Digital processes and expanded capabilities have produced vast new sources and types of data, everything from telematic information gathered from drivers and cars to data harvested from wearables and more, sources said.
“With the new data streams available with the digitization of everything, insurtechs are able to suck in data they never had access to before,” said Alex Schwarzkopf, co-founder and CEO of Pillar Technologies Inc. in New York.
Coupled with gains in computing strength required to crunch that data, the evolution is producing new information for users.
“There are types of data that we used to not be able to analyze that we now can analyze,” said Mark McLaughlin, global insurance director for IBM in Chicago.
Video footage from a drone or security camera, for example, requires vast computing resources to analyze for extractable and usable data, he said.
The sheer volume of the data is also unprecedented.
In 2015, there were 11 quadrillion terabytes of unstructured data captured, according to Andrew Johnston, global head of insurtech for Willis Re in London.
That will grow to 39 quadrillion terabytes in 2020 and 104 quadrillion in 2025, he said.
“The speed at which data is being captured is also increasing,” Mr. Johnston said. “That has to be met with some type of improvement in one’s ability to ingest, host and manage” the data.
“Nondiscrete data tends to be large data,” Mr. McLaughlin said. “It has to be in the cloud. You can’t build big enough data centers to hold all the data.
The average telematics device generates 4 gigabytes of data each year. Where am I going to put all that for my 1 million policyholders?”
