Majesco Insurance Blog

Lost in Translation Part 1

Dec 3, 2015 | By: | Topic(s): Business Transformation, Data & Analytics, Data Strategy, Digital Strategy, Explosion of New Data, Group & Worksite Benefits, L&A, P&C, Shifting Customer Expectations

 

Why Effective Data Science Storytelling Matters

Data Scientists are analytical. Their lingua franca is mathematical logic. Their value to insurers is in their ability to turn data into dollars and analytics into sense.

If you are a data scientist, you understand that this isn’t simple. The CFO may not understand the difference between sensitivity and specificity on a ROC curve and the head of Marketing may not need to know that a support vector machine constructs a hyper-plane in a high-dimensional space. You know these principles are important and you want to communicate that you chose the best machine-learning algorithm to solve a classification problem. What really matters, however, is the clarity and impact of business information. Without clear communication over advanced analysis, organizations are prone to poor decisions that can result in years of waste and profit loss. Can the data scientist move an insurance enterprise in the right direction with a compelling story?

The answer, of course, is “Yes.”

This is why effective storytelling matters to the data scientist and to the insurance organization. The organization needs the right information presented clearly enough to make informed next steps. At a high level, it simply means that data scientists and those they serve should understand one another. Understanding will require some translation.

All of this may seem logical and self-evident, but as we dig deeper, we’ll look at the complex details in this one relationship and its communication issues. Instead of simply looking at the difficulty of data storytelling, it may help to highlight all of the roles that the data scientist fills when he or she tells data’s story. There are at least four different translators present in just one data scientist and the organization needs to hear from all of them.

The Journalist — Data’s Unbiased Reporter
Most business decisions, hopefully, are made on the basis of information. The information transaction is a two-part process. First, there is a reporter of the information. Then there is a reader/viewer/listener. The reporter may be a person. It may be a report. It may be a dashboard. No matter how the information is transacted, its most vital aspects are an effective summary, truth and lack of bias. The data scientist, in the journalist’s role, turns numbers into clear words and concepts that are useful in making informed decisions. This summary is vital. The “scientist” within data science is also responsible for translation that keeps the data story real. Others within the insurance organization may have a vested interest in swaying decisions. By keeping data’s story unbiased, the data scientist fulfills the important role of trusted journalist, making sure that no key insight is twisted or lost in translation.

The Attorney — Data’s Advocate
Data, once collected, deserves to have its story told. If an insurer has gone to the trouble of modernizing big data collection, warehousing and analytics, it should finish the job by making certain that data is being “heard” by those who need to know. There will be many time-crunched executives who simply want to see the numbers or have them visualized. “Just give me the dashboard.” They are making an assumption, however, regarding their abilities to see and draw the proper inferences from the numbers. A data scientist who is adept at communicating, will bring data’s story to life, not to sway the jury, but to tell the complete, unvarnished story.

The Consultant — Data’s Concierge
The consultant isn’t in charge of running the business. The consultant’s job is to be an accessible expert. In many cases, the data scientist isn’t simply presenting analysis; he or she is answering questions. These may be informal, off the cuff questions or formal questions posed as an important point of inquiry. (e.g. “We are thinking of launching a new auto insurance product aimed at “low-risk” 20-30 year olds and we think we can sell 10,000 policies in the next year. Using key data indicators, can we forecast claims on 10,000 individuals in this age group?”)

The consultant is available, not just to answer that question, but to expertly navigate data sources and bring them to light. The consultant may suggest outside data streams to use for analysis, data indicators that may work well as filtering mechanisms, etc. The consultant researches, builds a framework and understands data’s backstory. Grasping the data’s backstory (the building blocks used to find conclusions) can uncover caveats and risks or it can also serve as a strong support for trust. Trust in data’s story, as we’ll discuss in our next installment, is crucial.

The Marketer — Data’s Salesperson
Effective data storytelling should cut through massive amounts of unimportant data to focus precisely upon insights that will make a difference. Just because the data scientist is a good communicator, doesn’t mean that all of the organization’s managers and executives are good listeners. In a device-distracted, over-worked, attention-deficit society, data science has to cut through data diversions and continually drive to the heart of what matters. Using marketing-oriented concepts, data scientists need to sell business leaders on the importance of data and which data threads are likely to yield good insights. Experienced data scientists often become masterful marketers, selling data’s value through compelling stories with targeted messaging to varied executive roles.

In summary, then, the data scientist works as the journalist, the attorney, the consultant and the marketer to place the right messages in front of insurance stakeholders, translated in ways that will bring meaning to the business. Data science translation is enhanced by the tools of data analysis and the capabilities of insurance systems for data gathering, neither of which can be overlooked.

The challenge for an insurer is in hiring or contracting data scientists or data partners who are also effective communicators. In Lost in Translation Part 2, we’ll look at what steps organizations and data scientists can take to improve data science storytelling.

About the Author

Leave a Reply

Subscribe to our Research and Communications