Kanban Method
Kanban improves delivery by visualizing work, limiting work in progress, and managing flow without imposing fixed sprints or new roles. It gives migration teams predictable throughput and early warning of bottlenecks.
Best Practice: Kanban Method
The Kanban Method, formalized by David J. Anderson, is an evolutionary, change-management approach for improving the flow of knowledge work. Unlike Scrum, it does not require fixed iterations or new roles; instead it starts with what you do now and improves it incrementally. Teams visualize their workflow on a board, limit work in progress (WIP), and manage flow so work moves smoothly from request to delivery. For modernization programs that mix planned migration work with unplanned incidents, Kanban's pull-based flow keeps throughput predictable.
Step-by-Step Implementation Guidance
- Visualize the workflow: map your real stages (for example, To Do, Analysis, Build, Review, Done) as columns on a board.
- Make policies explicit: write down entry and exit criteria for each column.
- Set WIP limits per column so the team finishes work before starting more.
- Manage flow: watch for bottlenecks where cards pile up and address them.
- Implement feedback loops such as a daily standup and a replenishment meeting.
- Measure flow with cycle time, throughput, and a cumulative flow diagram.
- Improve collaboratively using data and small, safe experiments.
Common Mistakes Teams Make When Ignoring This Practice
- No WIP limits, so everything is in progress and nothing finishes.
- A board that does not reflect how work actually flows.
- Treating the board as a to-do list rather than a flow system.
- Ignoring cycle-time data, so bottlenecks stay invisible.
- Pushing work into the system instead of pulling when capacity is free.
Tools and Techniques That Support This Practice
- Kanban boards in Jira, Trello, Azure DevOps, GitHub Projects, or physical boards.
- Cumulative flow diagrams and cycle-time scatterplots.
- WIP limits, classes of service, and explicit pull policies.
- Little's Law for relating WIP, throughput, and cycle time.
How This Practice Applies to Different Migration Types
- Cloud Migration: Track each workload as a card through assess, migrate, and validate stages with WIP limits to avoid overload.
- Database Migration: Visualize schema, ETL, and cutover tasks so the team finishes one safely before starting the next.
- SaaS Migration: Flow tenant onboarding and integration work through clear stages, limiting concurrent cutovers.
- Codebase Migration: Manage module-by-module refactors as a steady pull system to keep cycle time low and quality high.
Checklist
- The board mirrors the real workflow stages.
- Each column has explicit entry and exit policies.
- WIP limits are set and respected.
- Cycle time and throughput are measured.
- Bottlenecks are reviewed and addressed regularly.
- Work is pulled, not pushed.
- Improvements are based on flow data.