Content Certification

Content certification is the practice of marking which reports, dashboards, or datasets an organization stands behind — distinguishing content somebody maintains from content somebody once made. BI platforms provide the mechanism as an endorsement badge; the practice is what makes the badge mean something.

It exists because self-service has an asymmetry built into it. Giving people the ability to build reports without waiting works, and the count of reports rises. Nothing in any tool removes one. Five years on, a platform holds hundreds where a few dozen are used, several disagree, and the ones executives quote are not reliably the maintained ones.

What certification has to promise

A badge that means “this looks fine” teaches readers to ignore badges. Four commitments make it informative, and all four are checkable.

• A named owner. A person or a team, not a department. The test is whether someone can be asked a question about it next month.
• The definitions it uses. Which metric definition each number follows, so two certified reports cannot silently mean different things by “active customer.”
• A freshness commitment. Stated as an objective someone measures, not as a refresh schedule. A schedule says what was attempted; an objective says what readers may rely on.
• A review date. Certification without expiry becomes a claim about the past. A date forces a decision — recertify, or remove the mark.

Read together, those four make certification a commitment rather than a compliment, and that is the whole difference. They are also why the certified set should be small: each entry is an obligation somebody carries, and an organization that certifies two hundred reports has certified nothing.

Separate spaces do half the work

Before any endorsement scheme, the platform’s own structure helps. Power BI gives each user a personal workspace and separate collaborative workspaces that “reside within capacities”; other tools have the same split under different names. Used deliberately, it means a work in progress cannot be mistaken for the company’s answer, and moving something into shared space becomes the natural moment to ask who owns it.

That boundary matters more than it sounds, because the most common way a wrong number spreads is not a certified report being wrong. It is a personal exploration being found, shared, and treated as official by people who never saw where it came from.

Retirement has to be scheduled

Usage data tells you what nobody opens. Deleting it is still hard, because the person who built it may object and nobody wants to be the one who removed something. So make removal the default and the calendar the actor.

• Review on a fixed cadence, quarterly or twice a year, rather than when someone complains about clutter.
• Notify, then remove. Unused content gets a notice with a date; silence means it goes. Objection is cheap and keeps it, which is the point — the decision moves to the person who cares.
• Archive rather than delete where it is cheap, so a removal that turns out to be wrong is reversible and nobody has to be brave.

The measure that tells you the scheme is working is not how many items are certified. It is what share of actual viewing happens on certified content, and whether the share of shared content with a named owner is rising — the BI equivalent of catalog coverage. A certified set nobody reads means the badge is not where readers look.

What it cannot fix

Certification marks content; it does not reconcile it. If a certified report and an analyst’s own query disagree because the metric is defined in two places, no badge settles the argument — that belongs to the definition layer upstream, and the BI platform can only present the answer it is handed.

Two reports can also both be certified, both correct, and show different numbers because they were refreshed at different times. An extract is “saved separately from the original dataset” and carries its own refresh state, so stating the cutoff on the report matters as much as the endorsement.

Where certified content is consumed by other teams, the clearest framing is to treat it as a data product with a stated audience and support expectation, and to write the promise about definitions and freshness into a data contract rather than leaving it to a shared understanding.

How certification fits with models, refresh, capacity, and ownership in an enterprise BI platform is worked through in The Dashboard Is Slow and Nobody Owns It.

References: Microsoft Learn, What is Power BI Premium?; Tableau Help, Extract Your Data.