Every CDO eventually creates a Data GovernanceData GovernanceData governance is the set of policies, roles, and processes that ensure data is accurate, secure, well-defined, and used responsibly across an organization.View full definition → Council. Most of those councils become useless within 12 months.
They start with good intentions: bring together the right stakeholders, make decisions about data collectively, drive organization-wide change. Then reality sets in. The meeting has 25 attendees who aren't sure why they're there. Agenda items keep getting deferred. No one has clear authority to actually decide anything. The CDO spends three hours a month facilitating a conversation that produces no outcomes.
This doesn't have to be your story.
Too many people. A governance body with 25 members is a committee. A committee's natural state is inaction. Effective governance councils have 7-12 members, enough diversity of perspective, small enough for actual decision-making.
Wrong people. Governance councils fail when they're populated by representatives rather than decision-makers. If your CFO sends a financial analyst to the governance meeting instead of attending themselves, will never be a CFO priority. The council must include people with budget authority and business accountability.
No clear mandate. "We govern data" is not a mandate. "We make binding decisions on data definitions, data qualitydata qualityThe degree to which data is fit for purpose: accurate, complete, consistent, timely, valid and unique. Poor quality data undermines analytics, reporting and AI.View full definition → standards, and data access policies" is. Without explicit authority over specific decision types, the council becomes advisory, and advisory bodies rarely drive change.
Tactical focus instead of strategic. Many governance councils spend their time reviewing individual data issues ("Should we include returned orders in revenue?") instead of setting the policies that prevent those issues from arising. Tactical councils scale poorly. Strategic councils set rules that prevent problems.
Executive Steering Committee (quarterly): CDO + C-suite representatives. Makes strategic decisions: data strategy alignment, major investment approvals, organizational changes. Maximum 8 people.
Data Governance Council (monthly): CDO + domain data owners (one per major business function). Makes operational governance decisions: data definitions, data qualitydata qualityThe degree to which data is fit for purpose: accurate, complete, consistent, timely, valid and unique. Poor quality data undermines analytics, reporting and AI.View full definition → standards, data access policies, Data lineageData lineageData lineage maps how data moves and transforms across systems, from origin to consumption, showing where it came from, what changed it, and where it goes.View full definition → requirements. 8-12 people.
Domain Data Steward network (weekly/as-needed): Data stewards in each business domain, coordinated by the CDO team. Implements governance decisions, resolves data issues, owns data qualitydata qualityThe degree to which data is fit for purpose: accurate, complete, consistent, timely, valid and unique. Poor quality data undermines analytics, reporting and AI.View full definition → within their domain.
The key principle: decisions should be made at the lowest level that has enough context. The Steering Committee shouldn't be arguing about whether "customer" includes trial users. The domain stewards should resolve that and escalate only if they can't reachreachThe number of unique people exposed to your message in a given period. Unlike impressions, reach counts each person once, no matter how often they see it.View full definition → consensus.
Knowledge check
1. According to the lesson, what is the recommended size range for an effective Data Governance Council?
2. How often does the lesson recommend the Executive Steering Committee should meet?
3. The lesson contrasts 'tactical' and 'strategic' governance councils. Which activity best exemplifies a STRATEGIC council approach?
4. Select ALL the reasons the lesson identifies for why most governance councils fail.
Sélectionnez toutes les réponses correctes.
5. Select ALL the decision types the lesson assigns to the monthly Data Governance Council (not the Executive Steering Committee).
Sélectionnez toutes les réponses correctes.
A governance council that works meets monthly, makes 3-5 binding decisions per meeting, and tracks outcomes. Here's a concrete agenda structure:
Standing items (15 min):
Decision items (30 min):
Strategic items (15 min):
One practical rule: Every item on the agenda is either a decision or an escalation. Nothing is "for information only." If it's for information only, send an email.
ING Bank's data governancedata governanceData governance is the set of policies, roles, and processes that ensure data is accurate, secure, well-defined, and used responsibly across an organization.View full definition → program is one of the most cited in European financial services. Their approach: treat data governancedata governanceData governance is the set of policies, roles, and processes that ensure data is accurate, secure, well-defined, and used responsibly across an organization.View full definition → as a product, not a compliance function.
They built a Data GovernanceData GovernanceData governance is the set of policies, roles, and processes that ensure data is accurate, secure, well-defined, and used responsibly across an organization.View full definition → Council with explicit business ownership, each data domain (customer, product, transaction, risk) had a named executive owner accountable for data qualitydata qualityThe degree to which data is fit for purpose: accurate, complete, consistent, timely, valid and unique. Poor quality data undermines analytics, reporting and AI.View full definition →. Council meetings made binding decisions, tracked in a governance register. Outcomes were reported to the Management Board quarterly.
The result: data qualitydata qualityThe degree to which data is fit for purpose: accurate, complete, consistent, timely, valid and unique. Poor quality data undermines analytics, reporting and AI.View full definition → improved measurably across their core domains within 18 months, and regulatory data submissions became significantly more reliable. The governance council had teeth because it had accountability.