Microsoft Fabric: 4 Governance Tips

Key insights

• Semantic models & OneLake catalog: Make semantic models discoverable in OneLake so users request access instead of creating duplicates.
  That improves reuse and reduces duplicate content.
• Endorsements & Discoverability: Use endorsement and discoverability settings to mark trusted assets and direct users to approved data.
  Endorsements increase confidence and speed up analysis.
• Sensitivity labels & Domains: Apply default sensitivity labels at the domain level to enforce consistent protection and compliance.
  Automatic domain labels cut manual errors and simplify policy enforcement.
• Governance recommendations & Stale assets: Use OneLake catalog recommendations to surface missing endorsements, unused content, and stale items.
  Regular cleanup reduces storage costs and keeps the data estate reliable.
• Bulk workspace access & Role management: Bulk add, modify, or remove users across workspaces to simplify administration at scale.
  Be aware of role-management gotchas when changing permissions across many workspaces.
• Domains and ownership & Lineage and auditing: Organize assets by domain with clear owners and track lineage and audits to understand dependencies and impacts.
  Combining these practices with platform-native tools builds consistent governance and trusted data.

Overview: Practical governance from Guy in a Cube

In a recent YouTube video from Guy in a Cube, host Marthe Moengen outlines four practical governance features that she says can improve how organizations manage their Microsoft Fabric estates. The video focuses on real-world controls such as making semantic models discoverable, applying default sensitivity labels at the domain level, using the OneLake catalog for governance recommendations, and bulk-managing workspace access. Together, these steps aim to reduce duplicate content, boost discoverability, and strengthen compliance while simplifying administration at scale.

Marthe organizes the advice around common pain points that data teams face as Fabric usage grows, including orphaned assets, unclear ownership, and inconsistent protection of sensitive data. She demonstrates the features in the Fabric UI and highlights how each capability plugs into broader governance patterns like lineage, endorsements, and auditing. Consequently, the guidance is practical for teams that want faster wins while still building toward a longer-term governance program.

Making semantic models discoverable

The video emphasizes making semantic models visible in the OneLake catalog so users can request access instead of rebuilding similar assets. By surfacing certified or endorsed models, organizations can reduce duplication and guide users to trusted sources, which improves data consistency. Marthe notes that discoverability must pair with clear access workflows so users can request and receive permissions smoothly without resorting to ad hoc copies.

However, there is a tradeoff between openness and control: increasing discoverability can lead to more access requests and higher administrative load if approvals remain manual. To mitigate that, Marthe suggests combining discoverability with domain-based ownership so requests route to the correct owners, and then automating common approvals where appropriate. This balances better reuse against the need to avoid bottlenecks and maintain security.

Using sensitivity labels and domain governance

Another core point in the video is applying default sensitivity labels at the domains level to ensure consistent protection across a set of workspaces and assets. Marthe shows how domain defaults can help enforce organization-wide rules from the start, reducing the chance that highly sensitive data is published without controls. She also explains that sensitivity labels work best when combined with clear domain ownership and simple policies that business users can understand.

Nevertheless, default labels introduce tradeoffs between convenience and precision because broad defaults may over-classify benign data or under-protect niche cases. To address this challenge, teams must monitor label effectiveness and allow local, governed exceptions where justified, while retaining automated guardrails. Thus, domains and labels together form a pragmatic compromise between central control and local flexibility.

Leveraging catalog recommendations and cleanup

Marthe highlights the OneLake catalog’s governance recommendations as a tool for identifying missing endorsements, unused assets, and stale content so that administrators can prioritize cleanup work. The catalog flags opportunities such as assets that lack endorsements or show no recent activity, which helps keep the estate trustworthy and reduces unnecessary storage costs. She points out that using these recommendations regularly turns governance from a one-time task into an ongoing habit.

At the same time, relying on automated recommendations can create noise and false positives that teams must manage, and it may require human review to set priorities correctly. Organizations should therefore combine automated scans with routine governance checkpoints and clearly defined thresholds for action. That way, recommendations become a way to focus effort rather than an extra source of busywork.

Bulk workspace access management and admin tradeoffs

The final practical tip Marthe offers is bulk management of workspace access to simplify administration across multiple workspaces and reduce repetitive tasks. She demonstrates bulk add, modify, and remove operations that can save time for administrators managing large estates, and she highlights common pitfalls such as role misapplication and propagation delays. Moreover, she warns that bulk changes require careful planning and testing to prevent broad permission mistakes.

There is a clear tradeoff between efficiency and risk when performing mass updates: bulk operations speed administration but can amplify errors if safeguards are weak. To manage that risk, Marthe recommends staging changes in a test environment, using role templates, and auditing changes after they occur. Consequently, bulk tools can scale governance work effectively when they form part of a disciplined process.

Conclusions and practical next steps

Overall, the video from Guy in a Cube offers actionable guidance for teams that want to strengthen governance in Microsoft Fabric without overcomplicating daily operations. It stresses discoverability, domain-based defaults, catalog-driven recommendations, and bulk administration as levers that deliver practical gains while acknowledging tradeoffs in control and workload. For newsrooms and data teams alike, the guidance points to incremental changes that deliver measurable benefit.

In short, organizations should treat these features as parts of a wider governance program: enable discoverability with clear request paths, apply sensible domain-level protections, use catalog insights to drive cleanup, and operate bulk actions with appropriate safeguards. By doing so, teams can improve trust, reduce duplication, and keep their Fabric environments clean and manageable as they scale.

Keywords

Microsoft Fabric governance tips, Microsoft Fabric governance best practices, Microsoft Fabric security tips, Microsoft Fabric governance checklist, Microsoft Fabric compliance best practices, Microsoft Fabric administration guide, Microsoft Fabric data governance, How to govern Microsoft Fabric