- Where Insurance & Technology Meet

Data Science Implementations: Keep Your Eye on the MoneyBall

Data Science is a hot topic these days.  The ability to quickly turn data into actionable information using advanced analytics is seen as a major competitive advantage.

However, a sense of urgency to put analytics into place may be creating some longer term consequences.  We’d like your thoughts and experiences in implementing analytics.

Cheaper, Faster Silos

This blog has noted a number of examples of the development of departmental data/analytics silos within organizations (see, e.g. Analytics In P&C Insurance: Is a Good Start Good Enough?).

We have also noted the challenges of establishing effective data governance structures and procedures, and the implications of action in absence of direction (see Data vs. Data Governance? Chicken & Egg Dispute Resolved!).

There is one more concern.  Specialist knowledge replacing business expertise.

The rise of the Data Scientist

Until a few years, the term Insurance Data Scientist was rarely used.  Most organizations assumed that the actuaries would handle any complex analytic requirements.

Almost overnight, insurance got swept by a big data wave that was covering organizations across all sectors.  In a 2011 report, McKinsey predicted that by 2018,   “the United States alone could face a shortage of 140,000 to 190,000 people with deep analytical skills as well as 1.5 million managers and analysts with the know-how to use the analysis of big data to make effective decisions.”

Insurance companies started to find ways to fill its own perceived gap.  Now, many insurers have developed structural units of data scientists.  According to Adam Charlson and Kristen Bargem, writing in Insurance Networking News,  these employees are “mathematicians, computer scientists, academically trained scientists (astrophysicists, ecologists, biologists, etc.), hackers and software engineers who use the scientific method.”   Sometimes they are headed with a c-level officer, carrying a title such as “Chief Data Officer”.

So what’s the issue?

We see this trend as mainly positive.  We just see a little too much enthusiasm for the expertise, and not enough focus on the outcome.  In many respects the rise of Data Science mirrors the rise of Computer Science.  There is a lot of promise, a lot of delivery, and good value.  Until the science becomes a goal in and of itself.

Writing in, Florian Zettelmeyer, director of the Program on Data Analytics at Kellogg School of Business, and Matthias Bolling, a consultant with Egon Zehnder International, put it succinctly:  “big data isn’t a data science problem. … Big data is a leadership problem.”

The authors use the Moneyball book/movie story as the example.  The differentiator for Oakland A’s manager was not the use of data to pick good players. It was the use of data to find undervalued good players that created success.

Billy Beane had an economics problem, not a skills problem.  And, according to Zettelmeyer and Bolling, Beane “had the courage to use the insight gleaned from data analytics to drive the way he ran his business… and changed the organization so it could deliver on that potential.”

So it goes back to leadership….

Data science is very attractive, but it will not solve all of the issues of any insurance organization.  It must be pointed in the right direction, based on long term goals and objectives.

Kurt Vonnegut , Jr. wrote: We are what we pretend to be, so we must be careful about what we pretend to be.”  If our pretense is to be the best in data science, we may achieve that goal, at the expense of other objectives.

What do you think?

What is your experience here?  Are you looking at data science as a component of your strategy?  How is it positioned in your organization?