Every delivery date you promise is, quietly, a bet on capacity: that the right machines will have the right hours free to make the right parts in time. Capacity planning is how you stop guessing and start making that bet on purpose.
This guide covers the core ideas — load vs. capacity, takt time and the bottleneck — and how to move from a fragile spreadsheet to a plan that stays honest as reality changes.
What capacity planning actually answers
Strip away the jargon and capacity planning answers one question: can we make what we've committed to, on time, with the resources we have? If yes, by how much margin. If no, where exactly does it break, and what would it take to fix — overtime, an extra shift, a different sequence, or a later promise.
The core idea: load vs. capacity
Two numbers, compared:
- Capacity — how much you can make in a period. Roughly: working time × number of machines × a realistic efficiency. (A machine scheduled for 80 hours that truly runs at, say, 85% effective time gives you ~68 productive hours, not 80.)
- Load — how much you've been asked to make. The sum, across products, of quantity × cycle time.
When load is comfortably under capacity, you can deliver. When load exceeds capacity — even on one resource — something has to give, and the plan should tell you that before the week starts, not after it fails.
Takt time and the bottleneck
Takt time is the rhythm demand sets: available production time ÷ the quantity the customer needs. If you have 450 productive minutes and need 90 parts, takt is 5 minutes — one part must leave every 5 minutes to stay on schedule.
But a line only moves as fast as its slowest stage. The bottleneck — the resource with the least capacity relative to its load — sets the throughput of the whole line. An hour gained anywhere else is an illusion; an hour gained at the bottleneck is real output. Good capacity planning finds the bottleneck first.
Finite vs. infinite scheduling
A lot of "plans" are built with infinite capacity assumptions: they schedule work as if every resource could take any amount at once. It's optimistic and it's why plans slip. Finite capacity planning respects the real limit of each resource — it won't load a machine to 130% and pretend that's fine. The difference is the difference between a wish and a schedule.
Multi-stage reality: pull-through and buffers
Few plants make one thing in one step. A finished product pulls demand back through its bill of materials (BOM): to ship 1,000 assemblies you need 1,000 sub-assemblies, which need components, each with its own machine and cycle time. Downstream demand drives upstream load.
Between stages sit buffers — stock that absorbs variability so an upstream hiccup doesn't instantly starve the next step. A good plan accounts for what's already in the buffer, so you don't over-produce parts you already have.
Why spreadsheets break
Most capacity planning still lives in a spreadsheet, and for a first pass that's fine. But spreadsheets are static and fragile: they don't recalculate when demand shifts, they hide errors in copied formulas, they can't easily answer "what if we add a Saturday shift?", and by Wednesday they're out of date. The plan becomes a snapshot of last Friday's intentions.
From spreadsheet to a live schedule
The fix isn't a six-figure system — it's putting the same logic somewhere live and visual. Production Plan does exactly this in Power BI: it builds a capacity-aware, shift-by-shift schedule from your products, goals and cycle times; it explodes the BOM so downstream goals drive upstream demand; it accounts for buffers and flags supply risk; and it shows the whole flow on a pipeline map — all without DAX. The spreadsheet logic you trust becomes a plan that updates with your data.
The bottom line
Capacity planning is just load vs. capacity, honestly compared, with the bottleneck in focus and the BOM respected. Do that on paper once to understand it — then put it somewhere live, because a plan is only useful while it's still true.