Data Governance Office: how to set it up in 90 days
A Data Governance Office is the team that makes the rules about data work in everyday practice. It does not need to be large or costly: it needs a clear mandate, named owners and results people can see early. Here is a 90-day plan to start.
What it is (and how it differs from the committee)
The committee decides: it approves policies and settles disputes, and meets every few weeks. The office operates: it prepares decisions, maintains the glossary and catalog, monitors quality and answers questions from the rest of the company. It is the difference between a residents’ board and whoever runs the building every day.
What it does
- Policies and standards: data quality, access, classification and retention.
- Glossary and catalog: everyone calls things the same and knows where each dataset is.
- Quality: define rules, measure and manage incidents.
- Access and compliance: who sees what, and evidence for audits.
- Support and training: answer questions and accompany teams.
- Data for AI: make sure the data AI systems use is appropriate and authorized.
Who is in it
To start, you need one person leading (Data Governance Lead), a network of part-time stewards in each area, and links with IT, legal and data protection. It should report to an executive who backs it. Without sponsorship, the office remains a good document nobody applies.
90-day plan
Days 1 to 30: mandate and scope
- Secure an executive sponsor and a written statement of what the office can decide.
- Choose only three priority data domains (for example customers, products and finance).
- Name an owner (data owner) and a steward per domain.
- Set up the committee and fix its calendar.
Days 31 to 60: first rules
- Publish a glossary with the 20 most disputed terms.
- Define quality rules for one domain and measure the baseline.
- Approve an access policy and a data classification policy.
- Open a simple channel for anyone to raise questions or incidents.
Days 61 to 90: operate and measure
- Run the first full cycle: an incident detected, resolved and communicated.
- Publish the first indicators.
- Show management one visible result (a report that used to disagree and now matches).
- Decide which domains join next quarter.
How to measure whether it works
- Share of critical data with an assigned owner.
- Open quality incidents and average resolution time.
- Requests answered on time.
- People using the glossary or the catalog.
Common mistakes
- Starting with the tool: buying a catalog before knowing who is responsible for what.
- Trying to cover everything: three domains done well beat twenty done halfway.
- No sponsor: the office proposes, but nobody backs it.
- No visible results: after three months nobody remembers why it exists.
Where AI fits
The data office is the natural foundation of AI governance: AI systems are only as reliable as the data they use. That is why it should be coordinated with the AI governance committee and whoever holds the AI governance officer role.
Frequently asked questions
Does an SME need an office?
Not as a department. One lead with allocated time and stewards in each area is enough; what matters are the functions, not the org chart.
What is the difference between office and committee?
The committee decides and approves; the office prepares, operates and follows up.
How long until it shows results?
With a limited scope, the first visible result can arrive in around three months.
Informational content, not legal advice.
This gives you the judgment. If you want the finished document: Structure and Roles Pack.
What's your Data Governance maturity?
Free assessment with your priority gaps, plus the self-assessment quiz and savings calculator on the Data Governance path.