Increase CNC Profit Now: Where the Money Actually Goes
Most machine shops do not lose margin on the hourly rate. They lose it in setup hours, spindle idle time, scrap, and rework. This page explains the mechanics behind each leak and what a shop can realistically do about it. Read it if you buy machined parts or run a machining department and want to know which levers move cost and which ones just look busy.
In this article
- 1
- 2
- 3
- 4
- 5
- 6
- 7
How a machining hour turns into profit
A CNC hour has three parts: cutting time, non-cutting time, and dead time. Cutting time removes metal. Non-cutting time includes loading, probing, tool changes and repositioning. Dead time is waiting for a program, a fixture, an operator, or an inspection result. Only the first part is usually quoted at full rate, but all three consume the machine.
Profit per hour is therefore cutting time divided by total occupied time, multiplied by the margin on that hour. If a machine is occupied 100 hours a month but only cuts for 55 of them, the shop is carrying 45 hours of cost with no chip produced. Raising the cut share from 55% to 70% on the same machine adds roughly 27 cutting hours per month without buying anything.
This is why the fastest route to increase CNC profit now is not a new spindle. It is removing the work that surrounds the cut. Setup sheets that arrive late, fixtures that need shimming, and first-article inspection done on the machine all sit in the non-cutting or dead column.
The lever with the shortest payback is usually fixture and workholding design. A self-locating fixture that drops onto a receiver can cut 30 to 60 minutes of setup per job. On a 20-job month that is 10 to 20 machine hours recovered, which then can be sold again.
Setup time is the biggest single leak
Setup is expensive because it is paid at the machine rate but produces no parts. On a 3-axis vertical mill, a typical vise-and-stop setup for a new part runs 45 to 90 minutes: dial in the vise, touch off tools, find the datum, prove the program at reduced feed. If that job repeats monthly, the setup is repeated monthly.
Four things shorten setup without touching the cutting parameters. First, pre-set tooling off the machine so tool length offsets are already known. Second, standardize the datum across a family of parts so the program origin does not move. Third, use a receiver plate or zero-point system so the fixture repeats within a few microns. Fourth, keep a setup sheet with the actual offsets used last time.
The fifth item is scheduling. If a job runs on the machine that already holds the right fixture and the right tools, setup can drop to 10 or 15 minutes. Grouping parts by fixture family rather than by due date looks inefficient on a schedule board but usually raises throughput.
Watch the boundary here. On one-off prototypes, elaborate fixturing is a loss. For a single part, a vise and a dial indicator is the correct answer. Fixture investment only pays when the job repeats three times or more, or when the part is too flimsy to hold any other way.
Why 5-axis changes the cost curve on complex parts
A 3-axis machine can only reach the part from one direction at a time. Every new face means a new setup, a new datum, and a new chance to introduce position error. A 5-axis machine tilts the tool or the table, so the same part can be reached from many directions without re-clamping.
The gain is not in cutting speed. It is in eliminating the second and third setups, and in keeping one datum through the whole part. On a part with four angled faces and two bores that must be coaxial, that can remove two setups and the associated re-datum risk.
The trade is programming. Simultaneous 5-axis toolpaths need collision checking and a post-processor that the shop actually trusts. A poorly verified program can scrap a part faster than any 3-axis job, because a wrong tool vector moves in three directions at once.
So the rule is geometry-driven. If the part has features on three or more faces, deep pockets with undercuts, or bores at compound angles, 5-axis usually wins on total cost even at a higher hourly rate. If the part is a flat plate with holes, 3-axis is cheaper and that does not change.
Scrap and rework quietly eat the margin
Scrap is not just the material cost. It is the machine hour, the operator hour, the inspection hour, and the scheduling disruption. A part scrapped at the last operation can carry three times the cost of a part scrapped at the first.
The cheapest controls are also the least exciting. Verify the program in the air with the tool offset shifted up before the first cut. Probe the stock so the program adjusts to the actual blank instead of a nominal one. Check the first part fully, off the machine, before releasing the run.
For a shop aiming at ±0.005 mm, thermal drift matters. A machine that has been sitting cold will move during the first hour. Letting the spindle warm up, or running roughing first and finishing after warm-up, keeps the finishing pass inside tolerance.
In-process monitoring pays on long runs. Measuring a critical bore every 20 parts catches a worn tool before it produces a batch of undersized holes. The cost of the measuring time is far below the cost of scrapping 40 parts.
Quoting speed is a profit lever, not an admin task
A quote that takes four days to reach the buyer is often a quote that never gets used. Buyers work down a list and stop when they have enough answers. Speed wins work, and the shop that quotes in a few hours gets to choose which jobs it wants.
Fast quoting also improves job selection. A shop that can estimate cycle time and setup time quickly can decline the jobs that will lose money. Turning down the wrong job protects more margin than optimizing the right one by 5%.
The input side matters as much as the output. A DFM review that flags a 0.4 mm deep rib before the mold is cut, or a tolerance that requires a second operation, saves money before the first chip. That is why a drawing review should happen at quoting, not at production.
Confidentiality is part of the same transaction. Engineers will not send a full assembly drawing for review if the file handling is unclear. A named NDA and a secure upload path remove that hesitation and shorten the loop.
Typical time split on a 3-axis job versus a 5-axis job
Illustrative ranges for a mid-complexity aluminum bracket, one setup each.
| Activity | 3-axis, multiple setups | 5-axis, one setup |
|---|---|---|
| Setup and fixturing | 45–90 min per face | 20–40 min total |
| Tool changes | Higher, more re-clamping | Lower, fewer setups |
| Non-cutting share | Often 40–50% of occupied time | Typically 20–30% |
| Rework risk | Rises with each re-datum | Lower, one datum holds |
| Best fit | Simple parts, high volume | Complex geometry, tight position |
Which lever to pull first
Match the symptom on the left to the action on the right.
| Symptom | Likely cause | First action |
|---|---|---|
| Machine busy, output flat | Non-cutting share too high | Pre-set tools, standardize datums |
| Repeat job still takes long setup | No fixture family | Receiver plate or zero-point system |
| Scrap spikes on one feature | Tool wear or thermal drift | In-process check every 20 parts |
| Quotes take days to send | Manual estimating | Standard setup and cycle templates |
| Rework on multi-face parts | Multiple datums | Move to one-setup 5-axis where geometry allows |
| Margin falls on long runs | Cycle time not revisited | Re-cut speeds and feeds after 500 parts |
The trade, stated plainly
If your parts repeat, spend on fixturing and tool pre-setting first, because that is where the recovered hours are. If your parts are complex and low volume, spend on one-setup 5-axis capability instead, because setup elimination beats speed on that geometry. Doing both at once without measuring current setup and scrap hours is guesswork.
Questions engineers ask next
How do I know whether setup or cutting speed is the real bottleneck?
Log three numbers for two weeks: total occupied machine hours, actual cutting hours, and parts produced. If cutting hours are under about 60% of occupied hours, setup and idle time dominate and speed changes will barely show up in output.
If cutting hours are already above 75%, the remaining gain has to come from feeds, speeds, tool life or toolpath efficiency. Measuring first avoids spending money on the wrong side of the problem.
Does a higher shop rate always mean higher total part cost?
No. Hourly rate and total cost are different numbers. A shop at a higher rate that runs one setup, has low scrap, and ships on a stable schedule can land lower on total cost per good part than a cheaper shop that needs three setups.
The comparison should be cost per conforming part delivered, not rate per hour. Ask for cycle time, number of setups, and inspection plan, then multiply.
When is 5-axis not worth it?
When the part is prismatic with features on one or two faces, 3-axis is cheaper. The higher machine rate and longer programming time buy nothing if the geometry does not need the extra axes.
It is also a poor fit for very simple parts at very high volume, where a dedicated fixture on a 3-axis machine with a short cycle time will always win.
What tolerance can a shop realistically hold without a second operation?
On aluminum and stainless parts, ±0.005 mm is achievable on critical features with a warm machine, sharp tooling and in-process checks. General features on the same part are usually specified looser, around ±0.05 mm, because tightening everything raises cost without improving function.
The practical rule is to tighten only the features that locate or seal. Wide tolerances elsewhere keep the process stable and the price down.
How does inspection strategy affect cost?
Inspection time is real machine or operator time. Checking every dimension on every part is expensive and usually unnecessary. Checking the critical-to-function dimensions on the first part, then sampling during the run, catches drift at a fraction of the cost.
For regulated work, the inspection plan should be agreed before the run starts, so the reports match what the buyer actually needs.
Can a shop increase profit without buying any equipment?
Yes, and often that is the first move. Standardizing datums, pre-setting tools, writing down the offsets that worked, and grouping jobs by fixture family all raise the cutting share with no capital spend.
Equipment decisions get easier once those numbers are on the table, because the shop can see exactly which hours it is trying to recover.
Send a drawing, get a costed answer
We review the drawing, flag the features that drive cost, and return a quotation with a DFM analysis within 12 hours.
12-hour quoteDFM analysis included100% inspectionNDA on request
More machining notes
We publish setup notes, tooling trials and inspection data from the factory floor.
