Data Governance and Its Importance in the Modern Data Stack

Origins of data governance

To understand the origins of data governance, think about how governments have evolved over the course of time. Initially, members of our species were organized into tribes or clans. Through conquests, these groups formed monarchies with a leader at the top. As society evolved, kings or queens began establishing an aristocracy to provide support to their leader. But monarchical and aristocratic societies are not egalitarian, and so democracies evolved to give governance power to all within a society.

The result? A shared, collaborative approach to government. While societies cannot exist without governance, governance cannot be successful without input and discussion from everyone involved.

Any community needs good governance to thrive — the same is true for a data community within an organization. Traditionally, data governance has been approached from the top-down. A select Data Governance committee (think the monarch and aristocrats) would tell all other data users how data is to be managed, accessed, and used. But data governance has to evolve to become more democratic, which, in turn, increases process adoption and data usage.

The current state of data governance

The traditional top-down approach has earned data governance a reputation as being controlling, bureaucratic, and, oftentimes, boring. Historically, companies only enacted data governance for the sake of compliance. Regulations like GDPR describe how data must be secured and handled, and so organizations felt compelled to invest in data governance initiatives out of fear of violating the law, subjecting them to heavy fines.

This has resulted in a very unidimensional approach to governance; nothing more than a way to comply with regulations. The tragedy is that data governance has the power to be so much more – a way to maximize the value of data assets. This is discussed at length by Prukalpa Sankar, in her blog, which talks about how data governance is having an identity crisis; she writes, “At its heart, data governance isn’t about control. It’s about helping data teams work better together.”

Key challenges in modern data governance. Source: Gartner

Future of data governance

We must change the perception around data governance in order to unlock its full potential. The way to do that is to think about data governance less in the context of control, and more in terms of data enablement, empowering organizations to achieve business objectives through the use of data. When applied properly, data governance creates better data teams through collaboration. And collaboration is at the heart of true data democratization.

Key differences between traditional and modern data governance. Source: Gartner

Principles of modern data governance

Modern data governance requires a shift in thinking to a new approach designed for today’s data environment. Consider making these three adjustments when imagining data governance at your organization.

1. From data governance to “data and analytics” governance

According to the Open Group, data is an asset with real, measurable value. It is essential and indispensable in helping decision-makers form critical, accurate, and timely decisions. As such, data must be shared, easily accessible, and well-managed. In short, data must be governed — but so too must analytics.

Data assets go beyond raw data, dashboards, and models to include analytics. Analytics also have real, measurable value, and, as such, must be similarly governed.

A demo of Atlan's "data & analytics" governance capabilities

2. From a centralized to a community-led approach

We mentioned above that data governance has traditionally been executed using a top-down, or centralized, approach where a data governance committee tells an organization how data should be handled, managed, and leveraged. The problem with this is that across an organization, different teams will have different relationships with data.

Your product team and marketing team might use the same data sets, but they don’t necessarily use them in the same way. The product team might be using customer survey results to decide which product features to build, while the marketing team could be using the same data to decide which keywords to target for SEO. With a data culture that takes a community-led approach to data governance, both teams will have the opportunity to provide input on how that data is categorized and pulled to fit their unique use cases.

A collaborative approach is one in which the processes, policies, and roles that make up data governance are crowd-sourced. Obtaining input from your people makes them more likely to actually adopt the structure that you build. This is what we mean by collaboration leading to greater data democratization.

Embedded collaboration is the key to modern data governance. Source: Atlan

3. From an afterthought to a part of daily workflows

Shift from thinking of data governance as an additional step in your operations. Instead, think about it as something that’s integral to your organization’s daily workflow. You can start building insight into how to accomplish this by looking at your social metadata.

Social metadata provides information on how people are using data assets. It reveals which values, tables, codes, or models are being used most and least. If a critical set of values aren’t being used, then perhaps your people don’t have access (or visibility) to this data — or perhaps they don’t understand its worth. Knowing this, you can address the issue so that your teams will start incorporating these data elements into their daily workflows. Ultimately, this will lead to greater data governance engagement and better decision-making.

Data governance that's always on, intelligent, and action-oriented. Source: Atlan

A Guide to Building a Business Case for a Data Catalog

Download free ebook

Structuring data governance programs

Centralized data governance lacks feedback from teams, and, thus, data is unlikely to find mass adoption. Decentralized data governance, on the other hand, can lead to a situation where there are “too many cooks in the kitchen” — in other words, chaos.

Structure your data governance program to strike a balance between centralization and decentralization — between autocracy and democracy. There’s a reason why democracies in the modern world are usually structured as republics. Having citizens vote directly on every matter is an inefficient way to run a community. The same is true for your data community.

The ideal data governance program encompasses your people, processes, and technology. It’s one where data is viewed as an asset, and is readily visible and available, engendering trust and greater adoption. People can use it to collaborate and make well-informed, timely decisions, spurring innovative solutions.

Data governance: Best practices

Establishing a data governance program might seem like a heavy lift, but we’ve culled some tips to help you get yours off the ground. Here are eight best practices for implementing a solid data governance program:

1. Establish measurable goals for your organization
2. Understand what needs to be done to achieve compliance with regulations
3. Get buy-in from C-suite executives right from the start
4. Implement a robust data governance framework
5. Pinpoint every data domain
6. Provide training to all employees on the importance of data governance
7. Document progress
8. Invest in technology to support your data governance program

To read about these best practices in more detail, take a look at 8 Best Practices for a Robust Data Governance Program.

Data catalogs are going through a paradigm shift. Here’s all you need to know about a 3rd Generation Data Catalog

Download free ebook

Data governance: Framework

The right framework can set you on a path to data governance success. Here are six frequently referenced data governance frameworks:

1. DAMA DMBOK: This framework imagines data management as a wheel with data governance at the center (the hub) surrounded by nine knowledge areas.
2. The DGI data governance framework: This framework is made up of ten universal components that address the why-what-who-how of data governance.
3. McKinsey: This framework will have you reimagine your organizational design to include three core components: A data management office (DMO), domain-based roles, and a data council.
4. Eckerson: This framework places people at the heart of data governance by defining roles and outlining responsibilities.
5. The SAS data governance framework: This framework illustrates how organizations value their data using effective data management and governance practices.
6. PwC enterprise data governance framework: This framework takes conventional models such as DAMA DMBOK and DGI to the next level by accounting for next-gen data landscapes.

Be sure to spend some time reading this article to learn in-depth details about each of these frameworks and how they might bring benefits to your business.

Data governance framework: DAMA DMBOK. Source

Data governance: Examples

Here are a few examples so that you can see data governance in action.

Spotify

Spotify’s rollout of its popular Discover Weekly feature is a great example of good data governance. The world’s largest music streaming service established processes and procedures for handling listener music data so that it was in compliance with strict European privacy laws. This robust data governance program assured regulators and subscribers that information on listening habits were being used responsibly. Spotify’s data scientists then had latitude to use the data in creating a recommendation algorithm to give subscribers a more enjoyable experience. Now, every Monday, subscribers are treated to a Discover Weekly playlist of recommended songs they might like based on their listening habits.

According to O’Reilly, “Discover Weekly illustrates how data, properly governed, can create a well-loved brand and change the power dynamic in an entire industry.” It’s win-win-win for listeners, musicians, and Spotify.

CSE Insurance

CSE Insurance is a leading provider of property and casualty insurance across the United States. Despite having won several product awards, CSE had some major data governance challenges:

• Siloed data: Data was scattered across disparate sources.
• Mismatched metrics: Basic definitions & metrics were warped among users.
• Migrating documentation: Tableau governance efforts had to migrate seamlessly.

What CSE needed was a single source of truth powered by data governance policies and procedures that would ensure data was accessible and usable by the teams who needed it. They decided to adopt Atlan’s platform and it became the foundation of the company’s award-winning data-culture.

Financial Institutions

Financial institutions are heavily regulated and must comply with a litany of rules including on how data is managed. Data governance programs can do more than help these institutions stay within the boundaries of the law. Benefits include:

• Timely regulatory reporting adherence
• Increased market share
• Overall reduction in IT spending due to information and resource rationalization
• Improved analytics to gain a competitive advantage.

Financial institutions that implement strong data governance programs will achieve results beyond just following regulations — they build a foundation for business growth and expansion.

Data governance – the key to unlocking the potential of diverse modern data teams

Now is the time to shift the perception of data governance from control to collaboration. A proper data governance program doesn’t restrict team members but involves them in the process of identifying the proper tools, protocols, and workflows that will give them the freedom to discover, understand, trust, and drive greater value from data.

Go deeper by exploring the other components necessary to build a modern data stack.

Data Governance with Atlan

If you are evaluating and looking to deploy best-in-class data access governance for the modern data stack without compromising on data democratization? Do give Atlan a spin.

Atlan is a Third-generation data catalog built on the premise of embedded collaboration that is key in today’s modern workplace, borrowing principles from GitHub, Figma, Slack, Notion, Superhuman, and other modern tools that are commonplace today.