RACI of Roles
| Activity |
Data Experience Lead |
Pod Product Owner |
Domain Team |
Platform Team |
Data Governance |
| Data product vision and scope | C | A | R | I | I |
| Data product delivery | I | A | R | C | I |
| Data product operations | I | A | R | C | I |
| Enablement model and sequencing | A | C | I | C | I |
| Playbooks, templates, examples | A | C | I | C | C |
| Platform onboarding experience | C | I | I | A / R | C |
| Mesh discovery and UX standards | A | C | I | R | C |
| Contracts, lineage, SLAs standards | C | C | R | C | A |
| First-wave pod coaching | A | C | R | C | I |
| Ongoing pod support | I | A | R | C | I |
| Adoption metrics and signals | A | I | C | I | I |
| Showcases and cross-domain learning | A | C | C | I | I |
Legend: R Responsible. A Accountable. C Consulted. I Informed.
Dependency Management
Platform Dependencies
- Databricks workspace provisioning and access
- CI/CD pipeline templates availability
- Data catalog and discovery tooling readiness
Organizational Dependencies
- Executive sponsorship and decision-making authority
- Cross-functional resource allocation
- Legal and compliance framework clarity
Dependency tracking board and mitigation plans to be defined.
Risks and Assumptions
Primary Risks
- Role Misalignment: Existing roles interpret Data Mesh through their own lenses. Software engineering, analytics, and governance mindsets optimise for different outcomes, creating friction before delivery even starts.
- Task Framing vs Product Thinking: Pods treat data products as compliance work to finish, not products with users, lifecycle, and accountability.
- Cognitive Overload Under Delivery Pressure: Capacity-constrained teams struggle to absorb new roles, standards, and ways of working alongside feature delivery.
- Ownership Ambiguity: Unclear boundaries between platform, governance, and domains cause hesitation, rework, and slow decisions.
- Early Signal Amplification: Confusion or setbacks in early pods travel fast across countries and portfolios, shaping perception before capability matures.
- Self-Service Fallacy: Tools are assumed to replace enablement. Without shared understanding, self-service increases fragmentation rather than adoption.
Foundational Assumptions
- Core data platform on Databricks and AWS is production-ready and centrally supported
- Domain pods retain end-to-end ownership of data products, including quality and semantics
- Executive sponsorship remains active with visible support from CEO, CDO, and platform leadership
- Pods are capacity-constrained; protected capacity is allocated for data product work alongside feature delivery
- Enablement is high-touch in weeks 0-6, tapering after week 12 as pods stabilize
- Governance standards are defined upfront and embedded into tooling and playbooks
Mitigation plans and assumption validation checkpoints to be defined.