You added WIP limits to your board because a blog told you to. Then you picked the number out of thin air. Three is fine, right? Maybe five. Nobody knows, so the limit gets ignored by Wednesday and the "In Progress" column fills up like a sink with no drain.
Let's fix the guessing part. This post gives you a kanban WIP limit calculator, the math it runs on, and a way to keep the limit honest once it's set.
Key takeaways
- A WIP limit is a hard ceiling on how many cards live in a column at once. Its job is to expose bottlenecks, not to look tidy.
- The most defensible starting limit comes from Little's Law: WIP = throughput x cycle time. Match the limit to the flow you already have.
- A common practical floor is one to two cards per person in active columns. Start there if you have no flow data yet.
- Limits that are too high do nothing. Limits that are too low starve people. The calculator helps you land between the two.
- Setting the number is step one. The harder part is noticing when a column quietly blows past it.
What a WIP limit actually does
Work in progress (WIP) is every card a stage is holding at once. A WIP limit caps that count. When a column is full, you can't pull new work into it until something leaves.
That constraint feels annoying, and that's the point. A full column forces a conversation: why is nothing moving? Usually the answer is an upstream stage dumping work faster than the next stage can clear it. Without a limit, that pile-up hides in plain sight. With one, it stops the line and gets looked at.
So a WIP limit isn't a productivity hack. It's a smoke detector. The number matters because a detector set too sensitive goes off constantly, and one set too loose never goes off at all.
The kanban WIP limit calculator
[Interactive calculator embeds here. Enter your numbers below and it returns a suggested per-column limit.]
The calculator asks for four things:
- People working the active stages (not the whole org, just the folks who pull cards).
- Throughput: cards your team finishes per week, averaged over the last month or so.
- Target cycle time: how long, in days, you want a card to take from start to done.
- Number of active columns between "To Do" and "Done".
It returns a suggested limit per active column, plus a board-wide WIP ceiling. If you'd rather do it by hand, the next section is the same math without the form.
The formula behind it
The calculator runs on Little's Law, the one piece of queueing theory every kanban practitioner should keep in their pocket:
Average WIP = Average Throughput x Average Cycle Time
Say your team finishes 10 cards per week and a card takes 5 working days end to end. Convert throughput to the same unit as cycle time (10 cards / 5 working days = 2 cards per day). Then:
WIP = 2 cards/day x 5 days = 10 cards in flight, board-wide.
Spread 10 cards across, say, two active columns and you get a starting limit of about 5 per column. That's your board-wide WIP working backward into a per-column cap.
Now flip it. If you want a faster cycle time, lower the WIP. Drop the board ceiling to 6 cards and, at the same throughput, your cycle time falls to 3 days. Less work in flight, faster delivery. That trade is the whole reason WIP limits exist.
No flow data yet? Use the per-person rule
New board, no history to average? Start with the crew-size heuristic: one to two cards per person in each active column. A team of four working a single "In Progress" column starts around a limit of 4 to 6. It's rough, but it's grounded in something real, which already beats picking a round number you like the look of.
Run it tight for two weeks, watch where cards pile up, then recalculate with Little's Law once you have actual throughput.
Tuning the number after you set it
No limit is right on the first try. Here's how to read what your board tells you:
- The column is always full and cards barely move: the limit might be too high, or the next stage is the real bottleneck. Lower the cap on the downstream column first.
- The column is rarely near its limit: the cap is doing nothing. Tighten it until it occasionally pinches. A limit that never bites isn't a limit.
- People sit idle waiting for work: too low. Loosen by one and watch.
Adjust one column at a time. Change five things at once and you'll never know which one helped.
Putting your limit to work in Zoobbe
A number in a spreadsheet doesn't manage flow. A board does. Once the calculator gives you a per-column limit, build your stages as columns on a kanban board and run real cards through them.
The honest hard part of WIP limits is enforcement. It's easy to agree on a cap and then quietly break it on a busy Thursday. Zoobbe's AI Insights include bottleneck detection and workload balancing, so when a column starts swelling past the number you calculated, you get told instead of finding out three sprints later in a retro. The AI will even suggest tightening WIP when it spots a stage backing up.
You can also wire automation rules to fire when cards move between stages, so the moment work shifts you have a notification trail rather than a silent pile-up. The calculator gives you the target. The board keeps you honest about hitting it.
FAQ
What is a good WIP limit for a kanban board?
There's no universal number. Start with Little's Law (throughput x cycle time) if you have flow data, or one to two cards per person per active column if you don't. Then tune until the limit occasionally stops you, which means it's actually working.
How do I calculate WIP limits with Little's Law?
Multiply your average throughput by your average cycle time, in matching units. Ten cards a week at a 5-day cycle time means roughly 10 cards in flight board-wide. Divide that across your active columns for a per-column starting cap.
Should each column have its own WIP limit?
Active columns like "In Progress" and "Review" benefit most, because that's where work stalls. Queue columns like "To Do" and terminal columns like "Done" usually don't need a cap.
What happens when a WIP limit is reached?
You stop pulling new work into that column until a card leaves it. That pause is the feature. It pushes the team to finish what's started and surfaces whatever is blocking flow downstream.
Can WIP limits be too low?
Yes. If people sit idle waiting for permission to pull work, the limit is starving the team. Raise it by one and observe. The goal is steady flow, not maximum restriction.
Pick your number with the calculator, build your columns, and let the board tell you when the limit needs another nudge.
Photo by Patrick Perkins on Unsplash