tells us that the primary function of this new breed of scientist is to “make discoveries while swimming in data” and whose skillset is a hybrid “data hacker, analyst, communicator, and trusted adviser.” While the article notes that universities are behind creating degree programs to support this paradigm, the search I performed turned up more than 20 Master's programs in the United States specifically related to business analytics/informatics. So is there consensus in how to create a well-rounded data scientist in the academic community? No.

Rachel Alt-Simmons
Rachel Alt-Simmons, SAS

First, programs in actuarial science, mathematics, statistics, and operations research (to name a few) have been around for decades. These programs provide intensive training on mathematical methods – and that’s about it! The newer families of business analytic degree programs provide a survey of mathematical techniques, but incorporate classes in communication, visualization, project management, and technical writing. Here’s the challenge – the business analytic programs may not provide the mathematical depth, but do train students in the soft skills necessary to advance analytics within an organization. The mathematically intensive programs train students on the methodologies, but not on how to effectively communicate results outside of a scientific community. (Anyone who has ever had to present a statistical model analysis to a sales organization will appreciate this skill!) The answer to the challenge is that you need both sets of skills: deep mathematical expertise and the ability to communicate results to a broad community of consumers.

Since the ability to effectively visualize, communicate, and present are often considered soft skills (i.e., those are either innate or can be picked up on the job), organizations often don’t pay attention to evaluating and increasing those skill levels for key analytic resources. It’s great that the universities are beginning to incorporate these competencies into their programs, but organizations need to continue to build and develop these critical skills for their employees. In a famous example, visualization guru Edward Tufte makes a compelling argument that engineers could have stopped the launch of the space shuttle Challenger, which exploded just after takeoff, if they had effectively presented information back to NASA officials: “The charts were unconvincing; the arguments against the launch failed.” And while we might not be making life or death decisions with analytics in our business, it provides a compelling reason to elevate the importance of visual and verbal communication in our companies.

[For more on Balancing Creativity and Control: Bringing Process Discipline to Predictive Analytics.]

Another skill that doesn’t receive enough attention is the ability to sell a concept or idea. As an analyst in an insurance organization years ago, I presented a set of statistical model results to the sales organization that would be responsible for implementing the model in the field. I made the mistake of showing the actual decision tree model: The model made several key predictive splits on dollar amount variables that were difficult to explain non-mathematically. The meeting was almost derailed when the executives began to ponder why the split happened on a specific dollar amount and we lost sight of the true objective: the ability to predict which of our key producers were likely to stop selling our products. Outside of our good science, it turned out that the artistry (better graphics and communication tuned to a particular audience) was more critical in advancing this initiative.

So how do we turn our data scientists into artists? One insurer holds a monthly contest for the “best visual.” Analysts submit their graphs and analysis in a competition that is judged by a group of non-analyst peers. In addition to getting valuable and objective feedback, participants can win a variety of prizes – the fun of the competition encourages participation and raises the collective understanding of the entire organization. More formally, consider working with your human resources department to add communication, presentation and visualization classes to the standard corporate development offerings. Start developing job descriptions for analysts that include the progressive development of these competencies.

About the Author: Rachel Alt-Simmons is senior industry consultant for Insurance at Cary, N.C.-based SAS. Simmons has driven business intelligence initiatives at Travelers and Hartford Life and has been Research Director for Life & Annuity at research firm TowerGroup.

Register for Insurance & Technology Newsletters