DataOps is the latest agile process-oriented methodology used by insurers to improve data quality and reduce data analytics cycle time.
Insurers embracing DataOps can become the industry and sector disruptors rather than being the disrupted.
As industries across all sectors embrace corporate agility, and traditional vertical lines continue to blur, insurers are finding themselves in a digital transformation race.
The race, as expected, has become more of an ultra-marathon than a sprint, with a greater focus on time frame and pace than on short-term winners and losers.
DataOps is an emerging discipline recognizing the value to an insurer of data, both their own internal data and any other data available to them, in that digital marathon.
Unlike traditional marathons where runners compete at a slower, steadier pace, DataOps defines a framework where short, frequent activities ensure the optimal race is being run both strategically and tactically.
DataOps, as defined by Wikipedia, is “an automated, process-oriented methodology ... to improve the quality and reduce the cycle time of data analytics.”
It is especially important to highlight the two-headed DataOps value proposition—data quality and data analytics cycle time.
Since DataOps is new, specific insurance use cases are evolving. However, because DataOps is essentially a set of best principles, insurers can get started today in applying DataOps principles to their data usage.
Resources such as DataKitchen's The DataOps Cookbook can help improve insurers' understanding of DataOps thinking. The cookbook identifies specific techniques to begin applying product-independent principles, equating a data life cycle to a manufacturing process. In this context, it describes how statistical process control, a lean manufacturing quality tool, can be applied to any data used by an insurer. This approach then helps achieve the DataOps data quality value proposition.
Likewise, decreased data analytics cycle time resulting from DataOps is something any insurer can now begin visualizing, regardless of what types of data products are, or aren't, being used. Specifically, insurers using robotic process automation tools can get a head start on DataOps simply by applying the same analysis done for their current RPA implementations to their data process life cycle.
Insurance is an industry built on data. Insurers embracing DataOps can become the industry and sector disruptors rather than being the disrupted.
For insurers thinking that it may be too early to jump into DataOps or those waiting for others to stub their toes first, tools such as Celent's Demystifying Artificial Intelligence in Insurance: The Tools Supporting Data Science and the Rise of DataOps, can provide perspective.
The report helps insurers apply manufacturing terms not commonly used in the insurance industry today, such as industrialization and common assemblies, to their data pipeline. For example, the metadata of an insurer's data parallels a manufacturing bill of materials, where all components come together as a data inventory for use by an insurer's data team, independent of source.
It is important for insurers to leverage data to optimize the profit machines they already have, Tom Warden, former chief data officer at AIG Life and Retirement, said in a Forbes article.
“Best practice today is to embed more predictive modeling in all parts of the value chain. You don't necessarily need big data to do that. You need smart people, a disciplined approach and a culture of cooperation to monetize data-driven insights,” he said. In other words, you need DataOps.
Best’s Review columnist Gates Ouimette is founder and principal of ITconnecter. He can be reached at firstname.lastname@example.org.