Capacity Planner
Before you add labor or equipment, run the math. The bottleneck is rarely where you think it is. Machine, labor, and WIP/material capacity checked station by station against takt time, with what-if scenarios modeled before you spend a dollar.
Problems This Solves
Four ways capacity decisions go wrong
Three Dimensions
Every capacity model needs all three
Most capacity analyses look at only one dimension. A cell with plenty of machine time can still be constrained by labor, or by material that doesn't arrive. The skill addresses all three before it says which one limits the process.
Machines don't run at 100%. A machine available 8 hours per day with a 72% OEE has 5.76 productive hours, not 8. Where OEE data exists, the skill shows the theoretical and the OEE-adjusted analysis side by side. Using nameplate capacity makes the machine look 39% more capable than it is in this example.
Bottleneck
Only one station is the real constraint
System capacity equals the bottleneck's capacity. Adding capacity anywhere else won't increase output. This is the most common and most expensive mistake in capacity planning.
For each station, the skill calculates the capacity ratio: station cycle time ÷ takt time. Above 1.00 the station cannot meet takt, 0.85 to 1.00 is at risk, and below 0.85 there is buffer. The station with the highest ratio is the bottleneck, and the skill says whether machine, labor or WIP/material limits it. Cycle times are never averaged across stations.
What-If Scenarios
Which station breaks first when volume rises?
For each scenario the skill recalculates takt time and station loading from scratch, not as a percentage on top of the baseline. The first station whose ratio passes 1.00 shows what breaks and the action it requires.
Available time: 26,400 s per day (one 8-hour shift less breaks) · 4 stations, 1 operator each · Illustrative numbers.
| Station (cycle time) | Current: 200/day, takt 132 s | +10%: 220/day, takt 120 s | +20%: 240/day, takt 110 s | Breaks at |
|---|---|---|---|---|
| 1 Prep (95 s, labor) | 0.72 · OK | 0.79 · OK | 0.86 · At risk | +39% volume |
| 2 Press (108 s, machine) | 0.82 · OK | 0.90 · At risk | 0.98 · At risk | +22% volume |
| 3 Assembly (118 s, labor) | 0.89 · At risk | 0.98 · At risk | 1.07 · OVER | +12% volume ← bottleneck |
| 4 Test (80 s, machine) | 0.61 · OK | 0.67 · OK | 0.73 · OK | +65% volume |
Station 3 is already at 0.89 at current volume, and it is labor-bound: 118 s of work against a 104 s machine cycle. A 12% volume increase puts it over takt before any other station. The first move is to rebalance the labor content at Station 3, not to add an operator or a machine. Station 2 is the next constraint, at 244 units per day.
Capacity Gap Analysis
Required vs. available, for every dimension
The same cell at +10% volume (220 units per day).
| Dimension | Required | Available | Gap | Recommended Action |
|---|---|---|---|---|
| Labor (Station 3) | 118 s/unit × 220/day = 25,960 s | 26,400 s (1 operator × one shift) | +440 s | At risk (ratio 0.98). Rebalance work off Station 3 before adding anyone |
| Machine (Station 2) | 108 s/unit × 220/day = 23,760 s | 26,400 s | +2,640 s | At risk (ratio 0.90). Next constraint after Station 3 |
| WIP / material | [NEEDS GEMBA] | [NEEDS GEMBA] | [NEEDS GEMBA] | Not provided, so assumed not the constraint [ASSUMED]. Check for starved or blocked stations on the floor |
Examples
Manufacturing and clinical throughput planning
Assembly cell capacity model: 4 stations, 1 operator per station, 1 shift (26,400 s available after breaks). What-if at +10% and +20% volume. Station 3 identified as the bottleneck (labor-bound); it breaks at +12%.
Cell: Final Assembly
Stations: 4 · Operators: 1 per station · Shifts: 1 (26,400 s available)
Takt at current volume (200/day): 132 s
Bottleneck at current volume: Station 3, labor-bound (118 s), ratio 0.89, At risk
System capacity: 223 units/day (+12%)
Breaks at: +12% volume (about +24 units/day)
Recommended action: rebalance labor content at Station 3. Station 2 is the next constraint at 244 units/day
[NEEDS GEMBA: confirm Station 3 labor content with a timed observation; WIP and material not provided]Outpatient clinic throughput planning. The skill adapts vocabulary: "units" becomes "patient visits," "machine capacity" becomes "room capacity" (exam rooms × clinic hours ÷ average visit time, adjusted for how much of that time rooms are actually in use). Labor capacity models provider and MA time separately.
Room capacity can be the binding constraint in clinic throughput planning, not provider hours, so the skill checks both. Adding provider time without adding exam rooms shifts the bottleneck rather than relieving it. The skill models all three dimensions before any staffing or facility decision.
What you get
A real run on the sample data
Unedited output. The sample message at the top went to Claude Sonnet 4.6 (an earlier-generation model, so a newer one may word things differently) with this skill pasted in, the same way the steps below show, on September 29, 2026. Scroll inside the frame to read the whole reply.
Installation
Describe your cell. Get the capacity model.
resources/examples folder.Related Skills
Capacity planning connects to VSM and OEE
A VSM shows which steps have the most inventory and longest lead times. Capacity planning shows why. If machine capacity is the binding constraint, OEE analysis explains where the capacity is being lost.