Your scheduling board shows 70% utilization. Your production manager says the floor is maxed out. Your sales team keeps promising two-week lead times because “we have bandwidth.” Someone is working with bad numbers — and it’s probably everyone.
Capacity planning sounds straightforward: machine hours available minus machine hours booked equals open capacity. In practice, digital manufacturers rarely hit that simple math. The gap between theoretical capacity and actual throughput is where margins evaporate, lead times stretch, and customer trust erodes.
Why Theoretical Capacity Lies
Most shops calculate capacity by multiplying machine count by shift hours. A five-axis mill running two shifts? That’s 112 hours a week of “capacity.” But that number assumes:
- Zero setup or changeover time
- Zero maintenance windows
- Zero tool changes or probe cycles
- Zero first-article inspection delays
- Zero material staging or workholding time
- 100% operator attendance and efficiency
None of those hold. A 2023 study by the Association for Manufacturing Technology found average actual spindle utilization across CNC shops sits between 25-35%. The rest is “non-productive time” — necessary work that doesn’t show up on a capacity spreadsheet.
The Hidden Time Thieves
Setup and changeover are the biggest offenders. A “four-hour job” might need two hours of workholding design, fixture building, tool presetting, and first-piece inspection. That’s 50% overhead before a single chip flies.
Then there’s the cascade effect. One late material delivery pushes Job A into Job B’s slot. Job B’s operator waits. Job C’s setup gets rushed, causing a rework. Suddenly your 70% utilization day produced 40% good parts.
Additive shops face parallel ghosts: build plate prep, support removal, heat treatment scheduling, and post-process queueing. A 40-hour print might consume 60 hours of calendar time when you count the full workflow.
How Bad Data Distorts Quoting
When your capacity model is optimistic, your quotes inherit that optimism. You quote three-week lead times based on “open slots” that don’t exist. You price jobs assuming 85% machine efficiency when reality delivers 40%.
The result: you win the order, then scramble. Overtime costs eat margin. Expedited shipping kills profit. The customer gets their parts late — and remembers.
Conversely, pessimistic capacity models (padding everything 2x) make you uncompetitive. You quote six weeks when the shop could deliver in three. The customer goes elsewhere.
Building Capacity Data That Reflects Reality
Start by tracking actuals, not plans. For every job, capture:
- Calendar time from material receipt to ship
- Machine-on time vs. machine-available time
- Setup/changeover hours by operation type
- Rework and scrap rates per process
- Wait states: material, inspection, post-process, shipping
Run this for 20-30 jobs across your mix. Patterns emerge. You’ll learn that 3-axis milling averages 1.8x setup-to-run ratio. That your SLS printer spends 35% of calendar time in cooling and depowdering. That Thursday afternoons lose 15% throughput to maintenance.
Feed those ratios back into your quoting engine. When a new RFQ arrives, the system applies real-world multipliers — not wishful thinking.
Connecting Quoting to Floor Reality
The shops that close this loop don’t use spreadsheets. They use a system where quoting, scheduling, and execution share the same data model. When a quote is built, it pulls current machine availability, historical setup times for that process, and real throughput rates. When the job hits the floor, actuals feed back automatically — no manual entry, no Friday afternoon data cleanup.
That’s the architecture behind Solvi: instant quoting tied to a live MES, so capacity assumptions update with every completed job. The quote engine learns from the floor. The floor executes what the quote promised.
Start With One Process
You don’t need a full overhaul tomorrow. Pick your highest-volume process. Track actual calendar time vs. quoted lead time for the next 15 jobs. Compare machine-on hours to available hours. Note every wait state.
That dataset becomes your baseline. Next quote for that process uses real numbers. Margin improves. Lead times become credible. The floor stops drowning.
Then move to the next process.
Capacity Is a Living Number
Machine upgrades change it. New hires change it. Process improvements change it. A capacity model from six months ago is already wrong.
The shops that quote accurately and deliver on time treat capacity as a measured, updated metric — not a static calculation. They build systems that capture reality automatically and feed it forward.
Your scheduling board says 70%. Your team says 100%. The truth is in the timestamps. Start capturing them.
Solvi
Never turn work away
Post overflow with the price and terms attached, or fill idle machines with jobs from other shops. Free to post, free to join, no commission.