# Debunking the Myths: Data Governance in Tech Companies
Written on
Understanding the Core of Data Governance
What is the cornerstone of a robust, contemporary data organization? I recently posed this question in a LinkedIn poll to my professional network. The options I provided included Data Operations, Data Analytics, Data Strategy, and Data Governance. While these categories may not encompass all aspects of data management, they served as a useful framework for this impromptu poll. Here’s a brief overview of each choice:
- Data Operations: This pertains to the transportation, storage, and accessibility of data assets. It encompasses a substantial segment of the data lifecycle, from creation to user consumption.
- Data Strategy: While many view this as merely outlining an organization’s vision and mission regarding its data journey, I believe it should extend to include core values, performance metrics, and incentives. This broader perspective differentiates it from the common approach, which often focuses only on vision and a few strategic themes. (Note: I plan to delve deeper into this topic in a future article.)
- Data Analytics: This involves utilizing data to generate business value through various analytical methods. For simplicity, I consider everything from dashboarding and reporting to AI modeling under this umbrella, although some data preparation activities may align more with Data Operations.
- Data Governance: This refers to the management and oversight of the data lifecycle within an organization. Contrary to the common perception that it solely revolves around committees and policies, it encompasses compliance, standards, information sharing, and determining decision-making authority regarding data.
After a week of polling, I was taken aback by the results. Data Strategy narrowly edged out Data Governance as the top priority for building a strong data organization, with Data Governance coming in a close second, which was a pleasant surprise.
The LinkedIn poll results indicate a surprising preference for Data Strategy over Data Governance, despite my initial expectations that Data Analytics and Data Operations would rank higher. Personally, I believe that Data Governance is paramount among the four options presented, as it fundamentally shapes how organizations make decisions about data.
On any given day, large organizations may make hundreds, thousands, or even millions of data-related decisions. These range from end-user interactions to overarching data enterprise management, and I can assure you that the volume is significant.
As I scroll through my feed, I encounter advertisements from tech companies promoting data catalogs, access solutions, consulting services, and cloud offerings. These ads often suggest that they can eliminate the governance challenges associated with managing organizational data.
However, there’s a significant flaw in this narrative: many tech companies are misleading you about Data Governance.
Data Governance fundamentally revolves around decision-making concerning data, which are primarily people and process issues—not technology challenges. The belief that Data Governance can be entirely centralized and resolved through a singular tech solution is what I term "Data Communism," and it’s detrimental.
Exploring Governance Structures in Data Organizations
Not every data decision is dictated by a top-down approach from the Chief Information Officer (CIO) or Chief Data Officer (CDO). Hierarchical structures are just one method of organizing individuals to manage data. Below are various governance structures often referenced in data management discussions:
- Market Structure: In this model, independent entities compete to deliver value. Coordination costs are low, flexibility is high, and control over each actor is minimal. Relationships tend to be short-term and transactional, such as commercial data purchases bound by software licensing agreements.
- Bazaar or Open Exchange: Similar to the Market model, this structure sees independent data enthusiasts collaborate towards a common goal. Incentives are rooted in cooperation rather than competition, leading to more innovation. Relationships can be long-lasting, as long as the shared objective persists, exemplified by various open data communities.
- Hierarchy: In this structure, individuals depend on one another, with relationships being more rigid. Control is high, and information typically flows in a linear fashion. This can lead to bureaucratic inefficiencies and the potential for "Data Communism." Traditional CIO, CDO, or enterprise data functions often fit this model.
- Network Exchange: This model unites interdependent actors for mutual benefit, characterized by lower coordination costs and a moderate degree of flexibility and control. Examples include matrixed team structures and interagency working groups.
These structures illustrate how individuals organize themselves to make data-related decisions, whether autonomously or with intentional design.
For data leaders, managers, and engineers, I encourage you to observe the execution of data tasks within your organization. Map the teams involved in these tasks to the aforementioned governance structures, noting how people are organized to implement processes.
The emergence of these structures underscores that data decisions are primarily driven by people, followed by processes. Once a process is well-established, technology can be employed to automate and enhance it. It's crucial to understand that a tech solution is not a panacea for data governance challenges.
Chapter 2: The Reality of Data Governance in Tech
The first video, Why Tech Companies Are Lying to You About Jobs, discusses the misconceptions surrounding the job market in tech and highlights the importance of understanding data governance accurately.
The second video, Don't Lift and Shift: Data Governance to AI Governance with Karen Meppen, provides insights into the transition from traditional data governance to AI governance, emphasizing the importance of adapting governance practices in the digital age.