If the team has nine active tickets and keeps opening new ones, the best place to look may be the small column near the end. Count the items that have started and still need something. Then check which active items have been waiting the longest.
The Kanban Guide defines work in progress as items between entry into the started state and a finished state. Work item age measures elapsed time since an item's start while it is open. Cycle time uses the finish date, so it describes completed work. You can combine current ages with the recent times for similar finished work to decide which tickets deserve attention today.
Read an example board
| Stage | Active items | Oldest item age | What to ask |
|---|---|---|---|
| Building | 3 | 4 days | Is someone working on each item? |
| Review | 4 | 9 days | Who can clear the oldest review? |
| Delivery prep | 2 | 3 days | Is a release decision holding both items? |
Calculate age from the original start. Moving a card between columns leaves that date unchanged. The review stage is the first place to inspect in this example because it holds the most cards and the oldest one. It may still be a symptom of a missing reviewer, an approval that needs to be scheduled, or different work types sharing the same lane.
Try one change for a full review cycle
Set a provisional cap for the crowded stage. Atlassian's guide to WIP limits describes setting and adjusting a stage limit to surface bottlenecks. Agree on what the team does when it reaches the cap, such as pairing on an old review before pulling another build ticket. Keep a blocked item visible while it is consuming attention or occupying a started state.
Have the team agree how it will count elapsed time before comparing ticket ages. A calendar-day measure includes the weekend for every item. An internal report using working days will produce different numbers. Use the same convention when comparing open work to your recent finished items.
Check whether several old items share a reason
Take the four review cards from the sample and write down the last real action on each. The oldest has been waiting since a customer asked for new terms. Two other cards are waiting for a peer to approve a copy edit. The fourth moved into review this morning. A larger review column describes the queue, but a common solution may apply to only the middle two cards. Give those two to someone with the ability to review them, and route the oldest to the person who can resolve the contract issue.
Keep the board status and the actual dependency visible. Use a note for “waiting on the customer” with the date the question was sent and who is following up. If the ticket is already started under your team's agreement, keep counting it in the age and WIP measures even while it waits. Explain how you treat external review so a shorter internal review does not get interpreted as a longer one merely because they entered the same column.
Review the boundary of finished work. If delivery prep must complete before a customer can use the result, moving a ticket there has not finished it under that definition. Compare starts and finishes over the same review period. If the team started five items and finished only two, total active work has risen by three, assuming no cancellations or reclassifications. Look across the entire board to see where those items went.
Make one small intervention
Choose the provisional cap with the people who work in that stage. With four tickets in review, you could try a cap of three for the next week. Discuss the action to take when the count is three or higher. For example, hold new work that would need the same reviewer while one person handles an old approval. Keep existing cards visible when you begin the trial. Reassigning or hiding cards merely to get below the cap will make the comparison meaningless.
Pick one review moment and write down the question to answer. If the test is intended to reduce waits for peer reviews, compare those tickets' ages and completions with the previous week under the same convention. Check the downstream queue and whether other work was delayed while a reviewer helped. If the review stage emptied while delivery prep grew by the same amount, inspect the last stage before keeping the cap as a permanent rule.
Decide what evidence calls for a different change
| Pattern over the trial | Next check |
|---|---|
| Review stays full, and most cards wait for one specialist | Ask whether that person has time reserved for approvals or a trained substitute. |
| A long customer approval blocks one lane, while peer reviews move | Track the external wait explicitly and compare it with similar work. |
| Review empties, but finished output stays flat | Check the next stage and definition of finished before declaring the change successful. |
| Open work keeps growing at every stage | Compare starts and finishes, then limit commitments at intake. |
Bring a single card to the review meeting and ask what could move it today. The answer may be a call to a client, a handoff to a second reviewer, or a decision to return an incomplete request to its owner. If too many jobs begin with missing details, use the request form rehearsal to check which information is disappearing before work starts. Keep one week of observations with the board so the team can tell whether the change improved completed work under its actual conditions.
A card can vanish from review because it was cancelled. Record that outcome separately from successful completion, and decide how your board represents a restarted item before comparing weeks. If a card goes back to building for the same deliverable, preserve its original start so rework stays visible. If a request is replaced by genuinely different work, close the earlier item with a note and create a new one under the agreed policy. During the trial, use a small log for such exceptions. A small change in completed count may simply reflect how an unusual item moved through the workflow.



