Micro Focus is now part of OpenText. Learn more >

You are here

You are here

How to Balance Data Governance with Data Democracy

Vishal Singh Head of Products, Starburst Data
Photo by Piret Ilver on Unsplash

Harnessing data—and harnessing even more data—creates competitive advantages for businesses. It also creates a plethora of new challenges for IT leaders and teams. Data sprawl is rampant across business units, partner organizations, clouds, and even geographies. This leaves organizations on the defensive when it comes to deriving business value from data while complying with an increasing number of security and data-privacy regulations.

Businesses today have strong motivation to deliver high-quality customer service and experiences, using the data at their disposal to do so—as well as to maintain the privacy of consumers and partners, in lock-step with modern regulations. This means that while data democratization and data governance appear to work in opposition, organizations must strike a balance between them.

Defining Data Governance and Data Democratization 

Today, more data exists than ever before—and most organizations today seek greater access to their data. This has led to increased interest in the philosophy of data democratization—in other words, enabling the average end user to become a data consumer able to quickly and securely access data.

Data democracy will likely increase in popularity as data-consumption demands become more widespread among nontechnical users. Recent research from Boston Consulting Group (BCG), Red Hat, and Starburst Data suggests that data demands from nontechnical users is expected to increase over the next three years.

Yet at the same time, data security, privacy, and compliance requirements necessitate careful examination and control of who has access to what. So while the mandate of data democratization is to provide more access, it must be properly governed. 

Data governance—comprising internal standards, procedures, and frameworks—contributes to supporting more efficient workflows, maintaining legal compliance, ensuring consistency, and (critically, for our purposes) deriving business value from data. As such, data governance is a vital component of any data strategy. With an effective data-governance strategy—which includes strong leadership as well as clearly defined goals, policies, and access controls—organizations can improve efficiencies and build credibility with customers and partners.

Balancing Data Governance and Data Democratization 

Data democratization is important to an organization because it ensures an effective and efficient method of providing all users, regardless of technical expertise, the ability to analyze readily accessible and reliable data to influence data-driven decisions and drive real-time insights. This eliminates the frustration of requesting access, sorting information, or reaching out to IT for help. The organizational incentives are clear: functional business units are able and empowered to make informed data-driven decisions on their own—and, in so doing, boost productivity across the enterprise. 

BCG's research (referenced above) suggests that for larger companies with a more mature data stack, the total number of unique data vendors has nearly tripled in the last decade—up from about 50 to over 150 today. This means, first, that engineers are overwhelmed with complexity and, second, that data consumers don’t know where or how to find the data they need.

This is where effective data governance can work as an enabler of access rather than a deterrent. Indeed, effectively balancing data democracy with governance offers a solution to this problem by providing the best of both worlds—assurance that your data is being used correctly while driving data-driven outcomes across your organization. 

The inherent nature of data governance is to provision and control access to data. While data democratization seeks to provide more access to more data users across the organization, organizations need to reexamine their data-governance framework so that it is aligned with democratization but doesn’t lead to overprovisioning. It is essential throughout the entire process of democratizing data—including collecting, storing, accessing, and interpreting data—that data governance acts as an enabler and guide for data democratization.

Whereas data governance means more control, data democratization means more freedom. While seemingly at odds with one another, both are crucial to success. Here are a few tips for balancing the two. 

From Single Source of Truth to Single Point of Access 

From mainframes to client servers to the cloud, companies have long struggled to build and maintain the elusive "single source of truth." Hundreds of millions of dollars have been wasted just for organizations to find themselves in the same position as before. Data is everywhere, engineers are spending their valuable time just moving it around, and analysts and data scientists are twiddling their thumbs waiting. It is also a challenge to govern data while ensuring organization-wide access to insights created from distributed datasets.  

The solution to this problem lies in data federation, which makes data from multiple sources accessible under a uniform data model. This model acts as a "single point of access" such that organizations create a virtual database where data can be accessed where it already lives. This makes it easier for organizations to query data from different sources in one place.

With a single point of access, users can go to one location for searching, finding, and accessing every piece of data your organization has. This will make it easier to democratize data access because you won’t need to facilitate access across many different sources. It will also open new analytics possibilities while arming your data consumers with opportunities to increase the speed to insight and overall productivity.

Moreover, data-governance standards can be established at a single point of access, making it easier to manage fine-grained access controls at a single location—as opposed to managing those controls at each of the many places data is stored. 

Empower Data Consumers with Data Literacy 

While the promise of democratizing data is exciting, there are concerns that organizations must address to feel confident using data—as well as meeting data-governance standards surrounding accessing and trusting data. For example, according to BCG's research, only 45% of respondents said their company promotes data literacy among all employees. For data democratization and governance to work effectively, data literacy needs to be embedded throughout the organization so that everyone can learn, access, and understand data to leverage it for data-driven business decisions.  

Data literacy is the ability to read, work with, analyze, and argue with data. Increasing data literacy will not only help your business make data-driven decisions faster by removing barriers to understanding the data, but it will also ensure that you can trust the insights that emerge. Empowering your data consumers with data-literacy skills is essential to both data democratization and data governance.

There are a lot of steps you can take to increase data literacy, but evaluating your organization’s current data-literacy level is a good place to start. Identify the fluent data speakers and the gaps. People who are data-literate can help educate and train others to bridge the gap between data analysts and nontechnical teams.

Once you’ve identified where the gaps are and stressed the importance of data literacy, offer resources to all employees so they can develop their skills. Training, such as in spotting missed opportunities with data, can provide ongoing support for employees throughout their data journeys.

It's also critical to communicate to your entire organization why data literacy is important and should be prioritized. One way to do this is by showing employees how your organization uses data every day to make decisions and support operations.

Prioritize Data Quality 

Data literacy cannot be complete without data quality—that is to say, whether data is accurate, complete, consistent, reliable, relevant, and up to date. If data is of poor quality, it can lead to inaccurate conclusions and ineffective decision making—even if the person analyzing the data is highly data-literate. To make the most of your data, it is important to ensure that it is of good quality.

This is another area where data governance comes into play. You can improve data quality by establishing clear standards, developing specifications for what good data looks like such that it can meet business needs, and setting benchmarks for collecting, storing, and maintaining data. This should be done in addition to regularly monitoring data for accuracy, completeness, consistency, and timeliness. 

Organizations should also ensure data transparency by exposing data's limitations and quality to stakeholders—while collecting continuous feedback from them. This creates a continuous feedback loop and allows quality metrics to incorporated into business performance. By having good quality data and data consumers with data-literacy skills, you can ensure that your organization is making informed decisions that are based on accurate and relevant information.

Hire Data Product Managers

Data products—high-quality, curated, discoverable datasets enriched with metadata—empower teams to take ownership of their data. Data products offer greater data accessibility because they offer a more consumable version of the data. This leads to more people being able to derive value from that data—even if they don’t have a technical background.

While data products are more quickly and easily consumable, the data-product model relies on data owners to ensure proper governance in the form of fine-grained access controls. These access controls are necessary to help protect data (and are required to a certain degree to comply with regulatory requirements). Relying on data owners to implement these controls, however, can get messy and lead to overprovisioning without centralized oversight.

In the absence of a centralized data store, data-product managers can establish and standardize these controls at the domain level, ensuring that each domain is aligned on the governance framework.    

Data product managers and traditional product managers have a lot in common. The general duties of data product managers start with defining product strategy and direction for data analytics, data knowledge, and the data-science platform. From there, they work with data scientists, engineers, and designers to develop and curate data products that are easily consumable and drive business value. They also define requirements, prioritize user stories, and manage product launches by coordinating cross-functionally with other teams (such as marketing and sales).

Data product managers are also responsible for monitoring the performance of data products by tracking metrics such as performance, usage, retention, and satisfaction. They then use those insights to improve the data product continuously. As teams work remotely and move quickly, data-product managers provide the necessary governance to facilitate who has access to data and to what extent.

Managing the complexity of today's data environment is already a tall order, so it can be overwhelming to add in the responsibility of balancing data democratization and governance. However, with the right approach, you can lay an effective foundation for balancing the two—and thus empower data consumers so that your organization is up to the challenge.

Keep learning

Read more articles about: Enterprise ITData Management