Marketing data governance: how to get the committee to discuss decisions, not versions of the number
Definitions, owners and quality: the fundamentals for connecting marketing, sales and finance data in large organizations.
Marketing data governance is the set of responsibilities, definitions and controls that makes it possible to use information with context and confidence. In large organizations, its first result should be making decisions comparable across brands and units. A visually consistent dashboard does not fix metrics calculated in incompatible ways.
When marketing presents conversions, sales presents contracts and finance presents recognized revenue, all three numbers can be correct. The problem starts when they get the same name or are compared as if they represented the same event. Governance work makes these differences explicit before the executive meeting.
Start with the questions that move resources
Choose recurring decisions: increasing investment, prioritizing a region, revising a journey or expanding an offer. For each one, identify the required measures and the systems that produce them. This scoping is more useful than trying to organize the entire data estate at once.
Create a metrics dictionary with formula, unit of analysis, period, exclusions and owner. A qualified opportunity needs entry criteria. Revenue needs a definition compatible with the intended financial reading. Changes to these rules should be recorded to preserve historical interpretation.
Establish contracts between producers and users
- Source: the system responsible for the record and the conditions for updating it.
- Quality: required fields, tolerated duplicates and consistency checks.
- Timeliness: delivery frequency and known delay.
- Responsibility: who fixes failures and who approves changes to definitions.
- Use: purpose, access and sharing limits, assessed by the responsible areas.
The contract can start as simple documentation accompanied by automated validations. Its value lies in the routine of compliance. If a campaign starts using new parameters, whoever maintains the reports needs to know before channel classification stops working.
Identity requires design, not just integration
Person, device, business account and contract are different entities. An identifier available on one platform does not automatically resolve the association between them. The Google Analytics User-ID documentation describes a specific feature for associating activity with identifiers supplied by the organization; it does not replace the business's identity design.
Define which associations are verified, which are estimated and which remain unknown. Also document access, retention and disposal rules with the competent teams. The existence of a central repository does not, by itself, demonstrate data quality or the suitability of every use.
Show confidence alongside the indicator
In a hypothetical example, one unit might deliver complete data in a day while another takes a week. Comparing the most recent period without flagging that difference can suggest a drop that does not exist. The dashboard should display freshness, coverage and relevant alerts.
Track incidents, time to fix and recurrence of discrepancies. Prioritize failures by their impact on decisions, not just the volume of records affected. A small inconsistency in the margin rule can matter more than many incomplete secondary fields.
Do we need to replace all our tools to govern the data?
No. Definitions, responsibilities and controls can evolve on top of the existing infrastructure. Replacements should answer demonstrated limitations.
Who should own a metric?
The area responsible for the business meaning should work with whoever maintains the data. The Data practice connects this foundation to CRM and revenue integration.