Data governance is, strangely, very challenging. Challenging because it requires significant cultural and organizational alignment, arguably more soft skill work than technical execution. “Strangely” because there is no rational conclusion, where performing data oversight is optional. Yes, it is difficult, but organizations will dig in and tackle the hard things once that thing is accepted. Data governance is frequently not accepted as mandatory. For an organization to say, “Data governance is hard right now, so we will defer it,” is simply kicking the can down the road. To delay or refute because it is hard, or has shown a poor track record, is merely making excuses.

To be clear, oversight and management of the data fuel of our world is certainly a fundamental capability. Certainly, AI is well-understood to be high-risk and requires consideration of what you feed the model. Data governance, or whatever you want to call it, is a required capability.

Therefore, it is widely pursued, and there remains a strong demand and seeming justification for it. Yet, all available data – both qualitative and quantitative – suggests it’s failing. Recent work has been done to better understand this seeming contradiction: Why is something so necessary handled so poorly? What are the root causes? 

A recent article looking at the influencing forces for data governance, published by several practitioners (Thomas Redman, John Ladley, Mathias Vercauteren, Malcolm Hawker, Anne Marie Smith, and Aaron Wilkerson), identified several restraining forces that consistently impede data governance success:

• Knowledge gaps and lack of training
• Overreliance on technology
• As deployed, rarely works as expected or was overpromised
• A fundamental misunderstanding and a lack of agreement on the definition and scope of data governance
• Over-advertised as easier and more effective than reality
• Structural and organizational issues
• People-related challenges
• Difficulty in quantifying business benefits

These are countered by driving forces that urge data governance forward:

• Evolving regulations mean consistent demand
• External pressures and competitive demands
• Established resources
• Supportive data professionals
• Business stakeholder awareness

Our initial advice was to build on the driving forces; this article will focus on the restraining forces. We will explicitly look for the root cause of the restraints and suggest what actions can be taken.  

Root causes

The first step was a thorough root-cause analysis, utilizing the “5 Whys” technique. We made no assumptions and rigorously questioned our assertions, often to the point of absurdity on multiple occasions. This was done to maintain as much of a brutally objective stance, given the inherent limitations of any qualitative assessment. 

Our analysis highlighted 5 key themes when we felt we were getting near a root cause. These are: 

• Process (mechanistic vs organic)
• People and leadership
• Incentives and organizational  structure
• Lack of value quantification
• Outdated best practices

We cross-referenced the 5 themes of root causes with the restraining factors and deeply investigated each intersection to ensure a logical consistency between our initial observations and our proposed root causes.

We took the intersections and asked “why?” 5 times for each. Briefly, this resulted in a list of N root causes, which were then grouped into five areas. From the N root causes, we then boiled them down as far as we could.

Our results were interesting. There is a strong case to be made for a single area of root causality, though it is expressed in different perspectives. So, your authors submit the following: if we address this one area across three perspectives, we can finally ensure data governance success.  

The primary root cause is that organizations, leadership, governments, etc., do not understand that data should not be managed as an “IT thing.”

This is not new in the data field. It is foreign to any leader outside of the data field, including CIOs. There is a massive deficit in understanding the risks and the foundational presence of data within modern organizations. 

To elaborate, data is an organization or enterprise asset. That has also been hashed out.  This can mean many things, but in all aspects of the “asset” conversation, it means more discipline than organizations are willing to grasp. But data is not an asset that can be audited once a year, like office supplies. It is woven into the fabric of every activity and requires “supply chain” treatment.  It affects processes, people, business directions, cultures, and even society. It is far beyond an element of “Information Technology.” 

Dr, Tom Redman of our group stated it thusly in 2012 – “As companies devote increasing time and energy in gathering massive quantities of data, many neglect a critical first step: Get most responsibility for data out of the IT Department.”                                                      

Our colleague Doug Laney, author of the 2017 book Infonomics, describes it thus: “Separate the Information from the Technology.”  

Simply put, data is not a technical subject to be delegated; it has a core part in organizations (and society, for that matter) and requires enterprise-level strategic thinking.  Many practitioners in the data profession have emphasized this for decades. However, we never considered it as a root cause. It was always a preventative measure or a response. It is really a fundamental error to push data down the organization in terms of recognition and authority.  

This leads to the three perspectives that plague modern organizations:

1. Acknowledged but not engaged: Data governance is delegated too far down and isolated from core line capabilities. For example, financial auditing and GAAP are accepted as essential and have Board-level oversight. Data governance, which is conceptually identical, does not. 
2. Lip service over leadership: Leadership fails to embrace the real value and importance of DG. They refuse to demand data competency from their teams or ask for clear statements of value from their subordinates. We settle for  ‘buy-in.”  We rarely have true engagement. To be clear, data governance has value; it is the valuation and reporting of that value where the challenges lie.
3. The competency gap at both ends: At the top, leadership lacks the foundational data literacy required to ask the right questions and provide the right guidance for the data supply chain. At the opposite end of the organization, the technology staff hesitates to engage in enough business acumen to be able to understand what a quality product should look like.

The fallout from this is well known, and all the constraints to data governance can be tied back to this:

• Knowledge gaps and lack of training: Ignorance equates to no desire to learn. Too often, leadership declines to learn about data “things.” How can this be a standing practice when, in the following breath, leadership says, “We need to be data-driven”?
• Overreliance on technology: It is easier to throw the onus and blame on non-human technology. The treatment of data as a “technology play” means people at lower levels are asked to do what they know, which is buy and deploy technology.  Yet the largest constraints are human and cultural, not technological.     
• The expectation gap: Over-advertised as easier and more effective than reality and, as deployed, rarely works as expected or was overpromised – you cannot set solid expectations if your audience does not understand what you are talking about 
• Conceptual misunderstanding: There is a lack of agreement on what data governance is and is not. The basic concepts are missing and need to be introduced to all layers of organizations. The blame is not on leadership here. It is everyone involved with the concepts. Middle management is time-constrained, and data professionals are notorious for failure to execute or engage in good communication.
• Structural and organizational issues: Data governance has not integrated aspects of basic organization design because it is not considered a fundamental enterprise capability. 
• People-related challenges: Emphasis on tech means deemphasis on people 
• Difficulty in quantifying business benefits: You need to understand the investment areas to derive a benefit. Leadership often lacks a conceptual framework for data value; they struggle to view Data governance as a value-generator rather than a cost center. 

Next steps

There are some serious aspects to the analysis. True acceptance of oversight of data requires strategic transformation. Here is our root-cause-driven list of recommended activities:    

1. Educate leadership

The era where we delegated “data stuff” because it is technology needs to end. Effective leaders are at least an inch deep and a yard wide in their organizations (except for data).

• We need the CEO to get an inch deep on data.  
• CDOs must move beyond complaining about the lack of engagement and find the messages that will resonate. If leaders will not come to the training room, then data governance teams need to send an impact analysis to the board and await their response. 
• Training needs to be delivered at a corporate level – not IT departments.  
• Lastly, this means that certain sectors of management need to hitch up their trousers and talk about hard things.     

2. Document the business value of governance

CFOs or CEOs must make CDOs and CIOs accountable for documenting the business value of governance efforts – regardless of whether the accounting profession continues to waffle on the issue of whether data will be considered an asset by GAAP standards. Incentive plans for CDOs and CIOs, reviewed and approved by CEOs and boards, must optimally show alignment between improved data maturity and business outcomes.

3. Accept that data governance is not an IT issue

This perspective must be ingrained from the board down to the lowest level of technology – and even to supply chain partners or anyone else where there is a data agreement in place. Over time, data governance must cease to exist as its own separate discipline. Instead, it must be considered as an aspect of corporate governance, reflected in the daily actions of workers and non-workers.

4. Continue work on the true value of data

The accounting bodies have said “meh” for too long. The answer may or may not be “put data on the balance sheet,” but what we do now is simply not enough. If CDOs are serious about their organizations becoming data-driven, a great first step in that direction would be to create data to show the value of what they do. A refusal to do this is both hypocritical, and rightfully, an existential risk.

5. Cover the whole data supply chain

Data governance is a standard, required business capability, but data governance must cover the entire data supply chain, not just accessing data or analytics. 

6. Bring data governance closer to the business

Data governance must be considered in the same breath as process governance or financial governance. Separate the data governance needed for analytics (which data leaders can control) from the governance needed for business process optimization (which requires business partnership). Until organizations are ready to restructure and eliminate the idea that data is separate from “the business,” data leaders must focus their efforts on what they can control while advocating for structural integration.

There is a lot more work ahead, and the devil is still in the details. If there is a challenge or some more insight, please let us know. This is a process that the industry requires. Data governance will happen. But do we choose for it to happen ugly or smoothly?

About the Authors:

Malcolm Hawker is the CDO of Profisee and is a thought leader in the fields of Data Strategy, Master Data Management (MDM), and Data Governance. As a former Gartner analyst, Hawker has authored industry-defining research and has consulted some of the largest businesses in the world on their enterprise data and analytics strategies.

Having served as a Chief Product Officer, Head of IT, and strategic business consultant, Hawker is an industry leader with over 25 years of experience at the forefront of data-enabled business transformations. He is a frequent public speaker on data and analytics best practices, and he cherishes the opportunity to share practical and actionable insights on how companies can achieve their strategic imperatives by improving their approach to data management.

John Ladley is an experienced practitioner and author of several essential books in Data Management and Governance. He helps organizations define and transition to new business and data capabilities. His work and books enable alignment of business and data strategy, organizational change, and practical application of data technology to business problems.