Visual Board: Category Boundary
Visual Board establishes the first dependable fact in the process. It should represent work and policy on a shared visual board. For this article's focus on where the category's responsibility stops and the alternative's responsibility begins, cycle time is the quickest way to see whether unbounded work in progress is being caught early enough.
- Show the exact source that feeds Visual Board and explain why it is authoritative.
- Create unbounded work in progress before the demonstration begins; do not repair it in advance.
- Record the starting value for cycle time and the person responsible for responding.
Work Item: Primary Job
Work reaches Work Item after the initial record exists. Its job is to make entry and exit criteria explicit for each workflow state, without blurring who owns the next decision. Watch throughput while deliberately introducing stale cards; the behavior of that handoff reveals more than a feature list.
- Have one operator make entry and exit criteria explicit for each workflow state while another observes the handoff.
- Delay or interrupt Work Item and note which queue, alert, or owner becomes visible.
- Compare throughput before and after the interruption instead of relying on impressions.
Workflow State: Shared Handoff
Workflow State is the point where the system changes or enriches the working state. A credible design can pull new work only when capacity is available and still leave the earlier facts recoverable. If hidden queues appears, work-item age should expose the problem before downstream teams rely on it.
- Trace one representative record into, through, and out of Workflow State.
- Change a key value and verify that the earlier state remains explainable.
- Use work-item age to decide whether the transformation is complete and timely.
Work-in-Progress Limit: Different Control
Work-in-Progress Limit marks a business boundary, not merely another screen. The platform must limit concurrent work to expose congestion under an explicit rule. Test the boundary with expedite abuse, then determine whether WIP limit breaches gives the approver enough context to accept, reject, or reroute the case.
- Name the role allowed to approve the decision at Work-in-Progress Limit.
- Attempt an out-of-policy action and inspect the denial or escalation path.
- Require the approver to justify the outcome using retained facts, not memory.
Pull Signal: Failure Test
Pull Signal becomes important when ordinary processing stops being ordinary. It needs to flag blocked aging and expedited items while preserving the unresolved condition. A buyer should examine how unbounded work in progress is surfaced and whether cycle time changes soon enough for a responsible person to intervene.
- Stage unbounded work in progress during normal volume and observe how quickly it becomes actionable.
- Follow the exception until a named person accepts responsibility for it.
- Verify that correction improves cycle time without hiding the original failure.
Flow Policy: Decision Evidence
Flow Policy closes the loop by making the outcome visible to the next participant. It should improve flow using observed cycle and queue behavior and retain enough history to explain what happened. Use throughput to confirm recovery from stale cards, then ask a second reviewer to reconstruct the decision independently.
- Give the completed case to someone who did not participate in the test.
- Ask that reviewer to explain the sequence, decision, and remaining uncertainty.
- Accept the result only when throughput reconciles with the source and destination records.
