Shop floor decision guide

How to Justify a New CNC Machine

A new machine is a capacity bet, not a tooling purchase. This guide shows how to justify a new CNC machine with cycle-time data, loaded hourly rates, and a payback test that survives an audit. Written for shop owners, manufacturing engineers, and buyers who have to defend the number.

Get Instant Quote Talk to an Engineer

Cycle-time basisLoaded rate methodPayback test

On this page

What the numbers have to prove

A machine is not a cost savingIt buys capability or capacity. If the shop is not constrained, the payback will not appear.

Build the case on cycle timeMinutes saved per part times annual volume. Nothing else moves the number as fast.

Use loaded rates, not machine ratesAdd labor, floor space, power, tooling, and depreciation to get an honest hourly figure.

Test 3 payback windows12, 24, and 36 months. A case that only works at 36 months needs a strategic reason.

Name the triggerBottleneck, tolerance, or lead time. A case with no trigger is a wish list.

Scoping the decision

When a new machine is the right answer

Most requests for capital start as a feeling. The shop is busy, quotes are slipping, operators are working overtime. Before any spreadsheet opens, name the constraint. A machine tool fixes exactly one thing: throughput on a specific operation. If the bottleneck is deburring, inspection, or programming, a new spindle will not help.

Ask three questions. Which part number or family is late? Which operation on that part is late? How many hours per week is that operation short? If the answer is a single 5-axis operation that runs 40 hours a week and needs 60, you have a machine case. If the answer is "everything is busy," you have a staffing or scheduling case.

Tolerance is the second trigger. A part that needs a true position of 0.020 mm across four faces will not run reliably on an old 3-axis with three setups. The scrap rate alone can pay for a 5-axis. Count the rework hours and the concessions before you count the cycle time.

1
Capacity triggerOne operation runs more than 80% of available hours for 8+ weeks.
2
Tolerance triggerScrap or rework from stacked setups exceeds 3% of part cost.
3
Lead-time triggerQuoted lead time is losing orders you could otherwise win.
4
Capability triggerA feature cannot be made at all without a 4th or 5th axis.

Cost model

Build the loaded hourly rate

Machine rate is the number that kills capital requests. A machine that costs USD 120,000 over 5 years and runs 4,000 hours a year carries about USD 6/hour in depreciation. That is not the cost of running it. Add the operator, the floor space, the power, the coolant, the tooling, and the maintenance contract.

For a single-spindle mill with one operator, the loaded rate typically lands between USD 45 and USD 75 per hour in a Western shop, and lower in a high-volume contract shop. Use your own payroll and utility numbers. The point is to compare like with like: the new machine's loaded rate against the old machine's loaded rate, not against its purchase price.

Tooling changes the math more than people expect. A 5-axis with a 40-taper spindle and a good holder set can cut titanium with a 12 mm end mill at 0.5 mm radial engagement. An older 3-axis with a worn spindle may need 6 passes and a second setup. That is where the real cost difference lives.

1
DepreciationPurchase price minus salvage, divided by expected machine hours over 5-7 years.
2
LaborFully burdened operator cost divided by attended hours. Include overtime.
3
OverheadFloor space, power, compressed air, coolant, and disposal.
4
Tooling and maintenancePer-part insert and cutter cost plus annual service spend.

The payback test

From cycle time to payback

Now the arithmetic. Take the seconds saved per part, multiply by the annual volume, and divide by 3,600 to get hours saved. Multiply by the loaded rate to get annual savings. Divide the machine cost by annual savings to get payback in years. If the result is under 24 months, the case is strong. Between 24 and 36 months, it needs a second reason.

Example. A part runs 18 minutes on an old 3-axis with 3 setups. On a 5-axis it runs 11 minutes in one setup. That is 7 minutes saved, or 0.117 hours. At 6,000 parts a year, that is 700 hours. At a loaded rate of USD 60/hour, that is USD 42,000 a year. A USD 90,000 machine pays back in 2.1 years before tooling and installation.

Do not stop at the first part. Add the second and third part families that will move to the new machine. A machine that only runs one part is a single point of failure. A machine that runs three families with similar setups has a much shorter real payback and a better story for the finance team.

1
Seconds saved per partMeasure with a stopwatch, not an estimate. Include load and unload.
2
Annual volumeUse committed orders plus a realistic forecast. Do not use best-case.
3
Second and third partsList every family that can move. Group by fixture and material.
4
Scrap reductionFewer setups means fewer datum shifts. Count the rework hours saved.

Risks

What makes a machine case fail

The first failure mode is single-part dependency. If the payback only works because one customer keeps ordering, the case is fragile. Diversify the part list or shorten the payback window. A machine that runs three families with similar fixtures is a better bet than one that runs a single high-volume part.

The second is underestimating the ramp. Operators need time on a new control. Programmers need time on a new post. Fixtures need to be designed and proven. Budget 4 to 8 weeks of reduced output before the machine runs at full rate. If the payback assumes full rate from day one, the number is wrong.

The third is ignoring the old machine. If it stays, you carry two depreciation lines and two maintenance contracts. If it goes, you lose the backup capacity. Decide before you sign. Most shops keep the old machine for 6 to 12 months, then sell it or move it to a lower-tolerance cell.

1
Single-customer riskIf one customer is over 50% of the volume, shorten the payback window.
2
Ramp timeBudget 4-8 weeks at reduced output for training and fixture prove-out.
3
Old machine planDecide keep, sell, or repurpose before the new machine arrives.
4
Tooling budgetFirst tool package can be 5-10% of machine price. Include it.

Run the process

Step by step: how to justify a new CNC machine

Six steps from floor data to a signed capital request

1
1. Time the current operationStand at the machine for one full shift. Record load, cut, unload, and inspection time for 10 consecutive parts. Write down the average and the spread. A spread over 20% means the process is not stable, and a new machine will not fix that.
2
2. Collect the scrap and rework hoursPull 3 months of nonconformance records for the part family. Add the hours spent reworking and the cost of scrapped material. If rework is over 2% of part cost, include it in the savings column.
3
3. Estimate the new cycle timeGet a real quote with a cycle-time estimate from the machine builder or an experienced shop. Compare setup count, tool count, and fixture count. A 5-axis should cut setup count by at least half on a multi-face part.
4
4. Build the loaded rate for both machinesUse the same categories for old and new. Include labor, floor space, power, tooling, and depreciation. If the new machine needs a different operator skill level, use that labor rate.
5
5. Run the payback at 12, 24, and 36 monthsCalculate savings for each window. If 24 months is not reached, list the strategic reasons: new capability, new market, or replacing a machine that fails regularly. Quantify downtime hours per year.
6
6. Write a one-page summaryOne page, five numbers: current cycle time, new cycle time, annual hours saved, loaded rate, payback. Add the trigger and the risk. Finance teams approve one page. They do not approve twenty.

Decision table

Which machine type fits which trigger

Match the constraint to the machine, not the other way around

Trigger	Best fit	Typical payback	Watch out for
One operation over 80% load	3-axis or 4-axis mill	12-24 months	Fixture cost and tooling package
Multi-face part, 3+ setups	5-axis simultaneous	18-30 months	Post-processor and programmer training
Turned part with cross features	Mill-turn center	24-36 months	Bar feeder and chip management
Tight tolerance, high scrap	5-axis with in-process probing	12-24 months	Metrology and calibration time
Low volume, many variants	3-axis with quick-change vises	12-24 months	Setup discipline, not the machine
Hard material, long cycle	High-torque 5-axis	24-36 months	Spindle and tool holder cost

FAQs

Questions engineers ask before signing

How many parts a year do I need to justify a 5-axis?

It depends on the cycle-time gap and the setup count, not the part count alone. A part with 3 setups that drops to 1 setup can justify a 5-axis at 2,000 parts a year if the loaded rate is high enough.

Run the payback at 12, 24, and 36 months with your own volume. If it only works at 36 months, the case needs a strategic reason such as a new capability or a new market.

Should I use machine rate or loaded rate in the payback?

Use loaded rate. Machine rate ignores labor, floor space, power, and tooling. It makes every machine look cheap and every payback look short.

Use the same categories for the old machine and the new machine. If the new machine needs a different operator skill level, use that labor rate in both columns.

What cycle-time saving is realistic on a 5-axis?

On a multi-face part with 3 or more setups, cutting setup count in half is common. That often saves 30-50% of total floor-to-floor time, depending on how much of the cycle is actual cutting.

On a simple 2-setup part, the saving may be under 15%. If most of the cycle is cutting time and the part is already stable, a 5-axis may not pay back.

How do I account for scrap reduction?

Count the rework hours and the scrapped material cost from 3 months of records. Add both to the annual savings column. Fewer setups means fewer datum shifts, which is where most tolerance drift comes from.

If scrap is under 1% of part cost, leave it out. It will not move the payback and it weakens the case.

What ramp time should I budget after installation?

Budget 4 to 8 weeks of reduced output. Operators need time on the control, programmers need time on the post, and fixtures need prove-out. Plan for 50-70% of target output in the first month.

If the payback assumes full rate from day one, the real payback will be 2-4 months longer. Build that into the request.

Can I justify a machine on lead time alone?

Sometimes, but only if you can name the orders you lost. Count the quotes that went to a competitor because your lead time was 4 weeks longer. Multiply by the margin.

Lead time alone is a weak case unless the lost revenue is documented. Pair it with a cycle-time or capability argument for a stronger request.

Need a second opinion on the numbers?

Send us the part and the current cycle time. Our engineers will review the setup count, the tolerance stack, and the realistic cycle time on a 3-axis, 4-axis, or 5-axis machine. You get a quotation and a free DFM analysis within 12 hours.

12-hour quote100% inspectionNo minimum order quantity

Get Instant Quote Talk to an Engineer

Follow the shop