The CNC Machine Shop Traveler: Setup, Run, Deburr, Inspect

The Job That Made Money on Paper and Lost Money on the Floor

A quarter closes and the numbers don't add up. The 5-axis job you quoted at 40 hours actually ran 58. Nobody caught it in real time because nobody was tracking time by operation — just a start date stamped on a router and an end date scrawled next to it three weeks later. The setup that ran long because the fixture needed rework, the second op that sat waiting on a tool change, the deburr pass that took twice as long as estimated because the finish spec was tighter than the quote assumed — all of that detail evaporated the moment the part shipped. All that's left is a single number: 58 hours instead of 40, and no way to say why.

This is the ordinary failure mode of a CNC machine shop traveler that only carries a job number and a due date. It moves the part through the shop, but it doesn't carry enough structure to tell you where the hours went. A traveler built to actually cost the job — not just route it — has to break the work into operations that can each be measured against a standard: setup, run, deburr, inspect. This article works through that structure: what each operation needs to carry on the traveler, how setup and run time get tracked separately, where deburr and inspection get shortchanged on paper systems, and how the resulting data rolls up into an honest actual-vs-quoted number per job.

What a CNC Traveler Actually Has to Carry

A traveler for mill or lathe work is the order-specific document — paper or digital — that follows the part number through the shop and carries its routing, its standard times, and the record of what actually happened at each step. At minimum, a working machine shop traveler needs:

• Job and part identification — job number, part number, revision, quantity, due date, customer.
• The operation sequence — the ordered list of steps the part passes through (saw, mill op 1, mill op 2, deburr, inspect, ship), each tied to a specific work center or machine group.
• A standard time per operation — the quoted or estimated setup and run time the job was priced against.
• A space to log the actual — clock-in/clock-out or logged hours per operation, captured at the machine or kiosk, not reconstructed from memory at week's end.
• A place to record scrap or rework, tied to the operation that caused it — not a general note at the bottom of the router.

The machine shop job traveler template covers the full field-by-field layout; this piece focuses on the operation structure itself — specifically the four stages that show up on almost every CNC job and that get flattened into one lump number when the traveler isn't built to hold them apart.

Setup vs. Run: Why the Split Matters

Setup and run are different costs with different behavior, and a traveler that logs them as one number destroys the information a shop needs to quote the next job accurately.

Setup time is the fixed cost of getting the machine ready — fixturing, offsets, first-piece approval, tool changes. It happens once per job (or once per operation, on jobs that split across machines) regardless of quantity. Run time is the per-piece cost — the time the spindle is actually cutting, multiplied by however many pieces are in the lot.

A traveler that separates these two lets a shop see two different problems. If setup consistently runs long against standard, that's a fixturing or process-planning issue — the same one every time this part number comes back. If run time runs long, that's a feed/speed or tooling issue on the cut itself. Collapse both into a single "operation time" and both signals disappear into the same number, and the next quote for that part number gets built on the same bad assumption.

This is also where lot size matters. A 40-hour job that's mostly setup on a quantity of two behaves nothing like a 40-hour job that's mostly run time on a quantity of two hundred — but a traveler that doesn't split the two will quote both the same way next time. The setup vs. run time tracking breakdown goes deeper into how to structure the standard for each and how the split changes the quote for repeat work.

Deburr and Inspect: The Steps That Get Skipped on Paper

Deburr and inspection are the two operations most likely to disappear on a paper router, because they don't feel like "real" machine time — there's no spindle running, no obvious machine to log against. In practice they're where a meaningful share of the labor and a disproportionate share of the rework lives.

A traveler that treats deburr and inspect as first-class operations — with their own standard time, their own work center or station, and their own actual-vs-quoted comparison — catches two things a lumped traveler misses:

1. Deburr time creep. A finish spec that's tighter than the quote assumed, or a part geometry that traps burrs in hard-to-reach features, shows up as a deburr operation that consistently runs over standard. Without a dedicated line for it, that overage gets absorbed into whichever machining operation happens to be open when the traveler gets updated — usually the wrong one.
2. Where scrap actually originates. Scrap discovered at inspection was very likely caused two or three operations earlier. A traveler that logs the scrap event against the inspection step, rather than against the operation that actually produced the defect, makes root-cause tracing close to impossible. The traveler needs a field that lets the person logging the scrap point back at the causing operation, not just the operation where it was found.

Cost of poor quality is not a small line item industry-wide — the figure commonly cited across manufacturing is 15%–20% of sales, in a broad range of 5%–35% depending on the operation, and scrap and rework specifically can run up to 2.2% of annual revenue for the average manufacturer. Those are industry-wide figures, not a claim about any one shop's numbers, but they're a reasonable argument for why deburr and inspect deserve the same operation-level structure as the machining steps that precede them, rather than getting treated as a formality at the end of the router.

Building the Route: From Print to Standard Time

The routing is the template the traveler gets generated from — the ordered list of operations, work centers, and standard times for a given part number, built once and reused every time that part comes back. Getting the route right up front is most of the work; the traveler is just the vehicle that carries it through the shop and records what actually happened against it.

A workable route for a milled or turned part typically looks like: material prep/saw → primary mill or turn operation(s) → secondary operations (drill, tap, thread) → deburr → inspect → pack/ship. Each line gets a work center assignment and a standard time pulled from history on that part number, or estimated from a comparable one if it's new. The part routing template for machine shops walks through how to structure that sequence and where to source standard times when a part is running for the first time.

Rolling It Into a Job Cost

Once the operations are broken out — setup and run split, deburr and inspect standing as their own lines, scrap tagged to the operation that caused it — the traveler produces the raw material for an actual-vs-quoted comparison per operation and per job. That comparison is what tells a shop whether a job made money, not the invoice total.

The mechanics, illustrated with round numbers for a representative part: a mill operation quoted at 0.5 hours setup and 0.1 hours run per piece, on a lot of 20 pieces, has a standard time of 0.5 + (0.1 × 20) = 2.5 hours. If the logged actual comes in at 3.4 hours, that operation ran 0.9 hours over standard — and because setup and run were tracked separately, the shop can tell whether that overage sat in the setup (fixturing trouble) or the run (a slower-than-planned cut). Multiply the hour variance by the work center's burden rate — the fully loaded cost of running that machine and operator per hour, folding in labor, overhead, and machine cost — and the operation's dollar variance against quote falls out directly. Do that across every operation on the job and the job-level actual-vs-quoted number is just the sum. The CNC shop rate calculator covers how to build a defensible burden rate per work center, which is the other half of this math — the traveler supplies the hours, the rate supplies the dollars per hour.

Median annual wage for machinists and tool-and-die makers was $56,150 as of May 2024, according to BLS data — a useful anchor when sanity-checking a work center's labor component, though the fully loaded burden rate a shop actually costs against will run well past raw wage once overhead and machine cost are folded in.

Where This Breaks Down With Paper and Spreadsheets

A paper router with a single "hours" column, backed up by a spreadsheet somebody updates at week's end, can carry a job number and a due date just fine. What it can't carry is operation-level truth: which specific step ran long, whether the overage was setup or run, which operation actually caused a scrap part discovered three steps later. That reconstruction work — walking the floor, asking the operator who ran it, guessing at the split — is the same work a quarter-end margin surprise triggers, over and over, on the same handful of chronically underquoted operations. For the broader case for moving execution and costing off paper entirely, the execution and costing guide for small job shops lays out the full picture.

Getting a Working Traveler in Place

The structure above — job header, operation sequence with setup and run split, deburr and inspect as standing operations, scrap tagged to cause, standard time next to logged actual — is buildable in a spreadsheet or on paper. It's also the exact structure the Machining Shop Traveler System is built around, as a ready-to-use starting template for mill and lathe work rather than a blank page. Download it, adapt the operation list to the parts actually running through the shop, and the actual-vs-quoted math above becomes something the traveler produces automatically rather than something reconstructed after the fact.