Data Governance

Data Governance measures the quality of the data your other features rely on. It keeps the fields that matter complete, current, and correct, and it shows who is responsible when they are not. Scores, roadmaps, and recommendations are only as trustworthy as the data behind them, so this is the feature that earns that trust.

The three quality dimensions

Every tracked field is judged on three dimensions:

• Completeness. The field is filled in, not empty.
• Currentness. The value has been reviewed recently enough to still be trusted.
• Correctness. The value is actually valid, not just present. A real reference, an allowed choice, or a sensible range.

Core concepts

• Scope. The per-table governed unit. A scope names a table, a filter that selects which records it covers, the fields it tracks, and the review cycle those fields follow. For example, "critical business applications that are not obsolete".
• Charter. A container that groups related scopes and carries shared settings, such as a campaign deadline and the responsible steward.
• Rules. Each tracked field carries typed rules (not empty, in an allowed list, a valid reference, and so on) that are checked automatically rather than by hand.
• Score and gaps. A scope produces a data-quality score and a gap list, so a steward sees exactly what is missing, overdue, or wrong.

Defining what good data means

You govern data table by table. For each table you set up one or more scopes: choose the records it covers, name the fields it tracks, set who the steward is, and set the review and reminder cycles those fields follow. One table can carry several scopes, for example a lighter yearly review and a stricter quarterly review.

The data-quality widget

On any record covered by a scope, the data-quality widget shows the health of each tracked field. For every field it shows the current value, whether it is complete, whether it is current, and when the next review is due, with overdue fields flagged. You can confirm a single field as reviewed or confirm them all at once, and the completeness and currentness percentages update as gaps close.

How it works for you

• Continuous, not periodic. Rules are evaluated as data changes and roll up into a live score, instead of a manual review cycle.
• Trust and verification. A value's origin is recorded, so a trusted operational source can skip the manual review a hand-typed value still needs. A field can be re-checked when it changes, on a review cycle, or both.
• Campaigns and stewardship. Run a time-boxed drive to close gaps against a deadline, with the work landing in the responsible steward's queue.
• Readiness. Ask whether the data is good enough to run a given analysis, and get the gap list to close before you rely on the result.

Screenshots and step-by-step guides for authoring scopes and charters, and for running campaigns, will follow here.