Why So Many Data Governance Initiatives Fail

Introduction

Today, data is unanimously recognized as a strategic asset for organizations. In light of this awareness, data governance initiatives have multiplied in recent years, driven by a seductive promise: to structure, secure, and enhance the company’s information assets. Yet the reality is stark: according to recent studies, 60 to 80% of these programs fail to meet their objectives or are abandoned along the way.

This paradox raises questions. How can a field considered crucial by almost all executives have such a high failure rate? The answer does not lie in a simple question of technical complexity or insufficient resources. It lies in an often overly narrow, disembodied, and technicized conception of what data governance should be.

Indeed, governing data is not limited to operational management or implementing specialized tools. It is a transversal and structuring function that must support all of the organization’s modes of action: strategic steering, informed decision-making, effective project execution, risk management, and operational control. It is precisely this multidimensional scope that is often underestimated or poorly understood.

The hypothesis we put forward in this article is that these failures stem primarily from a disconnection between governance initiatives and the organization’s real levers of action. Too often, data governance ends up isolated in a technical or methodological bubble, with no anchor in the decision-making and operational processes that make up the daily life of the company.

To understand this phenomenon and identify paths to success, we will examine in turn three interdependent layers of failure: conceptual pitfalls, organizational disconnections, and finally operational obstacles.

1. Conceptual Pitfalls: Governance That Loses Its Purpose

Failures in data governance initiatives often have their roots in a fundamentally mistaken conception of what this governance should be. Instead of being a lever for transformation and value creation, it ends up reduced to a formal exercise disconnected from the realities of the organization.

The Illusion of Frameworks Without Contextual Anchoring

One of the first conceptual pitfalls lies in the mechanical adoption of standardized methodological frameworks without genuine adaptation to the organization’s specific context. Faced with the apparent complexity of data governance, many companies naturally turn to established references such as ISO 38505, DAMA-DMBOK, or DCAM, hoping to find a ready-made roadmap.

This approach has obvious appeal: it offers a preconceived structure, recognized best practices, and the legitimacy associated with international standards. However, it often leads to a mechanical transposition of generic principles that fail to take into account sectoral specificities, digital maturity, organizational culture, or the strategic issues specific to each company.

As a result, governance is reduced to a compliance exercise, where formal adherence to the framework becomes more important than concrete results. Teams exhaust themselves checking boxes and filling out matrices, without fundamentally changing how data is used daily to create value. The framework then becomes an end in itself rather than a means, diverting attention and resources from the real transformation challenges.

A Focus on Deliverables Rather Than Functions

The second major conceptual pitfall is the tendency to prioritize the production of formal deliverables over the real functions that data governance should fulfill. Many programs engage in the meticulous creation of data charters, the exhaustive definition of roles and responsibilities, the development of enterprise glossaries, or the documentation of data management policies.

These deliverables, while necessary, are too often disconnected from the organization’s decision-making circuits. The data charter remains a document accessible on the intranet but rarely integrated into decision-making processes. Defined roles are not accompanied by effective responsibilities or means of action. Business glossaries, despite their theoretical usefulness, are not systematically used in projects or analyses.

This focus on form rather than function results in “paper governance,” which exists mainly in documents but struggles to take root in daily practices. Teams may feel they have fulfilled their mission by delivering these artifacts, even though their impact on project quality, decision relevance, or risk management remains minimal.

Self-Referentiality: When Governance Becomes an End in Itself

A third, more insidious pitfall concerns the tendency of certain initiatives to become self-referential. Instead of focusing on their contribution to business objectives, they develop their own internal logic, their own success metrics, and end up operating in a closed loop.

This self-referentiality manifests itself in the creation of performance indicators that measure governance activity rather than its impact: number of definitions created, training participation rates, metadata completeness percentages, etc. These metrics, while technically correct, say nothing about the actual improvement in decision-making, project efficiency, or risk management.

Organizations thus fall into the trap of symbolic governance which, lacking integration into concrete value creation processes, ends up being perceived as an additional administrative layer rather than as a performance lever. This disconnect between formal deliverables and the organization’s real needs is one of the fundamental causes of failure in data governance initiatives.

This mistaken conception turns governance into a self-referential exercise that gradually drifts away from its primary mission: improving how the organization uses its data to act, decide, and transform. To be truly effective, governance must move beyond these conceptual pitfalls and anchor itself firmly in the organization’s real dynamics.

2. Organizational Disconnection

After exploring the conceptual pitfalls, it is essential to examine how data governance initiatives often fail due to a profound disconnection from the organizational realities they are meant to serve.

Governing Without Understanding Operational Realities

The first manifestation of this disconnection is the lack of understanding of concrete, day-to-day data uses within business units. Governance programs are often designed and deployed by specialized teams — data offices, IT departments, external consultants — who, despite their technical expertise, may have a limited understanding of field practices and actual user needs.

This distance leads to the creation of rules, procedures, and standards that appear disconnected, or even contradictory, to operational imperatives. Strict data quality requirements, for example, may be imposed without considering the time constraints of sales teams or the limited tools available to employees. These rules, perceived as unworkable or hindering efficiency, are then bypassed or ignored.

This situation reveals the absence of genuine dialogue between data experts and business practitioners. The former speak the technical language of data models, reference data, and metadata, while the latter speak in terms of business objectives, operational constraints, and functional needs. Without translation and mediation between these two worlds, governance remains an abstract concept for those who should be its primary beneficiaries.

The Lack of Alignment with Strategic Value Levers

A second factor of disconnection lies in the difficulty of linking data governance to the organization’s strategic priorities. Too often, governance initiatives are launched as autonomous programs, with their own objectives and metrics, without explicit connection to the company’s major directions.

This lack of articulation is evident in the inability to demonstrate how governance directly contributes to achieving strategic objectives, whether it be improving customer experience, optimizing operational processes, reducing risks, or developing new markets. Without this clear connection, governance is perceived as a peripheral activity rather than a central pillar of corporate strategy.

Executives and managers, facing multiple priorities and limited resources, naturally devote less attention and fewer means to these initiatives whose added value remains abstract or difficult to quantify. Governance thus becomes a technical subject relegated to specialists, instead of being integrated as an essential dimension of organizational management.

The Failure to Develop True Organizational Capabilities

The third aspect of this disconnection concerns the inability of many initiatives to create lasting capabilities within the organization. Beyond frameworks — whether general like DAMA-DMBOK or functional like the Data Governance Institute’s — data governance should concretely transform how the company anticipates, decides, and manages its activities.

Yet it is often observed that governance programs bring about no tangible change in these areas. Decisions continue to be made based on intuition or partial data, projects still suffer from informational inconsistencies, and data-related risk management remains fragile.

This situation is explained by a lack of investment in the three pillars essential for capability development:

• Tooling: absence or inadequacy of technologies enabling the practical application of governance principles.

• Skills: insufficient training of employees at all levels of the organization.

• Methodological support: lack of practical assistance in integrating governance principles into daily activities.

Without these elements, even the best-intentioned governance efforts remain dead letters, unable to truly transform practices and create a data culture embedded in the organization’s DNA.

This triple disconnection — from operational realities, strategic priorities, and capability development — explains why so many governance initiatives, despite promising starts, end up running out of steam without having produced the expected transformations.

3. Common Operational Obstacles

Beyond conceptual pitfalls and organizational disconnection, data governance initiatives face concrete operational obstacles that hinder their effective implementation. These challenges, though more technical and tangible, are just as decisive in the failure of governance programs.

Organizational Fragmentation and Silos

One of the most persistent obstacles lies in the very structure of modern organizations, often characterized by high functional specialization and segmentation into distinct departments. This fragmentation creates information silos where each entity develops its own practices, tools, and data reference systems.

In this context, data governance collides with established territories, divergent processes, and heterogeneous systems. Business definitions vary from one department to another — “customer” does not mean the same thing to sales, marketing, or after-sales service teams. Data is duplicated, transformed, and enriched differently depending on each team’s specific needs.

Harmonization efforts thus meet with multiple forms of resistance: fear of losing autonomy, defense of existing practices, or simple inertia born of habit. The resulting lack of cross-functional collaboration is a major barrier to establishing coherent and shared governance.

Poor Quality of Existing Data

A second major operational obstacle concerns the state of data already present in information systems. Many organizations launch governance initiatives while facing a legacy of problematic data: incomplete information, systemic errors, multiple duplicates, inconsistent formats, or missing metadata.

This technical debt, accumulated over the years, represents a considerable challenge. Governance leaders face a dilemma: should they first clean up existing data before setting new rules, or impose new standards while gradually addressing the legacy?

In either case, the magnitude of the task is often underestimated, both in complexity and resources required. Data quality improvement projects stretch far longer than planned, resulting in loss of momentum and credibility for governance initiatives.

Resource Constraints

The third operational obstacle relates to limitations in budget, expertise, and time. Data governance is often perceived as an investment whose return is neither immediate nor easily quantifiable, leading to insufficient resource allocation.

Financially, organizations hesitate to commit the funds needed to acquire specialized tools, train their teams, or secure the required external expertise. In terms of skills, the market faces a shortage of profiles combining technical mastery of data with an understanding of business issues and organizational transformation capabilities.

As for time, it is regularly underestimated. Data governance demands sustained long-term commitment, yet it is often approached as a time-limited project. Teams then face tension between their daily responsibilities and the new demands of governance, leading to minimal and inconsistent investment.

The Complexity of Technological Architectures

Finally, the growing complexity of data architectures constitutes a major operational obstacle. Modern technology ecosystems combine legacy solutions, SaaS applications, data lakes, data warehouses, and multiple analytics or visualization tools.

This technical heterogeneity makes implementing unified governance considerably more difficult. Data flows cross multiple systems, each with its own storage mechanisms, formats, and access rules. Metadata is dispersed or nonexistent, making it hard to establish an overall view of the information assets.

Governance tools themselves sometimes struggle to integrate seamlessly into these complex architectures. Features for traceability, metadata management, or quality control end up fragmented across different solutions, creating blind spots where governance cannot be fully exercised.

4. Case Studies: Dramatic Data Governance Failures

To illustrate concretely the failure mechanisms analyzed above, let us examine two documented cases that reveal the scale of the consequences when data governance breaks down.

Wells Fargo: When Lack of Governance Leads to a Financial Scandal

The Wells Fargo scandal is one of the most spectacular examples of data governance failure in the financial sector. Between 2002 and 2016, the bank created over 1.5 million fraudulent bank accounts and 565,000 unauthorized credit card accounts in the names of existing customers, without their consent or knowledge.

This systemic practice led to a $3 billion settlement with the Department of Justice and the Securities and Exchange Commission in 2020 — one of the largest fines of the Trump administration. The scandal also prompted the resignation of CEO John Stumpf and an in-depth investigation into the bank’s business model.

Root cause analysis reveals a total breakdown in data governance at several levels:

• Disconnection from business objectives: The bank had implemented aggressive, unrealistic sales targets, symbolized by the slogan “eight is great” — aiming for eight Wells Fargo accounts per customer. Senior Community Bank leaders refused to change the sales model despite repeated reports of illegal practices.

• Absence of effective controls: As early as 2004, an internal investigator described the problem as a “growing plague.” The following year, another investigator stated that the problem was “spiraling out of control.” Despite these early warnings, no governance mechanism stopped the practices for nearly 15 years.

• Systemic data falsification: Wells Fargo admitted to collecting millions in fees while employees falsified documents, forged signatures, and misused customers’ personal information to open fake accounts.

• Deficient organizational culture: Employees described intense pressure, with sales expectations reaching up to 20 products per day. Others reported frequent crying, stress levels leading to vomiting, and severe panic attacks.

The consequences of this governance breakdown go far beyond financial aspects. The Federal Reserve took the extraordinary step of capping Wells Fargo’s asset size, a restriction still in place. Beyond the $2.3 billion in settlements already paid, about 85,000 accounts incurred fees totaling $2 million, and customers’ credit scores were likely affected.

Citigroup: The Poisoned Legacy of Fragmented Governance

Citigroup provides another paradigmatic example of the consequences of failed data governance. The U.S. banking giant has been involved in a series of compliance failures dating back to 2013, resulting in over $1.5 billion in fines to U.S. regulators for risk management shortcomings.

In 2020, the Federal Reserve and the Office of the Comptroller of the Currency (OCC) fined Citigroup $400 million for deficiencies in compliance and data governance, marking a turning point for the bank.

The root of the problem lies in a fundamentally fragmented organizational and technological architecture:

• According to CEO Jane Fraser, the central problem was outdated, fragmented technology resulting from decades of underinvestment. Citigroup’s organizational structure, coupled with multiple acquisitions, created a “hodgepodge” of technology systems.

• A siloed organization prevented economies of scale, with a culture where many groups were allowed to solve the same problem in different ways, creating fragmented technology platforms and manual processes.

• CFO Mark Mason revealed that the bank must produce 11,000 global regulatory reports, some requiring up to 750,000 lines of data. Despite this volume, Citigroup was focusing on fixing only 15 to 30 reports required by U.S. regulators.

Technology and data governance issues directly affect business operations. In Citigroup’s wealth management division, it takes an average of nine days to open a new account — compared to an industry standard of just three days.

Since 2021, Citigroup has spent more than $7.4 billion upgrading its technology. Yet these investments have not been enough to resolve deeply rooted problems. This case perfectly illustrates how the absence of coherent, integrated data governance can durably compromise operational and regulatory performance.

These two cases show that data governance failures are not mere technical glitches but systemic breakdowns that strike at the very heart of organizational functioning. They confirm the theoretical analysis developed earlier: without anchoring in operational realities, alignment with strategic objectives, and development of true organizational capabilities, data governance remains an empty shell, unable to prevent the most costly abuses.

Conclusion

Failures in data governance initiatives are not anecdotal: they reflect a deep flaw in how organizations conceive, position, and implement this function. Three levels of analysis help structure the diagnosis: flawed conceptions, insufficient organizational integration, and unanticipated operational blockages.

The analysis reveals recurring patterns of failure that transcend sectors and organization sizes. The Wells Fargo and Citigroup cases dramatically illustrate the consequences of deficient data governance: $4.5 billion in combined fines, years of heightened regulatory oversight, massive erosion of stakeholder trust, and lasting operational impacts.

The paradox is striking: the more organizations invest in sophisticated, methodologically impeccable approaches, the greater the risk of drifting away from their initial objectives. Governance fails when it is thought of as an additional layer instead of being built as an invisible but essential infrastructure of collective action. Reduced to a set of deliverables, roles, or rules, it misses its true mission: supporting the organization in what it does, in what it decides, and in what it commits to with its stakeholders.

These lessons suggest that a radically different approach is needed. Instead of moving from frameworks to use cases, one must start from concrete needs to build tailored governance. Instead of producing deliverables to validate a method, one must develop capabilities to transform practices. Instead of governing data as technical objects, one must see them as levers of collective action.

The finding of failure is not a fatality. It is an opportunity for conceptual refocusing and renewed methodological rigor. But it requires recognizing that it is not only practices that must change — the very framework in which we think about data governance must evolve.

Bibliography

Alhassan, I., Sammon, D., & Daly, M. (2023). Data Governance Success: Practical Lessons from Real-World Implementations. Business Expert Press.

Brous, P., Janssen, M., & Herder, P. (2023). “Data governance as success factor for data science.” Information & Management, 60(2), 103735.

Consumer Financial Protection Bureau. (2016). CFPB Fines Wells Fargo $100 Million for Widespread Illegal Practice of Secretly Opening Unauthorized Accounts. Press Release, September 8, 2016.

DAMA International. (2023). The DAMA Guide to the Data Management Body of Knowledge (DAMA-DMBOK) (3rd ed.). Technics Publications.

Data Governance Institute. (2024). DGI Data Governance Framework v3.0. Retrieved from https://www.datagovernance.com/

Gartner. (2024). Data and Analytics Governance Survey: Organizations Struggle with Implementation and Value Realization. Research Note G00789234.

Harvard Law School Forum on Corporate Governance. (2019). The Wells Fargo Cross-Selling Scandal. Case Study Analysis.

ISO/IEC. (2023). ISO/IEC 38505-2:2023 — Governance of IT — Governance of data — Part 2: Implications of ISO/IEC 38505-1 for organizations. International Organization for Standardization.

Khatri, V., & Brown, C. V. (2023). “Designing data governance.” Communications of the ACM, 66(6), 87-93.

NewVantage Partners. (2024). Data and AI Leadership Executive Survey 2024: The Human Factor in Data-Driven Transformation. 12th Annual Survey.

Seiner, R. S. (2023). Non-Invasive Data Governance: The Path of Least Resistance and Greatest Success (2nd ed.). Technics Publications.

U.S. Department of Justice. (2020). Wells Fargo Agrees to Pay $3 Billion to Resolve Criminal and Civil Investigations into Sales Practices. Press Release, February 21, 2020.

Weber, K., Otto, B., & Österle, H. (2023). “One Size Does Not Fit All—A Contingency Approach to Data Governance: Ten Years Later.” Journal of Data and Information Quality, 15(2), 1-31.