Capacity isn’t the bottleneck. Coordination is.
When a board is late, the instinct is to blame the factory. Look at the calendar of a typical build and most of the days are spent somewhere else.
October 14, 2026
Ask a hardware team why a prototype run slipped and you will usually hear about capacity: the fab was full, the assembler had a queue, a line was down. Sometimes that is true. More often the elapsed time is made of smaller waits that never appear on a production schedule.
Where the calendar goes
A PCB order passes through quoting, design-for-manufacturing review, component sourcing, fabrication, assembly, test and shipping. Only a few of those steps involve a machine making something. The rest involve people asking questions of other people: is this drill size acceptable, can this part be substituted, which stackup do you want, is the BOM current.
Each question is cheap to answer and expensive to wait for. A one-day reply cycle repeated across three or four exchanges adds days that look like “lead time” but contain no manufacturing at all.
Sequential handoffs hide parallel work
Fabrication and component sourcing are independent. Boards can be built while parts are being bought, and neither needs to wait for the other until assembly. In practice they are often sequenced because one party owns the quote and the next step starts only once the previous is confirmed.
The fastest factory in the world cannot start a job that has not been fully specified.
What coordination can do
Check early and automatically
Most manufacturability issues are mechanical: a trace-to-space value below a process limit, a missing drill file, a footprint that disagrees with the placement data. Rules like these can be checked as soon as files arrive, so the first question a designer sees is the last one they would otherwise have received on day four.
Route by fit, not by habit
Not every factory is equally good at every board. A fine-pitch HDI board and a simple two-layer panel have different best homes. Matching jobs to capability and available capacity keeps specialised lines busy with the work they are best at, and keeps simple jobs from queueing behind complicated ones.
Make status visible
A large share of status emails exist because nobody can see where a job is. Shared, current status removes a category of waiting and a category of work.
A fair caveat
Capacity does matter, particularly for specialised processes and during demand spikes. The claim here is narrower: before adding capacity, it is worth measuring how much of an order’s elapsed time is spent on the line at all. For many prototype builds the answer is a minority.