To democratize data and analytics is to make them available to everyone. It is an admirable goal and one with its roots in the earliest days of the self-service movement. If an organization is to truly be data-driven, it follows that all key decisions—from tactical operational priorities to strategic vision—must be data-informed. So where is democratization going wrong?
Self-service tools—from early OLAP tools to modern, sophisticated data visualization suites—aim to make the creation of insight easy and intuitive. This trend continues today with the advent of augmented analytics.
Augmented analytics applications utilize AI in numerous ways: using natural language generation to explain the method and conclusions of an analytics model in plain English (or your language of choice), to monitor and automatically visualize emerging trends of interest, or even to automate the selection and tuning of the appropriate AI algorithm based on the characteristics of the input. These are all worthy endeavors which make the creation of insight more intuitive and accessible to a broader array of consumers.
Herein lies the rub: In this context, democratization refers to providing the ability to create insight to everyone. This implies that everyone is interested and, indeed, motivated to seek out their own data-driven insights in the context of their day-to-day work. This is simply not the case and sets an unrealistic expectation for how employees may broadly interact with data and analytics. Most people aren’t interested in self-service in this context. They are, however, interested in having targeted information delivered when and where it can be acted upon.
Therefore, here is a better definition of democratization: the ability to make insight available to everyone. This expansion is critical as it speaks to how insights are pervasively deployed in data-driven organizations. Specifically, analytics is pervasive when analytically derived insights permeate and support all aspects of business decision making, and when analytics systems are deployed as an integral part of key business processes.
From this viewpoint, democratization ensures that self-service expands to incorporate the delivery of analytics outputs (aka insights). In other words, a comprehensive self-service strategy must include the routine, reliable delivery of focused, discrete insights to employees just in time in the context of their daily tasks. This could be accomplished by presenting an executive dashboard with key emerging strategic indicators, providing a targeted list of next best actions to a customer service agent, or highlighting areas of potential machine failures in a heads-up display to a technician wearing virtual reality goggles.
This expanded view of democratization further requires that the purview of analytics and data governance programs include the following:
- Data literacy programs that incorporate education on how to properly interpret and apply analytics recommendations, metrics, or insights surfaced to employees in the context of their day-to-day work, as well as encouraging employees to identify insights that would further their individual work efforts.
- Analytics outputs—including inferred information such as customer behaviors, preferences, and such—that are subject to the same stringent privacy and usage policies as raw data assets.
- Incorporation of data-driven insights as an integral part of business processes and applications that is a fundamental component of business process design. Not only does this embed analytics within core business operations but it encourages employees to think critically and proactively identify opportunities to utilize data more pervasively.
Democratization is not dead. Nor is it a flawed endeavor. But without expanding the definition of democratization beyond the mere means of analytics production (analytics and data tools) to consider access to the insights produced, there is the risk of missing the largest population that exists within any organization: the analytics consumer.