Kanban in a SAFe/SCRUM World

Facilitator: Mike Lonergan

• Canonical example: an escalation or operations team
• First, measure what proportion of the team's time is spent on unplanned (Sprint Planning) work
• Ask: are escalations always the priority?
• If not, how do you decide which get put off?
• What are the prioritisation characteristics?
• Do the execs/stakeholders agree on the prioritisation approach?
• Assertion: SAFe does not require SCRUM/exclude Kanban
• Kanban requires metrics - usual metrics are:
• Cycle time
• Lead time
• WIP

• 
  

• If you pursue SCRUMban instead of SCRUM, keep the Retros and keep the Demos on a regular schedule
• Ask dependent teams: how long does it usually take to get your dependencies (stuff delivered from my team) met?
• Question for the team: is it possible to reserve any capacity for this team in the next Program Increment?
• Research: Look at your historical data (reactive vs features)
• Research: what constraints exist that keep team from improving cycle time of reactive issues

• Ask dependent teams: how unpredictable (how variable) has been the cycle time in getting dependencies met?
• Ask my team: what is working well?
• Research: what proportion of features in past sprints have had a bounded calendar deadline?
• Read PDF: Essential Kanban Condensed (David J. Anderson)
• Situational analysis: "ladder of inference" (study and decompose our assumptions)
• Computer Associates published a study on WIP limits and productivity
• Rule of thumb: start WIP limit at half the team size
• Challenge: break your dependencies