Actuaries have always been the people carriers have
used to deal with and understand data and determine
the best way to turn that data into better pricing and
risk decision-making, according to Jeffrey Goldberg vice
president, research and consulting at Novarica. But in
today’s world the discussion has changed to turning that
data into a structured technology rather than an internal
“It is easier for insurers to understand the value for
this kind of usage when it comes to pricing and risk de-
cisions because the industry understands this is so foun-
dational vs. some other areas,” says Goldberg. “Insurers
often forget about using data when it comes to claims
processing and marketing. Insurers struggle with that.”
The problem faced by carriers is there are two sides
to the issue of Big Data. One is the data source, accord-
ing to Goldberg. He refers to the data that has high
velocity, high volume, and high variety. The second issue
involves the technology used to process Big Data in a
way that allows companies full access to its potential.
The technology often requires either extensive training or
hiring people who can bring in that expertise—usually multiple
people, according to Greenberg.
“It’s not enough to hire a single engineer; you typically need
people to understand how to program against it, people on the
infrastructure side who can maintain a database, which requires
a different set of knowledge than a Microsoft or Oracle data
base,” he says.
This is definitely in the wheelhouse for large insurers,
according to Greenberg. He feels it is a lot to ask of mid-tier and
small companies to expand the staff when they are uncertain
what the value could be. That issue alone often prevents midsize
insurers from adopting Big Data plans.
Areas such as telematics, weather, and geo-spatial data are
discussed regularly by insurers, but what Goldberg sees is these
carriers are getting a weather data feed, but all they are really
going to use is the date, the temperature, and conditions.
“They are putting it in their database, but it’s not really a Big
Data project; it’s more like a normal data project,” he says. “The
value of that is high, but the question is what do you have?”
The weather database gives access to weather fluctuations
every minute. Goldberg believes Big Data in this form requires
a new set of skills to process the information and interpret it.
“Without a clear use-case value, it’s hard to justify the ex-
The Answer Is: Big Data
The question? How can insurance companies achieve greater value with their ana-
lytics and predictive modeling tools?
By Robert Regis Hyle
Current and Expected Data Source Usage by Insurers
K Already using this
K Probably Not
K Definitely Not
K (Did not answer)
Source: Novarica Survey of 58 Insurer CIOs and Senior IT Execs, 2015Q2
penditure,” says Goldberg.
That’s not to say there isn’t value to be gained from weather
data, but if the use case isn’t clear, data providers will need to
prove it out first before other insurers can put some numbers
around what they are gaining from it.
“When insurers invest in better decision-making they are
going to make better risk decisions and those who don’t invest
are going to be adversely selected and have to offer lower prices
for worse risks and higher prices for good risks,” says Goldberg.
Goldberg believes the insurance industry has crossed the
hurdle of data analytics and predictive modeling using traditional technology. Most insurers also are putting the time in to
fix data sources and augment third-party data when possible.
Future surveys will show more insurers use this data across
the whole business for different areas of investment, but given
the financial differences between top tier and smaller insurers,
Goldberg contends it will continue to be haves vs. have-nots.
Goldberg believes large insurers will continue to invest in
Big Data, but it remains unclear if midsize and smaller carriers
will invest in the technology.
“It will be like Big-Data-in-a-box approach,” he says.
“Whether it is SaaS or something you install, we’ll see more of
that available. They will start to build out tools that allow you to
easily deploy the whole environment and manage it.” ITA
Third-Party consumer or business data
Audio data (e.g. voice recordings)
Historical stock market data
Physical sensor data (e.g. motion sensors)
0% 20% 40% 60% 80% 100%