Power BI is consistently expanding its footprint in the enterprise environment. One lesser-known but highly valuable feature is the capacity to certify datasets, which enables organizations to signal to end users that certain datasets are more reliable. We are going to explore this valuable yet often overlooked feature, and to discuss the difference of promoted and certified dataset and use cases of when to use which.

Endorsement in Power BI comes in two forms: promotion and certification.

Promotion: This type of endorsement allows users to flag content they find valuable, useful, and ready for others to access. It facilitates the collaborative sharing of content across the organization.

Any content owner or member with write permissions in the workspace where the content resides can promote the content when they deem it suitable for sharing.

Certification: This type of endorsement indicates that the content aligns with the organization’s quality standards and can be trusted as reliable and authoritative across the enterprise.

Certification is granted by a designated group of reviewers selected by the Power BI administrator. Content owners who want their content certified but don’t have the authority to certify it themselves must follow their organization’s process for requesting certification.

Certification option is available only if a Power BI administrator has that feature enabled for your organization.

Once you promoted or certified a dataset, you will be able to see them in the OneLake datahub.

All Fabric items can be endorsed except for Power BI dashboards.

Deciding whether to endorse an app or a report

Endorsement is meant to be used for the content items (apps, reports, semantic models, and dataflows) you want people to find, use, and possibly re-share.

When you’re sharing data with a broad audience, a Power BI best practice is to share that data via an app. If you’re following this best practice, you want people to be able to find that app easily. In such cases then, you should endorse the app.

If you prefer to share reports directly, you should consider endorsing the report itself.

Regardless of whether you’re sharing a report or an app, if the underlying semantic models are clean and suitable for sharing, endorsing those semantic models is a smart move. The same goes for dataflows.