Power BI content certification and endorsement

In organisations with mature Power BI usage, dozens to hundreds of reports and semantic models exist. Users struggle to know which are authoritative — which to use, which to trust, which to build on. Endorsement is the Power BI / Fabric mechanism for designating trusted content; certification is the most formal level. Used well, it dramatically improves report-discovery and reduces "everyone builds their own version of revenue."

The endorsement levels

Power BI / Fabric supports three states:

Not endorsed

The default — content the author created, not formally vetted. Most content sits here. Fine for personal exploration; doesn't carry weight as authoritative.

Promoted

Content the author or their team promotes as a useful broader resource. Self-service endorsement — the content owner says "this is worth others using." No formal review required.

Useful for:

• Team-shared models the team uses internally.
• Departmental content of broader interest.

Certified

Content certified by a designated authority (typically the Power BI / Fabric admin team or a Centre of Excellence) as authoritative.

Used for:

• Single source of truth semantic models that the organisation should standardise on.
• Reports and dashboards with executive consumption.
• Reference dataflows centrally maintained.

Certified content carries the strongest endorsement and is what users should default to using.

Why endorsement matters

Without endorsement:

• Five different "revenue" reports exist; users don't know which to trust.
• Multiple teams rebuild the same data preparation.
• Decisions get made on stale or unrepresentative data.
• New users wander through content with no guidance.

With endorsement:

• Certified content surfaces prominently in user search.
• Promoted content appears next.
• Users know which content to start from.
• Authoritative source is unambiguous.

The certification programme

Setting up certification requires operational discipline:

Define standards

What does "certified" mean in your organisation? Examples:

• Documented business rules — model logic clearly explained.
• Data lineage — source-of-truth traceable.
• Refresh schedule verified and reliable.
• Security model implemented appropriately.
• Performance acceptable for expected user load.
• Owner accountable for the content.

Designate certifiers

Who can certify content?

• Power BI / Fabric administrators — typically a small group.
• CoE team — Centre of Excellence designation.
• Business-domain experts in some organisations.

Configure in Power BI admin portal → Tenant settings → Certification.

Submission and review process

For content to become certified:

1. Author submits content with documentation.
2. Certifier reviews against standards.
3. Feedback if changes needed.
4. Certification applied when standards met.
5. Periodic recertification — annual review.

Without process discipline, certification either doesn't happen or proliferates.

Communicate the programme

• Documentation of what certification means and how to get it.
• Training for authors on the certification path.
• Onboarding new authors explaining the programme.

Operational considerations

• Don't over-certify — if everything is certified, certification means nothing.
• Don't under-certify — too few certified items means users still wander.
• Maintain certified content — certified content that goes stale erodes trust.
• Recertification cadence — at least annually.
• Decertification path — content that no longer meets standards gets demoted.

Featured content

Beyond endorsement, featured content is a separate concept:

• Specific content pinned as visible to all users.
• Shows up in users' Home / Featured sections.
• Used for organisation-wide priority content.

Featured + certified is a powerful combination — high-trust, high-visibility.

Sensitivity labels for analytics

Combine endorsement with sensitivity labels on semantic models and reports:

• Certified models with appropriate sensitivity labels.
• Labels travel with exports (Excel exports inherit the label).
• DLP policies can act on labelled analytics content.

This produces a coherent governance picture — content is endorsed AND classified.

Cultural shift

The biggest barrier to a certification programme isn't tooling — it's culture. Common challenges:

• Authors don't want certification overhead.
• No clear owner for the certification process.
• Standards too strict — nobody gets certified.
• Standards too loose — everything gets certified.
• No enforcement — uncertified content gets used anyway.

A working programme balances rigor with accessibility. Iterate.

Where this fits

For organisations with 10+ Power BI authors and many shared reports, certification is one of the highest-leverage governance investments. The cost is operational (the review process); the benefit is durable trust in shared analytics.

For organisations still in early Power BI adoption, certification can wait — first establish the practice; then add governance.

Power BI / Fabric certification programmes mature over time. Start with promoted content (low overhead); evolve to formal certification as your analytics culture matures.