← Back to Field Notes

March 21, 2025 · The Bloomfield Team

Why Your Machine Utilization Numbers Are Misleading

A shop manager reports 85% machine utilization. The number sounds healthy. The machines appear busy. Capacity appears tight. The conclusion drawn from that number, usually that the shop needs another machine, may be entirely wrong.

The problem is what "utilization" actually measures versus what people think it measures. In most shops, utilization means the machine was assigned to a job. It does not mean the spindle was turning. When you break down what happens during that 85% "utilized" time, the picture changes.

In the chart above, a machine reporting 85% utilization is actually cutting metal 48% of available hours. Setup consumes 18%. Tool changes, inspection, and various forms of waiting eat the rest. The machine is "utilized" for 85% of the day. It is productive for less than half.

The Three Metrics That Actually Matter

Spindle uptime measures the percentage of available hours where the spindle is actually rotating and removing material. This is the truest measure of productive capacity. World-class job shops run 55% to 65% spindle uptime on complex work. Most shops run 35% to 50% and do not realize it because they are looking at utilization instead.

OEE (Overall Equipment Effectiveness) multiplies three factors: availability (percentage of scheduled time the machine is available, excluding breakdowns and planned downtime), performance (actual throughput versus theoretical maximum), and quality (good parts as a percentage of total parts produced). OEE of 60% is average for job shop manufacturing. OEE of 85% is world-class. The gap between those two numbers on a single machine running $125 per hour fully burdened represents roughly $65,000 per year in recoverable capacity.

Revenue per spindle hour measures how much money each hour of cutting time generates. This metric exposes the difference between running a machine busy and running a machine profitably. A machine cutting $80 per hour jobs all day has lower revenue per spindle hour than a machine cutting $200 per hour jobs 60% of the day, even though the first machine has higher utilization.

Where the Hidden Capacity Lives

Most shops looking at 85% utilization believe they need to buy a machine to add capacity. The data usually tells a different story. The hidden capacity lives in three places.

Setup time reduction. If setup consumes 18% of available time, cutting that to 12% through standardized tooling, documented setup procedures, and offline preparation recovers 6 percentage points of capacity. On a machine running 4,000 hours per year, that is 240 hours of additional cutting time, equivalent to $30,000 in revenue capacity at $125 per hour.

Queue time elimination. Jobs waiting for a machine because the scheduler did not sequence work to minimize changeovers create artificial capacity constraints. Scheduling similar parts back-to-back on a machine, grouping by material or fixturing requirements, reduces the setup count and increases cutting time without any capital investment.

Unplanned downtime reduction. A basic preventive maintenance program that cuts unplanned downtime by 30% recovers 2 to 4 percentage points of capacity on most machines. That does not sound dramatic until you multiply it across 15 machines and a full year.

The Capital Expenditure Trap

Buying a new CNC machine costs $150,000 to $500,000 depending on type and capability. Installation, tooling, fixturing, and training add 20% to 40% on top. The total investment to add one machine of capacity is typically $200,000 to $700,000.

Recovering 10 percentage points of productive capacity across an existing 15-machine shop through setup reduction, scheduling improvement, and downtime reduction costs a fraction of that and produces results in weeks instead of the 6 to 12 months a new machine requires from order to productive operation.

The right time to buy a machine is when productive cutting time on existing equipment is above 60% and the shop has exhausted the operational improvements that recover hidden capacity. For most shops reporting 85% "utilization," that threshold has not been reached. The numbers are simply measuring the wrong thing.

How to Start Measuring What Matters

Modern CNC controllers from Haas, Mazak, DMG MORI, and others can report spindle-on time, feed rate overrides, and cycle counts through their built-in monitoring interfaces. The data is available. Most shops do not collect it.

Start with one machine. Track three numbers for 30 days: total available hours, hours where the spindle was actively cutting, and number of setups performed. Divide spindle cutting hours by available hours. That percentage is your real productive utilization on that machine.

If the number surprises you, it should. The gap between what you thought your machines were doing and what they are actually doing is where the next increment of growth lives. No capital required.

For a broader view of how production data connects to operational decisions, see our complete guide to production visibility.