--- title: "Four-Phase CCoE Rollout" description: "The month-by-month build plan for the CCoE: what happens in each phase, what good progress looks like, and which signals show the CCoE is on track." sidebar: order: 3 label: "Four-Phase Rollout" source_url: "https://framework.stackit.cloud/advisory/cloud-centre-of-excellence/rollout-phases/" source_file: "docs/advisory/cloud-centre-of-excellence/rollout-phases.mdx" --- ## The CCoE develops — in four phases A CCoE does not become fully operational on day 1. It goes through four maturity phases, each with different priorities and different success signals. ## Phase 1: Foundation (Months 1–3) **Focus:** Build the team, establish the mandate, set first standards **What happens:** - CCoE core team recruited and released from other duties (minimum 4 FTE) - Charter adopted and communicated - First technical decisions made: IaC toolchain (Terraform + STACKIT Provider), Git workflow, CI/CD principles - STACKIT accounts and organisational structure created - First guardrails implemented as policy-as-code (geo-restriction, mandatory tags, encryption) - Track A training started for all IT staff **What good progress looks like:** - Charter is signed ✓ - STACKIT organisational structure in place ✓ - 3 core guardrails implemented and tested ✓ - CCoE team at full strength ✓ - Track A running for all IT staff ✓ **Warning signs:** - Charter signing delayed by >6 weeks → mandate problem, not a technology problem - Team cannot be staffed → rethink recruitment strategy, bring in a partner - Guardrails are blocking every deployment → too restrictive, calibration needed ## Phase 2: Foundations (Months 4–8) **Focus:** Build the landing zone, onboard first workloads, activate FinOps **What happens:** - Landing zone fully deployed as Terraform code (hub-and-spoke network, IAM structure, centralised logging) - First pilot team onboarded — 2–3 workloads migrated as lighthouse projects - IaC module library starts (5–10 standard modules for common resource types) - FinOps basics: tagging standard implemented, showback reports for first business units - Cloud Champions nominated in first teams - Track C/D training started for Platform Engineers and Architects **What good progress looks like:** - Landing zone fully deployed, all guardrails active ✓ - First workload team independently (with support) in the cloud ✓ - Tagging compliance >80 % for new resources ✓ - First showback reports delivered to business unit ✓ - > 5 Terraform modules in the library ✓ **Warning signs:** - Landing zone still not finished after 6 months → reduce scope, MVP first, then iterate - Pilot team has been waiting weeks for CCoE approvals → decision processes too slow, build self-service - FinOps tagging being ignored → escalate, harden the guardrail ## Phase 3: Scale (Months 9–18) **Focus:** Migration waves, expand self-service, decentralise the CCoE **What happens:** - Migration waves 1–N: structured workload migration process - Self-service catalogue: teams can deploy standard resources without CCoE tickets - CCoE transitions from "operational support" to "enablement and governance" - FinOps maturity increases: from showback to chargeback - Communities of Practice (Cloud Guilds) are active and self-organised - STACKIT advanced features are being used: Managed Kubernetes, Managed Databases **What good progress looks like:** - Teams deploy independently without CCoE tickets ✓ - > 50 % of target workloads migrated ✓ - Guardrail compliance >99 % ✓ - Cloud costs within budget (±10 %) ✓ - CCoE spends \<30 % of time on operational support ✓ **Warning signs:** - CCoE is still an operational bottleneck → self-service not developed far enough - Migration speed declining → resource bottleneck or technical hurdles — run a retrospective - Costs rising uncontrolled → FinOps escalation, activate budget alerts ## Phase 4: Institutionalisation (Month 18+) **Focus:** Cloud is the norm, CCoE is an innovation driver **What happens:** - All target workloads migrated or consciously excluded - CCoE focus shifts from migration to innovation: cloud-native modernisation, evaluating new STACKIT services - FinOps fully operational: chargeback active, budget accuracy ±5 % - Cloud competency embedded in breadth — no single point of failure in the CCoE - Regular platform upgrades, Kubernetes version management, service evaluations **Signals that Phase 4 has been reached:** - New projects start automatically in the cloud — nobody asks whether on-premises is possible - CCoE conducts proactive platform reviews without teams having to wait - External audits confirm compliance without remediation work - Business units ask the CCoE for new cloud capabilities, not just for help with problems ## What is realistic after 12 months? **A well-performing CCoE after 12 months:** - 60–80 % of target workloads migrated - Landing zone stable, no unplanned guardrail exceptions - 15–25 % cloud cost savings through FinOps optimisation realised - 4–6 Cloud Guilds active, all technical teams have Cloud Champions - CCoE spends >50 % of time on enablement and innovation **A dysfunctional CCoE after 12 months:** - CCoE is a migration bottleneck, teams are waiting for approvals - Guardrails are being bypassed because they are too restrictive or too slow - FinOps exists on paper but nobody looks at the costs - Cloud knowledge is concentrated in 2–3 CCoE people ## Practical steps 1. **Define phase timeline** with date targets for your transformation 2. **Agree phase-gate criteria** for each phase transition with the CIO 3. **Monthly retrospective** in the CCoE team: where are we in the phase, what is blocking, what should we accelerate?