Spindle utilization is the number a lot of job shops reach for first, because it’s the easiest thing to see: is the machine cutting metal right now, yes or no. It’s a real, useful number, and it’s a different one from OEE or TEEP, see TEEP vs OEE vs utilization for how the three relate. It’s also, on its own, a misleading way to judge how well a job shop is actually running, because a high-mix shop’s economics depend on things spindle time doesn’t measure at all.
What spindle utilization actually tells you
Spindle utilization, or run time as a share of scheduled time, answers exactly one question: was the machine cutting when it was supposed to be able to cut. That’s worth knowing. A machine sitting idle when it should be running is losing capacity in the most direct possible way, and a low utilization number is a legitimate signal something’s wrong with scheduling, material flow, or staffing.
What it doesn’t tell you is whether the machine made money while it was cutting, whether the parts it made were any good, or how much of the day around that cutting time was spent on things a job shop can’t avoid: quoting, setup, fixturing, and waiting on the next job to actually show up on the floor.
The job shop reality spindle time misses
A single-part production line and a high-mix job shop have fundamentally different cost structures, and applying a single-part shop’s utilization expectations to a job shop sets a target that was never realistic to begin with.
Setup and changeover eat a real, unavoidable share of the day. A shop running twelve different parts a week on the same machine has fixturing, tool changes, and program verification baked into its schedule as a fact of the business, not a failure to eliminate. A machine with 65% spindle utilization because a third of its scheduled time went to legitimate setup work for different jobs isn’t underperforming the way a single-part line at 65% would be. It’s running high-mix work the way high-mix work actually runs.
Quoting and non-billable work don’t show up in spindle time at all. Time spent proving out a new part, running a first article, or cutting a sample for a quote that may or may not turn into an order is real machine time, real labor, and often real material cost, and none of it shows up as productive spindle time even though it’s essential to keeping the shop’s pipeline full. A shop that only watches spindle utilization has no visibility into how much of its capacity is going toward work that hasn’t been sold yet.
Waiting on the next job is a scheduling problem, not a machine problem, and spindle time blends the two together. A machine sitting idle because there’s no work queued for it at all is a demand or a sales problem. A machine sitting idle because the next job’s material hasn’t arrived, or the fixture from the last job hasn’t been swapped, is an operations problem. Both produce the exact same dip in spindle utilization, and the number alone can’t tell you which one you’re looking at.
The numbers worth watching alongside spindle time
None of this means spindle utilization is useless, it means it needs company. A few numbers that round it out into something a job shop can actually act on:
Utilization broken into categories, instead of a single blended percentage. Splitting scheduled time into cutting, changeover, and idle, at minimum, turns “utilization is 58%” into “58% cutting, 18% changeover, 24% idle,” which immediately tells you whether the gap is a setup problem or a scheduling problem, two completely different fixes.
Changeover cost by product pairing. Since setup is unavoidable in a job shop, the useful question isn’t how to eliminate it, it’s which specific transitions are costing the most, covered in depth in changeover time you can actually cut. A shop that knows its worst offending product pair can fix that specific transition instead of chasing a vague “setup takes too long” complaint.
OEE on the machine’s actual scheduled jobs, not spindle time alone, folds in whether the parts made during that cutting time were good and at the right speed, which spindle utilization by itself says nothing about. A machine cutting metal 70% of the time while running slow and scrapping parts looks fine on utilization and is actually losing real capacity.
Quoted-but-unbilled hours, tracked separately. Even a rough tally of hours spent on samples and first articles gives a shop a real number for how much capacity goes toward winning future work, instead of that time disappearing invisibly into the idle bucket or getting absorbed into whatever job happened to be open on the schedule at the time.
The mistake worth avoiding
The failure mode isn’t tracking spindle utilization, it’s using a single-part shop’s 85% target as the bar for a high-mix job shop, then treating every point below that as evidence of a problem. A job shop with a healthy mix of setup, quoting, and genuine changeover complexity might have a ceiling closer to 65% or 70% purely by the nature of the work, no scheduling fix required to close that gap, because the gap was never waste in the first place.
A worked comparison, two shops at the same utilization number
Two shops both report 65% spindle utilization on their busiest CNC. On paper, they look identical. In practice, they’re not even close to the same problem.
Shop A splits its 65% and finds 20% legitimate changeover across eight different jobs a week, 10% quoting and first article work tied to real quotes in the pipeline, and 5% idle waiting on a specific material delivery that’s been chronically late from one supplier. Nothing here points at the machine or the operator. The fix is a supplier conversation and possibly pre-staging material earlier, not a faster process.
Shop B splits the same 65% and finds only 5% changeover, because it mostly runs long batches of a few repeat parts, alongside 30% idle time with no job queued at all. That’s a demand or scheduling problem hiding behind a utilization number that, on its own, looked exactly like Shop A’s. Two shops, the same headline percentage, two completely different root causes, and a completely different fix each. Only the category breakdown, not the single blended number, tells them apart.
The number worth chasing isn’t a borrowed industry target. It’s your own utilization, broken into categories, tracked over time, with the idle and changeover slices shrinking where they can, and left alone where they’re the real cost of running high-mix work well.
Quick recap
- Spindle utilization answers one question: was the machine cutting when scheduled, nothing about quality, speed, or why it wasn’t cutting
- A job shop’s setup, quoting, and material wait time are real, often unavoidable, and none of it shows up as spindle time
- A single blended utilization number can’t tell scheduling problems apart from legitimate setup and non-billable work
- Split utilization into categories, cutting, changeover, idle, at minimum, before deciding where the real gap is
- Watch OEE alongside utilization, a machine that’s cutting often can still be losing capacity to slow cycles or scrap
- Don’t chase a single-part shop’s 85% target on high-mix work, the real ceiling is often lower and that’s not a failure