The emergence of Software-as-a-Service in the industry has brought new data privacy and ownership concerns into the contracting process for insurers. All vendors, even startups, understand that insurers’ data can’t be shared without explicit permission. However, not all vendors recognize that the same approach must apply to aggregated data. It’s possible for vendors to look across multiple SaaS insurer clients to create a large data set that represents the industry trends, but—even when blinded—this data must be treated with the same careful ownership rules in place.
Sometimes this kind of aggregated data set is intentional. Contributory approaches require insurers to provide and allow use of their data set as part of the participation agreement, usually as a way for smaller insurers to pool their data to compete with the kind of analytics only large insurers could otherwise perform.
But I’ve talked to several InsureTechs who provide SaaS tools to insurers that are not, at their core, about contributory data, and these InsureTechs made it clear that they do not understand the full scope of ownership issues around insurers’ data. They expect they’ll be able to define better predictive models and public trend reports by running analytics across their whole client base.
Since the competitor data itself wouldn’t be shared, just the resulting formula-based models, the InsureTechs I spoke with didn’t understand why I was challenging this approach. It’s not directly sharing data, after all.
It’s true that several midsized insurers who sell standard lines of business and take a general market approach could end up helping each other in an aggregation scenario. Imagine, however, there’s a small or midsized insurer that does not have a significant percentage of the overall market, but has a large share of a very specific market demographic (such as jewelers or flower shops). While the insurer’s data set might be too small to determine general trends about a line of business, its data might be better at identifying trends for those distinct demographic groups than any other insurer.
A vendor that utilizes that data in an aggregation scenario will end up unwittingly helping other insurers compete in that distinct demographic against the small insurer who brought all of the knowledge to the table. Aggregated data, even if used exclusively to infer trends and insight, must have the same ownership and privacy rules applied to it as the raw data itself. Data sharing in any context must be made explicit or not happen at all.
Insurer CIOs need to ask how any vendors hosting their data or generating new data/transactions will treat that data at an aggregated/blinded level. If you don’t want a vendor using your data to build models and trends or to otherwise help all clients, spell it out in the contract.
If it’s not explicitly stated, assume a vendor—especially a startup—might not understand the extent of data ownership needs in insurance and will use it for cross-client data aggregation.
Add new comment