Guides
How do you govern multi-BU Marketing Cloud?
Install shared operating standards: naming, suppression truth, QA, authentication, escalation, and orphaned-BU policing. Federate creative. Do not federate chaos. Freelancer free-for-alls do not survive enterprise BU counts.
August 16, 2026·10 min read
Side-by-side
| Control | Federated OK? | Centralize? |
|---|---|---|
| Creative / content craft | Yes | Shared modules help |
| Consent / suppression | No | Yes |
| Auth / domains / SAP | No | Yes |
| Journey business logic | Yes, with standards | QA gates central |
| Incident escalation | No | Yes |
| BU folder archaeology | No | Policing required |
Scale without a freelance clearinghouse
Enterprises with many business units cannot let every BU invent sending standards. Shared naming, shared suppression truth, and shared escalation paths are deliverability and throughput controls, not bureaucracy for its own sake.
RHG’s enterprise work includes multi-BU patterns at IHOP-scale Marketing Cloud estates (~260 internal BUs). The lesson is not “hire more freelancers.” The lesson is an operating standard with teeth.
What must be shared
Federate brand expression. Centralize these.
- Naming conventions for data extensions, journeys, and content
- Suppression and consent truth
- Send classification discipline
- Authentication / SAP / domain change control
- QA gates before major launches
- After-hours escalation path
- Orphaned BU detection and cleanup
What can stay local
Creative, local calendar priorities, and BU-specific journey logic that respects shared standards. Local teams still need a path to request exceptions without inventing shadow systems.
Ownership model
Governance without an operating unit becomes a PDF. Someone has to run the weekly rhythm, police forks, and absorb peak weeks with people who already know the stack.
That can be a strong internal core, a Cell, or a hybrid with clear boundaries. See in-house vs outsourced Cell. Fake-managed retainers that invoice one line while every BU freelances standards will fail at scale.
How to install standards without a revolt
Start with one painful shared control: suppression truth or naming. Publish the standard, offer migration help, measure reopen tickets and launch slips before/after. Widen once the rhythm is real. Big-bang governance resets recreate old handoffs at higher volume.
RFP and partner language
Ask partners how they handle naming standards, shared QA, and orphaned BUs. Vague “we support multi-BU” answers are not enough. Use RFP questions and the scorecard.
FAQ
Does multi-BU mean we need a huge SI?
Not necessarily. You need an operating standard and a bench that enforces it. Large SI waves without ops ownership still leave orphaned BUs.
How do we handle BU exceptions?
Time-boxed, documented exceptions with an expiry and an owner. Permanent forks become the new standard by accident.
Where does deliverability fit?
Shared sending standards are deliverability controls. Preference and auth chaos across BUs is how reputation burns.
Keep reading
Next step
Talk to the Cell
Bring a painful queue, a stalled Data Cloud activation, or a staffing RFP. We will tell you which offer fits.