Profits of CNC Processing Plants: Where the Margin Actually Comes From

• 1
  Why CNC processing plants do not make money the way people assume
• 2
  Where cycle time hides money in CNC processing plants
• 3
  Scrap rate, rework and inspection cost per part
• 4
  Economies of scale and where they stop working
• 5
  Energy, maintenance and people: the costs that scale with the plant
• 6
  What a stable quote looks like from the outside
• 7
  Five cost drivers and how they behave at different batch sizes

A common assumption is that a CNC processing plant earns its money on the machine hour rate. That is only half of it. The rate covers the machine, the operator, the floor space and the depreciation on the spindle. What decides whether the job is profitable is how many of those hours the part actually needs, and how many hours were spent before the first good part came off.

Take a simple aluminium bracket. If the fixture takes four hours to build and the cycle runs at 12 minutes, the shop needs roughly 20 parts just to pay back the setup. At one prototype, that setup is the whole job. At 500 parts, it is 0.5 minutes of setup per part and the margin comes from the cycle instead. Same drawing, two completely different businesses.

This is why two shops can quote the same part 40% apart and both be honest. One is amortising setup across a batch that suits its fixtures. The other is quoting a one-off with a fixture built from scratch, on a machine that could have been running a repeat order. Neither number is wrong. They describe different production systems.

The practical takeaway for a buyer is simple. Ask what the quote assumes about batch size, fixture reuse and inspection method. If the shop cannot answer, the price is a guess and it will move later.

• 1
  Rate is not marginMachine hour rate covers cost. Setup, scrap and inspection decide profit.
• 2
  Batch size changes everythingThe same fixture cost spread over 1 part or 500 parts.
• 3
  Ask what the quote assumesBatch size, fixture reuse and inspection method.

Cycle time is the most visible cost and the easiest to misread. A 40 minute cycle on a three-axis machine and a 12 minute cycle on a five-axis machine can cost the same per part if the five-axis rate is roughly three times higher. What matters is spindle hours multiplied by the rate, plus the number of setups the part needs before it is done.

Every additional setup adds three things: fixturing time, a new datum to hold, and a new chance to scrap the part. A part that runs in two setups instead of five saves hours across a 200 piece order and removes three tolerance stacks. On tight features, like a bore and a face that must stay square to each other, fewer setups usually means better yield as well as lower cost.

Setup count is also where five-axis pays for itself. With 16 simultaneous 5-axis machining centers, complex geometry that would need four or five operations on three-axis machines can often be cut in one or two. The saving is not the spindle speed. It is the four fixtures that never get built.

Cycle time reduction has a limit, though. Push the feed too hard on a thin wall and the part moves. Push the finish too fine and you add polishing hours downstream. The profitable shop finds the cycle that holds the tolerance with the least secondary work, not the fastest cycle on the stopwatch.

• 1
  Count setups, not just minutesEach setup adds fixturing, a datum and scrap risk.
• 2
  Five-axis pays in fixturesOne operation replaces four fixtures that are never built.
• 3
  Faster is not always cheaperA cycle that needs polishing downstream costs more.

Scrap is expensive in a way the quote rarely shows. A scrapped part costs the material, the machine hours already spent, the setup time already spent and the capacity that could have run something else. On a part with 90 minutes of machining, one scrapped piece can wipe out the margin on several good ones.

The countermeasure is process control, not more inspection at the end. Checking raw material before it reaches the machine, monitoring the first article and watching critical dimensions in process catches a drift while the part is still fixable. Final inspection then confirms what the process already did. On our floor that sequence runs on every order, with reports available on request.

Inspection cost scales with what you can automate. A caliper check on one dimension is cheap. A full layout on a CMM is not. For features held to ±0.005 mm, the inspection plan matters as much as the machining plan. Parts that need 100% CMM verification should be quoted with that labour included, not discovered later.

A qualification rate of 99.99% is not a slogan. It is the result of catching drift early, so a 10,000 piece run does not carry a hidden rework bill inside the unit price.

• 1
  Scrap costs capacityLost machine hours cannot be recovered on another job.
• 2
  Control beats inspectionCatch drift in process, confirm at final inspection.
• 3
  Match inspection to tolerance±0.005 mm features need a planned measurement method.

Volume reduces unit cost, but not evenly. The first drop is steep because setup and programming are spread wider. The second drop is smaller. By the time a shop is running 10,000 parts, most of the saving has already happened and further gains come from tooling, material buying and automation rather than from the machining itself.

High mix works against this. Ten different parts at 200 pieces each carry ten setups, ten programs and ten inspection plans. The shop is efficient at the machine but spends its engineering hours on changeovers. That is a real cost, and it shows up as a longer quote turnaround rather than a higher unit price.

The mix that suits a plant depends on its equipment spread. Shops with many three-axis machines do well on simple parts in large batches. Shops built around mill-turn and five-axis centers handle complex, low-volume work without rebuilding fixtures every time. Neither layout is better. They are tuned for different order books.

For a buyer, the useful question is not who is cheapest at 1,000 pieces. It is who stays stable from one piece to 10,000. A plant with no minimum order quantity and a 3–5 day shipping window on standard work is usually set up to absorb that range.

• 1
  Savings flatten outMost unit cost falls before 1,000 pieces.
• 2
  High mix costs engineeringChangeovers consume skilled hours, not machine hours.
• 3
  Match shop to batchThree-axis for simple volume, five-axis for complex low volume.

Machine hours carry hidden overhead. Spindles, chillers, compressors and coolant systems draw power whether or not the part is complex. A shop running older machines at partial load pays for that in the unit rate. Newer equipment with better duty cycles cuts the energy per part, but the saving only appears if utilisation is high.

Maintenance is the other slow cost. Preventive schedules on spindles, way covers and tool holders keep tolerance stable. Skip them and the machine drifts, the scrap rate climbs and a rebuild lands in one quarter instead of being spread across the year. A plant that tracks tool life and spindle hours usually quotes tighter than one that reacts to failures.

People cost is where training shows up. A programmer who knows the material and the fixture design writes a process that runs the first time. A shop that treats programming as clerical work pays for it in prove-outs and trial cuts. Skilled staff are expensive per hour and cheap per part.

Our floor runs 150 technicians across three wholly-owned plants and 7,600 m² of production space. That is enough depth to keep programming, machining, finishing and inspection under one roof, which removes the coordination cost of moving parts between vendors.

• 1
  Utilisation decides energy costThe same machine costs more per part when half loaded.
• 2
  Preventive work protects toleranceSpindle and fixture maintenance keeps yield stable.
• 3
  Skilled programming cuts prove-outsFewer trial cuts means fewer charged hours.

A quote that will hold usually comes with a DFM note. It points at a wall thickness, a corner radius or a tolerance that drives cost, and it offers a change that removes the cost without losing function. If the note says a 0.4 mm radius needs a 0.4 mm cutter and a longer cycle, that is a shop showing its arithmetic.

Lead time tells you something about the order book. A plant that can return a quotation and a free DFM analysis within 12 hours and start production within 24 hours is not short of capacity. It has the process planning already done. Beware of a very low price paired with a long, vague lead time. That combination usually means the setup has not been thought through.

Certification is a filter, not a differentiator. ISO 9001:2015, IATF 16949:2016, ISO 13485:2016 and ISO 27001:2022 each imply a documented system. For medical or automotive work those systems are the entry ticket. They do not tell you whether the shop quotes accurately, but they do tell you whether scrap and complaints get tracked.

The most useful signal is repeatability across a mixed order book, from a single prototype to a 10,000 piece run, without the unit price collapsing or the delivery date moving.

• 1
  DFM note means real costingThe shop can name what drives the price.
• 2
  Fast, specific lead timeVague dates usually hide an unplanned setup.
• 3
  Certificates are a floorThey show a system exists, not how well it runs.