The bit is rarely the expensive part
Run the numbers this page opens with. A 38 dollar cutter with a measured life of 900 linear feet costs about 4.2 cents a foot. A part with 96 inches of toolpath cut half an inch deep in two passes is 16 feet of travel, so just under 68 cents of cutter goes into every part. Over 240 parts that is 3,840 feet, which is 4.267 cutter lives, so you keep five on the shelf and change four times mid-run.
Four changes at six minutes is 24 minutes of stopped machine. At 65 dollars an hour that is 26 dollars, against roughly 162 dollars of cutter actually consumed. On this job the cutters win. Push the life down to 300 feet and the picture flips hard: twelve changes, over an hour of downtime, and a strong argument for a better cutter at twice the price.
Depth passes are the multiplier nobody counts
The cut length per part is a drawing property. The linear feet the cutter travels is not — it is the cut length times the pass count, and the pass count is set by how deep you dare go. Taking half an inch in one pass instead of two halves the tooling cost per part outright. It also roughly doubles the load on the cutter, and whether the same cutter survives that trade at the same life figure is precisely the thing you have to test rather than assume. What the arithmetic can tell you is the size of the prize: if two passes at 900 feet of life and one pass at 500 feet of life are both real for your setup, the single pass is cheaper per part despite the shorter life, because it only travels half as far.
Amortised cost and cash outlay are different numbers
This page prints both on purpose. The amortised figure is the fraction of a cutter career this run actually consumed, and it is the right number for quoting a job. The cash outlay is the number of whole cutters you have to own to get through it without stopping, and it is the right number for the purchase order. On a short run they diverge a lot: a run that uses 0.4 of a cutter still needs one cutter bought.
Resharpening changes the arithmetic more than the price suggests
A resharpen that costs a third of a new cutter and gives most of a new life is a large lever on cost per part, because it multiplies the denominator of the whole calculation. Two resharpens turn 900 feet into 2,700 for well under three times the money. What it does not change is the number of times the machine stops, since a resharpened cutter still has to be swapped, and on a job where downtime dominates that is the term that matters. It also does not survive a diameter change: a reground cutter is smaller than it was, and whether that matters depends on whether your programs are cutting to size or to a tool-diameter offset.
Where this sits
The same pass count drives the run time on the CNC router job time calculator, and the same change minutes appear there as tool changes. Finishing passes with a ball nose add a lot of linear feet for very little depth — the scallop and stepover calculator shows how many. Spindle speed for a given cutter is on the router bit speed calculator and the chipload side on the machining speeds and feeds calculator. If the comparison is against a different process entirely, the cost per part crossover calculator puts them side by side.
Questions people ask
How do I measure tool life in linear feet?
Mark one cutter so you can tell it apart, and keep a tally sheet on the machine. Every time it runs a sheet, add the cut length of that sheet in feet, which your CAM already reports. Run it until the edge quality reaches whatever point makes you change it — fuzz on plywood, burning in hardwood, a chipped corner, a change in the sound. The running total at that moment is the life figure. Do it three times before you trust the average, because the spread between cutters is real. If different jobs use different depths of cut, log the total travel rather than the sheet count, since that is what the cutter actually experiences.
Should the change downtime be at the shop rate or the labour rate?
It depends on whether the machine has anything else to do. If the operator is standing there and the machine is the bottleneck, the full shop rate is right, because that stopped minute is a minute of capacity you cannot sell. If the operator has other work and the machine is not the constraint, only the labour part is real and the rest is an accounting fiction. Most small shops are in the first case more often than they admit, which is why change downtime so often turns out to be the larger of the two cost lines on this page.
Why does the parts-per-cutter figure round down?
Because a part that is three quarters cut when the cutter is done is not three quarters of a part, it is a scrapped blank and a cutter change in the middle of a program. The rounded-down figure is the number you can plan on. The fractional remainder is not lost, though: the cutter lives consumed line above uses the exact figure, so the amortised cost does account for the tail end of every cutter career even where it cannot be turned into a whole part.
Does this work for endmills in metal?
The structure does, but linear feet is a poor wear proxy for metal cutting, where life is normally tracked in cutting time at a given surface speed rather than in distance. If you want to use this page for milling, convert your life figure into feet at the feed you actually run and be aware that changing the feed changes both the distance and the life in ways that do not cancel. Coolant, coating and the specific alloy move the answer more than anything on this page, and none of them are inputs here.
My cutters die faster than the linear feet suggest they should. What else is going on?
The usual suspects are heat and corners. A cutter that is not clearing chips recuts them, and recut chips carry heat back into the edge, which is what actually kills carbide. Insufficient extraction, too small a chipload, or a slot cut at full width with nowhere for the chips to go all produce this. The other one is engagement: an interior corner momentarily wraps the cutter far past the nominal width of cut, and a program full of small pockets spends a surprising share of its distance in that condition. Neither shows up in a linear-feet figure measured on straight cuts, which is why the number to use is the one measured on your actual parts.