2. PURPOSE
Share AI and Store PODs experience
about Scrum practice and transition to
Kanban.
3. OUTCOMES
By the end of this meeting:
• We will understand challenges of scrum practice
for AI pod and how Kanban practice might solve
some of them, so we can share our experience to
support other teams.
• Opportunity for you to ask question about
Kanban.
• Opportunity to have a discussion with your peers
about interests and challenges.
4. What's in it for me?
• You learn how Kanban process solves some of
existing scrum practice challenges for infra teams.
• You will see what steps we took to move from
Scrum to Kanban
• You will see our Kanban board design and how it
can help to support Agile values at POD level.
• Kanban process might not be for all PODs.
5. Agenda
• Scrum process challenges for AI POD/ Store
POD
• Kanban Vs Scrum in a nutshell
• Kanban in a nutshell
• Kanban Principals
• Kanban Board( Visualizing Workflow)
• Kanban Metrics
• Your Feed Back through Idea Board
6. Scrum process challenges for AI POD
AI/ Seamless PODs Structure:
• 10 engineers as core team
• 5 engineers as helpers
• 3 technical managers
• Consist of two pods ( AI and Oracle Retail)
• Supporting 15 to 20 projects in each iteration
7. Scrum process challenges for AI POD
We never fully committed scrum:
• Priority changes impacted the sprint plan
• We never finished what we planned for.
• We never had an iteration planning that we
could commit
• We never had a defined sprint backlog (
stories coming to sprint everyday)
8. Scrum process challenges for AI POD
• We didn’t have the ability to plan in release
level because we didn’t know all application
requests and they were not prioritized before
they get to POD.
• We had more user stories added during sprint
than when the sprint started.
9. Scrum process challenges for AI POD
• Metrics in Sprint Reports were not effective:
– Mainly based on velocity of entire POD capacity
but in reality each engineering practice has it is
own capacity and velocity.
– User story sizing was not effective( place holders
with 10 points staying in progress for two weeks)
so measuring velocity and cycle time was not
based on realistic data.
12. Kanban vs Scrum in a Nutshell
• Responsiveness to Change: When priorities
change very frequently, Kanban is ideal (15 to
20 projects needs support from AI POD in
each sprint. Their priorities changes and there
are last minute requests. Commitment to two
weeks delivery is not feasible).
• Removing 2 weeks commitment for work.
13. Kanban vs Scrum in a Nutshell
• Removing Sprint backlog and just have a
general backlog
• Removing sprint planning meeting. Story
huddles and backlog grooming will continue
• Retrospective meetings will continue
• Scrum measure velocity as main metric for
improvement opportunities, Kanban measure
cycle time as main metric for improvement
opportunities
14. Kanban in a Nutshell
Kanban Principals:
– Visualize the workflow
– Limit Work-in-Progress (WIP)
– Manage Flow( Limited to Jira GAP workflow)
– Make Policies Explicit( Working Agreement)
– Implement Feedback Loops
– Improve Collaboratively, Evolve Experimentally
15. Kanban in a Nutshell
Kanban board( Visualizing workflow)
– Identify Sources of Work
– Identify Types of Work ( Class Of Services)
– Identify Outputs
– Map your Workflow
– Design your Board
– Create Policies for each column on the board
– AI Pod Board:
https://jira.gid.gap.com/secure/RapidBoard.jspa?rapidView=4230
– Store Pod Board:
https://jira.gid.gap.com/secure/RapidBoard.jspa?rapidView=3910
16. Kanban in a Nutshell
Kanban Metrics:
– Cycle Time – The time between starting on the
work and when item is ready for delivery.
• Cycle Time Depends on the Rate of the Bottleneck.
• The rate of the bottleneck is the rate of the entire
system.
– Lead Time – The time it takes to complete a
particular unit of work. The total time the client is
waiting for an item to be delivered.