A 90-day plan for a DevOps transformation

Fast delivery starts with a clear picture everyone can see. I begin by collecting three core metrics—deployment frequency, lead time, and change fail rate—so the numbers tell us where we stand.

console.log("hello");

Days 0-30: Take the snapshot

• List bottlenecks in current pipelines/environments; name owners.
• Map every manual prod step; list automation opportunities.
• Pick a pilot repo and lock in automated tests + static analysis + container build.

Days 30-60: Define the standard

• Create a shared template for all services: build, test, quality gate, image naming, versioning.
• Publish a reusable pipeline/action library plus a sample repo for teams to copy.
• Trial trunk-based branching + feature flags to keep changes small.

Days 60-90: Productize the platform

• Build shared IaC modules (network, database, observability) and make them self-service.
• Share deployment insights on a dashboard; summarize red areas weekly.
• Simplify the post-mortem template so it’s fillable in 30 minutes.

Metrics to watch

• Deployment frequency: How many times per week to prod?
• Lead time: Merge to live.
• Change fail rate: Rollback/incident rate for live changes.
• Mean time to recovery: Detecting impact to restoring service health.