Only 26% of data governance programs incorporate culture and communication into their strategy, according to the 2025 Gartner Evolution of Data and Analytics Governance for AI Survey of 62 organizations. The majority prioritize trust, risk, security, and technical controls instead. At the Gartner Data & Analytics Summit in Orlando, Gartner analyst Nate Novosel presented seven practical improvements for data governance execution, with a theme that runs through all of them: governance programs do not fail because of bad policies. They fail because the organization does not change its behavior to follow them. Gartner predicts that by 2027, 60% of organizations that fail to address cultural challenges in data governance will also fail to govern artificial intelligence successfully.
Why Does 74% of Data Governance Ignore Culture?
Three-quarters of data governance programs focus on compliance, policies, and technical controls while omitting the behavior change that determines whether anyone follows the rules. According to a 2025 Gartner survey presented at the summit, trust, risk, and security topped the priority list at 61%, followed by structure, roles, policies, and processes. Culture and communication ranked near the bottom at 26%, and outcomes and business value scored only 34%.
The gap between policy and practice shows up in participation. One unproductive meeting is enough to lose a governance board member permanently. If attendees feel they did not contribute or that the discussion could have been an email, they will not return. Stewards with full-time jobs see governance as extra work with no recognition, so they stop showing up. Without participation, programs collapse within 12 to 18 months.
Gartner predicts that by 2027, 60% of organizations that have failed to address the cultural challenges associated with data governance will fail to govern AI successfully.
What Is Minimum Effective Data Governance?
Minimum effective data governance targets the bare minimum data quality required to realize return on investment from specific business decisions, rather than pursuing enterprise-wide data perfection. Gartner calls this the “trust model”: identifying the necessary quality threshold for specific data to deliver value, then improving that data to meet the threshold and moving to the next target.
The traditional approach starts from data issues, assembles a large central governance body, runs a multi-year cleanup, and promises value at the end. That sequence leads to collapse because stakeholders cannot see results and lose interest before the program delivers anything.
The minimum effective approach reverses the sequence: start from the business outcome, identify the high-value data behind the decisions that drive that outcome, fix the data to the minimum acceptable quality, demonstrate the improvement, and move on. Where traditional programs attempt multi-year, domain-by-domain cleanups that delay results until completion, minimum effective governance delivers incremental improvements in weeks to months by targeting only the high-value data linked to specific outcomes.
The trade-off is real. Lower-priority data may continue to degrade while resources focus on high-value targets. But the alternative, attempting enterprise-wide governance and losing the program entirely within 18 months, produces worse outcomes.
How Do You Build on Existing Stewardship?
Organizations should find people already doing informal stewardship work and formalize their role, rather than assigning new stewards who lack context and relationships. In every team, someone is already fixing data to make reports accurate, ensuring records are entered correctly, or maintaining quality that others depend on. That person is doing governance work whether the title exists or not.
Exact Sciences demonstrated a three-phase approach to formalize existing stewardship. Phase one: thank the person. Stewardship is a thankless job, and acknowledging the work generates immediate goodwill. Phase two: standardize and streamline what they are already doing, reducing the time and effort required. Phase three: transition to full stewardship with additional responsibilities, but only after the first two phases build trust and demonstrate support.
The worst mistake a governance team can make is to assume nobody is doing anything and announce that data quality is terrible. The person who has been keeping data usable will be personally offended, and the program loses its most natural ally.
Ovintiv positioned stewardship as a value proposition rather than a compliance task. Stewards share data assets, reduce hassle, codify workarounds, manage exceptions, and preempt escalations. Framing the role around what stewards contribute, rather than what they must enforce, changes participation dynamics entirely.
How Do You Keep a Data Governance Program Alive?
Governance programs sustain momentum through incremental wins, visible celebrations of improvements, and meeting structures that make every participant feel their presence is irreplaceable. The signal that a program is working is five words: “that was great, what’s next?”
When a targeted data improvement produces visible results, a better report, more reliable decisions, higher data accuracy, and someone asks what else governance can fix, momentum is building. Celebrating that success at the next governance board meeting costs nothing and creates energy that keeps participants engaged. No one criticizes a meeting that opens with a win.
Meeting structure itself determines whether people keep showing up. Every agenda item should be perennially urgent and important: progress on issues from the prior meeting, compliance updates, a prioritized list of data quality improvements awaiting action. The committee should avoid detailed status reviews and operational details that can be handled offline.
Each participant should represent a unique voice, a perspective only they can provide. If someone believes the meeting can function without them, they will eventually stop attending. Composing the committee so that each member brings an irreplaceable viewpoint, whether compliance, operations, IT, or data and analytics leadership, creates accountability through necessity.
The modern outcome-driven approach works backward from the traditional model: business strategy and outcomes first, then prioritized data scope, then minimum effective governance, then demonstrable business impact, then repeat. That cycle, sustained through culture change and visible wins, is what keeps governance programs alive past the first year.
FAQ
1. What is minimum effective data governance?
Minimum effective data governance targets the bare minimum data quality needed to realize return on investment from specific business decisions. Instead of pursuing enterprise-wide data perfection through multi-year programs, organizations identify the quality threshold required for high-value data, improve it to that threshold, and move to the next target. Gartner calls this the “trust model.” The approach avoids large central governance programs that typically collapse within 12 to 18 months due to stakeholder fatigue.
2. Why does data governance culture affect AI readiness?
Gartner predicts that by 2027, 60% of organizations that fail to address cultural challenges in data governance will also fail to govern AI successfully. A 2025 Gartner survey of 62 organizations found that only 26% incorporate culture and communication into their governance strategy. Without behavior change, governance exists as policy on paper but not in practice, and AI initiatives that depend on governed data inherit the gap.
3. How do you find data stewards in an organization?
Organizations should identify people already doing informal stewardship work rather than assigning new stewards. In every team, someone is typically fixing data quality issues, maintaining reports, or ensuring accurate records. Exact Sciences demonstrated a three-phase approach: acknowledge the person’s existing contribution, standardize and streamline their processes to reduce effort, then transition to formalized stewardship with additional responsibilities.
4. What makes a data governance meeting effective?
Effective governance meetings require three elements: every agenda item must be perennially urgent and important, including progress updates, compliance changes, and a prioritized queue of data quality improvements. Each participant must represent a unique voice or perspective that only they can provide. The meeting should produce decisions and celebrate wins, not deliver status updates that could have been an email. Losing one participant to a meeting perceived as unproductive can trigger further attrition.
This article is based on a presentation by Nate Novosel, Sally Parker, and Sarah Turkaly, Gartner analysts, at the Gartner Data & Analytics Summit in Orlando (2026). Lucian Fogoros of IIoT World attended the event. AI tools were used to help summarize and organize the content. Reviewed and edited by the IIoT World editorial team.
