Skip to content

Work in Progress Limits

Summary

Limiting Work In Progress (WIP) increases throughput, improves predictability, and reduces stress by finishing the most valuable work sooner instead of diluting focus across many partially‑done items.

Why Limit WIP

  • Faster Completion: Partially done work delivers zero value and finishing earlier accelerates feedback loops.
  • Lower Context Switching Cost: Humans pay a heavy cognitive tax when juggling many threads, and fewer active tickets equals higher sustained quality.
  • Expose Bottlenecks: When a column hits its limit you must improve flow (help upstream) instead of silently starting more work.
  • Enable Swarming: Reserved capacity makes it cheap to swarm critical incidents without derailing weeks of in‑flight work.
  • Reduce Stress: Predictable small queues beat long hidden backlogs of half‑finished tasks.

Policies

  • Individual Limit: A person’s name should appear as active Shepherd on at most 2 tickets.
  • Column Limits: All intermediate columns have numeric limits sized to team capacity. If the next column is full you stop and help a ticket ahead of you finish. The execution columns should sum to 1.5 * Team Size as a starting point.
  • Interrupt Rail Limit: Only one active Interrupt at a time. Others wait in Ready which protects the Planned rail from interruptions.
  • Swarm Handling: Swarm items temporarily override individual limits. Once stabilized, excess helpers peel off.
  • Ticket Size: Tickets must be small enough to fit under the SLA (commonly < 2 weeks) so WIP represents near‑term deliverables, not projects. Small tickets mean that someone is frequently becoming idle and free to work on Swarms, etc.

Managing Flow in Practice

Look for columns consistently at or over limit, aging tickets compared to neighbors, repeated backward movement, and spikes in blocked markers.

When a limit blocks movement:

  1. Highlight the blockage in standup.
  2. Redirect available capacity to tickets closest to Done.
  3. If this is a chronic condition, adjust the process and culture not just the WIP number.

Flow Theory (Lightweight)

Little’s Law:

Average Throughput ≈ WIP / Average Cycle Time

For a fixed WIP, reducing cycle time increases throughput. For a fixed cycle time, excess WIP only increases delay. Therefore we cap WIP intentionally and focus on finishing work rather than starting more.

Anti‑Patterns

  • Increasing WIP limits instead of solving root causes.
  • Large tickets that “camp” in Engineer for weeks.
  • Hidden parallel work and side branches not visualized on the board.
  • Interrupt rail frequently with more than one active ticket.

← Step 7: Meetings | Step 9: Communication →