Design an Upstream Kanban process for product ideas before development

WHAT THIS TESTS: Whether you understand Kanban as a system for balancing demand against capability across the whole value stream, not just the delivery portion. Interviewers want to see that you distinguish between uncommitted options and committed work items, and that you can design explicit policies that prevent poorly formed or low-value work from entering the downstream pipeline. They also want evidence that you understand how unmanaged demand creates congestion that destroys predictability regardless of how well the delivery team performs.

A GOOD ANSWER COVERS: First, visualization of the discovery funnel using an upstream Kanban board where ideas enter as options rather than committed work. Second, explicit triage policies that evaluate options against organizational goals, urgency, and available capacity, often using a triage table or similar framework. Third, a defined commitment point that separates discovery from delivery, ensuring work is only pulled downstream when it has passed a go or no-go decision and meets a ready threshold. Fourth, refinement and discard loops that allow the team to kill weak options early rather than letting them accumulate as expensive inventory. Fifth, the mechanism for predictability: by filtering noise and refining work before it hits development, the downstream system receives higher-quality inputs with reduced variability, preventing overload and enabling smoother flow and more reliable cycle times.

COMMON WRONG ANSWERS: Calling upstream Kanban just a backlog or a to-do list for product managers. Treating every idea as a committed work item rather than an option that can expire if it does not prove valuable. Ignoring WIP limits in discovery and allowing infinite ideas to pile up in front of the team. Failing to explain how upstream filtering affects downstream metrics like cycle time, throughput, or service level expectations. Suggesting rigid upfront planning instead of options-based decision making that preserves the ability to defer commitment.

LIKELY FOLLOW-UPS: How do you prevent the upstream board from becoming a graveyard of stale ideas? What specific policies would you use to sequence options when capacity is constrained? How do you align upstream triage with portfolio-level priorities and funding decisions? What signals tell you that an option is ready to cross the commitment point? How would you measure the health of the upstream system?

ONE CONCRETE EXAMPLE: A platform team receives fifty feature requests per quarter. Without upstream Kanban, all fifty enter the backlog and the team context-switches constantly between poorly defined initiatives. With upstream Kanban, requests enter an option-discovery board with a WIP limit of ten active options. Each option moves through triage columns like Intake, Vetting, and Ready to Commit. Policies require a defined business outcome and a rough t-shirt size before an option can advance. The product owner runs a weekly triage meeting using urgency and value criteria. Options that fail vetting are discarded or parked. Only when an option reaches the commitment point and a delivery team has capacity does it convert to a committed work item. Downstream, the team sees a 40 percent drop in unplanned work and cycle time becomes more predictable because the input rate now matches delivery capability.

Source: djaa.com