Kanban capacity planning helps a team decide what it can responsibly start, not simply what it would like to finish. This weekly method shows how to calculate usable capacity, compare it with committed work, account for meetings and interruptions, and adjust a kanban board before overload becomes a delivery problem.
Overview
Capacity planning is the practice of matching planned work to the time and attention a team actually has available. In a kanban system, it complements flow-based practices such as visualizing work, limiting work in progress, and measuring cycle time. The goal is not to create a rigid promise for every task. It is to make a transparent, revisable decision about the work most likely to move through the system during the next planning period.
A useful weekly process answers four questions:
- Who is available to work, and for how much of the week?
- How much time should be reserved for meetings, support, administration, and interruptions?
- What work is important enough to commit to first?
- What should be deferred when planned demand exceeds usable capacity?
Capacity is not the same as utilization. Filling every available hour leaves no room for defects, questions, incidents, or the coordination required to finish work. A practical plan includes a buffer, especially for technology teams that handle production support or incoming requests.
Your kanban board should make the assumptions visible. Useful fields include owner, work type, size or estimate, priority, due date, blocked status, and expected effort. A separate availability field or weekly planning view can record leave, part-time schedules, shared responsibilities, and planned support rotations. For a broader review of the board itself, use a kanban board audit checklist.
How to estimate
1. Set the planning period
Use one week for near-term planning. A weekly horizon is short enough for availability to be reasonably known and long enough for a team to move a meaningful set of work. Longer forecasts can use the same method, but should rely more heavily on historical throughput and ranges rather than detailed task-level commitments.
2. Calculate gross availability
For each person, multiply scheduled workdays by scheduled hours per day. Then adjust for part-time schedules, leave, public holidays, or other known absences.
Gross availability = scheduled workdays × scheduled hours per day × availability percentage
If a developer normally works five eight-hour days but will spend two days on planned leave, gross availability is 24 hours: three days × eight hours.
3. Remove fixed commitments
Subtract recurring meetings, planning sessions, on-call duties, training, and other commitments that cannot be used for the planned work. Keep the calculation honest: if a meeting usually creates follow-up work, include a reasonable allowance for that work as well.
Net working capacity = gross availability − fixed commitments
4. Reserve an interruption buffer
Apply a buffer to net capacity rather than planning against every remaining hour. For example, a team may choose to reserve 15% to 25% for support, defects, coordination, and unexpected work. The appropriate percentage depends on the team’s operating conditions; use a stated assumption and revise it using experience.
Usable capacity = net working capacity × (1 − buffer percentage)
5. Estimate committed work consistently
Estimate work in hours only when the team can make reasonably consistent effort estimates. Otherwise, use relative sizes such as small, medium, and large, and compare the planned set with historical throughput. Avoid mixing points from one team or workflow with hours from another unless the relationship has been deliberately calibrated.
For an hour-based plan, add the estimates of work that is ready, prioritized, and expected to enter active work during the week. For a throughput-based forecast, review how many items the team normally finishes in a comparable period, then select a conservative range rather than treating the average as a guarantee. A WIP limit can prevent the team from starting more work simply because capacity appears available on paper.
6. Compare demand with capacity
Calculate the difference between usable capacity and committed work:
Capacity gap = usable capacity − committed work
A positive result indicates planned spare capacity, although it does not necessarily justify adding another item. A result near zero calls for careful monitoring. A negative result means the commitment exceeds the stated capacity and requires a decision: reduce scope, defer an item, assign the work differently, or revise the assumptions with an explicit reason.
Inputs and assumptions
Use the following worksheet during weekly planning. Keep the inputs on the board, in a linked document, or in a repeatable planning view so that the calculation can be revisited without rebuilding it.
| Input | What to record | Why it matters |
|---|---|---|
| Planning period | Start and end dates | Defines the availability being measured. |
| Scheduled time | Workdays and hours per person | Establishes gross availability. |
| Absences | Leave, holidays, appointments, or other known gaps | Prevents overestimating availability. |
| Fixed commitments | Meetings, support shifts, training, and operations work | Separates planned delivery time from reserved time. |
| Buffer | A stated percentage or number of hours | Leaves room for unplanned work and coordination. |
| Ready work | Prioritized tasks with owners and estimates | Defines the candidate commitment. |
| Dependencies | Blocked items, approvals, or external inputs | Stops nominal capacity from being mistaken for flow capacity. |
Use assumptions that match the workflow. A developer assigned to incident response may need a larger interruption buffer than someone working on a protected project. A shared database administrator may have several teams drawing on the same capacity, so requests should be ranked across the whole queue rather than accepted separately by each team.
When work is frequently interrupted, track the interruption category on the kanban board. Distinguishing support, defects, unplanned requests, and blocked work makes the next estimate more useful. If tasks arrive from meetings, a reviewable meeting-notes-to-action-items workflow can capture them without automatically turning every discussion into a weekly commitment.
Worked examples
Full-time team member
A full-time engineer has five eight-hour workdays, two hours of recurring meetings, and a 20% interruption buffer.
| Calculation | Result |
|---|---|
| Gross availability: 5 × 8 | 40 hours |
| Less fixed commitments | 38 hours |
| Usable capacity: 38 × 0.80 | 30.4 hours |
The team should normally commit to no more than about 30 hours of estimated work for this person. If the candidate tasks total 36 hours, the difference is not solved by changing the label on a task. The team should defer lower-priority work, reduce scope, or identify another owner.
Part-time team member
A part-time analyst works three six-hour days. One hour is reserved for recurring coordination, and the team uses a 15% buffer.
| Calculation | Result |
|---|---|
| Gross availability: 3 × 6 | 18 hours |
| Less fixed commitments | 17 hours |
| Usable capacity: 17 × 0.85 | 14.45 hours |
Round the result to a practical planning value, such as 14 hours, and avoid assigning work that depends on availability outside the person’s scheduled days. Recording working days on the board is as important as recording total hours when handoffs or response times matter.
Shared team member
A platform specialist has 32 available hours after leave and fixed commitments. The specialist supports two teams, with 12 hours already reserved for operational work. Applying a 20% buffer to the remaining 20 hours produces 16 usable hours for new planned requests.
| Calculation | Result |
|---|---|
| Availability after fixed commitments | 32 hours |
| Less operational allocation | 20 hours |
| Usable new-work capacity: 20 × 0.80 | 16 hours |
If Team A requests 12 hours and Team B requests 10 hours, the combined demand is 22 hours against 16 available hours. Use a shared priority rule to select work, rather than allowing both queues to assume full access. A task intake and prioritization matrix can help make that trade-off explicit.
For client-facing or multi-stage workflows, map capacity to the relevant workflow stage, not only to individuals. A team may have enough development hours but insufficient review, testing, or approval capacity. The client onboarding kanban workflow illustrates why stage-specific ownership and queues matter.
When to recalculate
Recalculate capacity at the start of each weekly planning cycle and whenever a material assumption changes. Revisit it immediately when someone takes leave, joins or leaves the team, changes working hours, enters an on-call rotation, or becomes responsible for a new operational queue.
Also recalculate when committed work changes significantly. A production incident, urgent defect, dependency failure, or newly approved priority can consume capacity that was previously assigned to planned work. Move deferred items back to the backlog with a clear reason instead of leaving them in progress without an updated expectation.
Every four to six weeks, compare planned capacity with actual outcomes. Look for patterns:
- Planned hours consistently exceed completed work.
- Interruptions regularly consume more than the stated buffer.
- Tasks remain active because they are too large or poorly defined.
- A particular workflow stage becomes a recurring bottleneck.
- Estimates are accurate for some work types but not others.
Use those observations to adjust the buffer, split work, improve intake criteria, or change WIP limits. Do not use the review to pressure individuals into higher utilization. Capacity planning is most useful when it improves the team’s decisions and makes trade-offs visible.
To use this method in your next planning cycle, create a weekly board view with availability, fixed commitments, buffer, usable capacity, and committed work. Enter each person’s current inputs, total the demand by workflow stage, and defer the lowest-priority items whenever demand exceeds capacity. Revisit the calculation after the week ends, record what changed, and carry the revised assumptions into the next cycle. That small feedback loop turns a static capacity planning template into a practical team workload management system.