Are We Doing Data Governance Backwards?

Image credit: iStockphoto/Birdlkportfolio

Mention data governance, and people tend to think about compliance and potential barriers to innovation. It makes the concept a difficult sell at a time when companies are looking to innovate for survival. Business leaders who are used to evaluating projects based on potential revenues will also find it challenging to understand the actual value — akin to the challenges that face security management practices, where one only sees its value when it fails.

However, the value proposition for good data governance is changing. As regulations become stricter, consumer data awareness increases, and new data-dependent projects like AI become mainstream, data governance can unlock significant organizational value.

A moderated CDOTrends virtual roundtable, organized in partnership with Informatica, explored these issues. Participants included:

●      Paul Loke, chief innovation officer at the Accountant General’s Department

●      Joshua Au, head of data center at A*STAR

●      Celine Le Contonnec, chief innovation officer at the Bank of Singapore

●      Steven Ng, head and senior manager for data center and cloud services at BASF South East Asia

●      Benjamin Kei, regional head of IT at CGS-CIMB Securities

●      Ajay Shah, director and head of ASEAN TTS, client operations, at Citibank Singapore

●      Elvin Lee, vice president for IT at Citibank Singapore

●      Zhang Songhua, head of legal compliance-data science at DBS

●      Lee Giok Leng, head of IT at GuocoLand Group

●      Jon Teo, healthcare and data governance specialist for APJ at Informatica

●      Priti Jauhari, head of supply chain technology at Johnson & Johnson

●      Charmayne Chung, head of strategic digital alliances at Maybank

●      Raymond Leong, senior manager and data analyst at Tan Chong Group

The discussion was held under Chatham House Rule and offered four critical insights:

1. Governance holes create data fragmentation headaches

Data growth is exploding both locally and on the cloud but the growth is not even. It creates silos and gravity wells, leading to what the industry calls data fragmentation. The faster the data growth, the wider the data fragmentation gaps become — made worse with mergers and acquisitions.

Data fragmentation adds other risks, especially data privacy risks. Without having proper privacy controls in place, companies are exposed to strict data privacy law actions.

A company-wide data governance framework can help to close the data governance holes. However, participants admitted that it takes top-down and bottom-up approaches.

Companies also need to expand the definition of data governance from being compliance-driven. Although risk management and compliance departments are often its first proponents, all departments benefit from a robust framework. Whether they see it as such is where the challenge lies.

In the opening presentation, Jon Teo, regional data governance specialist at Informatica, shared that modern data governance platforms already assume a fragmented IT landscape. These approaches leverage near-universal connectivity to establish an enterprise-wide catalog of data assets, which forms the foundation for holistic data governance efforts.

Legacy data can pose another problem, as it may not be ready for data governance and require considerable effort to make it ready. Spending money on legacy data governance can be a tough call at a time where many companies need to operate lean.

Regardless, all participants agreed that it is a problem that all companies need to address. Some participants saw opportunities in leveraging automation and new tools to reduce the effort and time-to-value.

2. Data sharing and data lineage is becoming a critical factor

Participants noted that with data becoming critical for rapid decision making, the ability to understand where the data comes from, the quality of the data, and its impacts on other operations become crucial. It is one reason why data lineage is becoming a hot-button topic.

However, most participants noted that in large organizations, data lineage can be complex. Achieving a dynamic, transparent understanding of how data flows within a large organization is labor-intensive and incomplete due to a hybrid IT landscape. A multi-cloud world makes having a unified view of data also difficult. Others opined that having good data landscape visibility can be difficult due to organizational reasons such as staff turnover and poor documentation.

Some departments, who see themselves as data owners, are also reluctant to share data for various reasons, ranging from privacy fears to competitive behaviors. Conglomerates have it worse; they may have competing businesses who may not value sharing data with other companies. Fostering transparent data sharing while adhering to the necessary safeguards is not simple to implement.

Yet, it is an issue that needs addressing. One way, participants noted, is to create data quality-focused committees and centers of excellence. It allows different stakeholders to understand the value of data sharing and data hygiene. Top-down commitment by the chief executive officers and the boards is also essential to break down the departments’ walls.

One participant also argued that companies need to see data governance as a long-term commitment and make it part of their DNA. Another highlighted how his company ensured strong adherence to data governance practices with AI assistants reminding data users.

Participants concluded that having a robust data governance program and a flexible data platform can address these concerns. It can also help scale governance to address different stakeholders’ needs over time.

3. Fostering data culture is a vital prerequisite 

All participants noted that to modernize data governance, you need the right culture. They said that current data governance initiatives sometimes fail because the culture is not ready. Some data owners may also resist governance efforts if they do not see the value or are afraid of losing control over their data.

Teo highlighted companies taking a fresh approach to building a data culture.  He provided Informatica’s work with the Bank of Ireland as an example. The bank’s chief data officer focused on clear, engaging communication to make the business value of good data clear to everyone. These cultural efforts helped to break down negative perceptions and encourage better data ownership amongst the business functions.

Other participants noted that vital awareness programs and internal campaigns can help get the company culture ready for strong data governance, similar to how companies reinforce their organizational values.  Along with cultural change, there is also the need to increase data literacy as data use and exploration become increasingly democratized.

With the right culture, implementing a robust data governance framework will be easier. Participants agreed that culture is the best starting point for data governance — and not the other way around.

4. Stop looking for the magic formula; do governance now

Participants advised that companies need to address data governance issues. There is no reason to wait for a blueprint as each company has its own challenges and issues to address — from its platforms and organizational structure to the company culture.

Participants noted two approaches to data governance that are polar opposites: having a pool of usable data for users governed by a robust framework and having the users’ apps directly “pull” data from departmental data stores governed by robust frameworks.

Both approaches are idealistic and take time to implement. Instead, one participant suggested taking a hybrid approach where critical data in a standard data layer is ready for analysis. It can then be scaled to the rest of the organization.

Whatever the final approach, participants agreed that the traditional practice of creating a robust data governance framework after the data is created is not going to work. Companies need to get more proactive in deploying data governance platforms and frameworks, prioritizing high-value use cases and datasets.

They also noted that having the right data governance is becoming more urgent. And with data-dependent initiatives like AI and real-time IoT-based analytics rolling out, it becomes vital for companies to get their governance right quickly — or face a host of data-related challenges soon that can derail their digitalization ambitions.

Winston Thomas is the editor-in-chief of CDOTrends. He is always curious about all things digital, including new digital business models, the widening impact of AI/ML, unproven singularity theories, proven data science success stories, lurking cybersecurity dangers, and reimagining the digital experience. You can reach him at [email protected]

Image credit: iStockphoto/Birdlkportfolio