If the objective is for organizations to treat data as a strategic asset, if doing so is the ultimate means of increasing ROI on data-driven investments, then the means of doing so must be by readily sharing data—across colleagues, business units, and even enterprises.
The benefits of doing so are beyond dispute. According to Gartner, in less than two years firms promoting data sharing will outperform those that don’t on most business metrics (1). Additionally, businesses that engage in inter-company data sharing with partners will create three times greater measurable economic benefit than companies that don’t, while analytics and data teams are twice as likely to produce measurable benefits from sharing data externally than they would otherwise.
The counterpoint to the acknowledged esteem of sharing data is the varying data governance, data privacy, and regulatory compliance complexities that would seem to make doing so inherently risky. Failure to adhere to protocols for any of these concerns could result in pricy litigation, penalties, and loss of reputation, all of which could considerably undermine the gains of data sharing.
Today, however, organizations can not only readily exchange data across users and companies, but also do so in way that redoubles their ability to follow mandates for governance, privacy, and regulations. The solution is a delegated data governance model with a distributed architecture for what Privacera CEO Balaji Ganesan termed governed data sharing, which allows firms to “share data as they would share Google docs.”
The difference is they can do so across the tools and platforms of their choice with centralized data governance policies embedded in each of those source systems. Consequently, they can now reap the above data sharing advantages without the risks of doing so in a manner defying governance or regulations.
Centralized Governance Policies The fulcrum on which governed data sharing hinges is the capacity to create centralized policies that are enforced no matter where data—and access to that data—happen to be. Solutions specializing in this space deliver organizations “an easier way to create policies,” Ganesan remarked. “Instead of writing policies in 10 different databases and 10 different applications, which is how DBAs used to write policies by writing code or SQL statements, you can abstract it [and] make a unified UI layer where all through APIs you can write to those tools in a fairly simple manner.”
With this approach, organizations don’t have to manually code data governance, data access, and security policies into each respective application (like Databricks) users want to share data.
