Label Printing Cost Per Thousand: How the Calculation Actually Works

The reorder that shouldn't have cost that much

A converter reprints a label they last ran eighteen months ago — same die, same four colors, same 3" x 2" size, quantity bumped from 5,000 to 25,000. The customer expects the per-thousand price to drop, and it does, but not by the amount anyone predicted off the top of their head. The estimator pulls up last year's job folder, finds a number scrawled in a margin, and has no idea which line items in that old quote were fixed costs that don't repeat and which ones scale with volume. Multiply that uncertainty across forty quotes a month and you understand why so much flexo pricing still runs on gut feel dressed up as a spreadsheet formula.

This piece breaks the label printing cost per thousand calculation into its actual parts — substrate, plate, makeready waste, die, and press time — so you can see exactly why the per-label number moves as much as it does between a short run and a long one, and rebuild that number for any job without guessing.

What "cost per thousand" actually bundles together

Cost per thousand (CPM) is not one number — it's a sum. Every flexo job carries a mix of costs that scale with quantity and costs that don't move no matter how many labels you print. Substrate and ink are variable: print twice as many labels, buy roughly twice as much material. Plates and dies are fixed: a plate set costs the same whether the run is 2,000 labels or 200,000. Makeready waste sits in between — it's a roughly fixed quantity of wasted material and time that gets spread over however many good labels come off the press.

The label printing cost per thousand calculation is really the exercise of adding those fixed and variable pieces together, then dividing by run length in thousands. Because the fixed pieces don't shrink, CPM always falls as quantity rises — the question an estimator actually has to answer is by how much, and whether that drop changes which press should run the job at all.

How substrate cost turns into an MSI-based number

In the US, converters price roll and sheet substrate by MSI — thousand square inches — rather than by the sheet or the roll. The formula is straightforward: multiply label width by height to get area in square inches, multiply by quantity, and divide by 1,000.

Worked example (round numbers, not a market rate): a 3" x 2" label has 6 square inches of area. At 25,000 labels, that's 25,000 × 6 ÷ 1,000 = 150 MSI. At an illustrative $28/MSI, that's $4,200 in substrate — before ink, plate, or press time enter the picture. Run the same math at 5,000 labels and substrate alone drops to $840 — but as the next section shows, that's not the whole substrate story, because waste doesn't scale down with the run.

For the full MSI mechanics — how converters land on a dollar-per-MSI figure and where substrate suppliers quote differently — see how cost per MSI works for labels.

How a per-colour plate cost changes the per-label math

Plates are priced per colour, and a plate set is a fixed cost: it doesn't care whether the run is 2,000 labels or 200,000. That's the mechanism behind the single most common surprise in flexo pricing — why the same job, same die, same colours, prices so differently at different quantities.

Worked example: four plates at $150 each is $600 in plate cost, full stop. Spread over a 5,000-label run, that's $600 ÷ 5 = $120 per thousand. Spread over 25,000, it's $600 ÷ 25 = $24 per thousand. Nothing about the plates changed — only how many labels absorbed the same fixed number changed. This is also the mechanical reason a flexo-versus-digital crossover point exists at all: as plate and die costs amortise across more units, flexo's per-label cost keeps falling in a way digital's per-impression or per-coverage cost does not. Where that crossover actually lands depends on colour count, label geometry, and your shop's configured plate cost — not a fixed number you can borrow from another converter.

Why makeready waste hits short runs hardest

Every flexo job burns some material and press time before the press is producing saleable labels — registering colour, adjusting tension, dialing in impression. That waste is roughly a fixed quantity per job, measured in feet or labels, largely independent of how long the run is afterward.

Worked example: say a job wastes 1,200 labels getting to colour, regardless of whether the total run is 5,000 or 25,000. On the 5,000-label run, that waste is 24% of total labels produced. On the 25,000-label run, it's under 5%. The substrate and ink spent on those wasted labels still has to be recovered — and because it's spread across far fewer good labels on the short run, it inflates the per-thousand number disproportionately there.

A short run doesn't just have less volume to spread fixed costs over — it also has the same waste as the long run, spread over less volume twice. That's why a small reorder can carry a bigger CPM swing than the plate line alone would suggest.

How die cost amortises into the per-thousand number

A rotary die is tooling, not consumable — bought once, then reused on every future run of that SKU as long as it's kept on file. That makes the first run of a new label the most expensive one to quote, and every reorder materially cheaper, because the die line drops to zero once it's already paid for.

Worked example: a $750 die on a brand-new job, spread over 5,000 labels, adds $150 per thousand. Spread over 25,000, it's $30 per thousand. On the reorder eighteen months later — same die already on the shelf — that line disappears from the quote entirely. This is exactly why a converter can quietly lose margin on reorders priced from memory instead of from the file: if the estimator doesn't know (or doesn't check) that the die cost already amortised out, the reorder gets priced as if it's still carrying that cost, and the customer overpays or the shop underprices the next new job to compensate.

Building the full CPM: a worked example

Putting the pieces from above together — substrate (including waste), plate, die, and a loaded press-time rate — here's how the same 3" x 2", four-colour label prices out at two quantities. These are illustrative, round inputs to demonstrate the method, not published rates:

5,000-label run:

• Substrate (5,000 good + 1,200 waste = 6,200 labels; 37.2 MSI at $28) ≈ $208/M
• Plate set ($600 fixed) ≈ $120/M
• Die ($750 fixed, new job) ≈ $150/M
• Press time (at an illustrative $105/hr loaded rate, ~0.31 hr) ≈ $7/M
• Approximate CPM: ~$485/M

25,000-label run:

• Substrate (25,000 good + 1,200 waste = 26,200 labels; 157.2 MSI at $28) ≈ $176/M
• Plate set ($600 fixed) ≈ $24/M
• Die ($750 fixed, new job) ≈ $30/M
• Press time (~1.31 hr at the same rate) ≈ $6/M
• Approximate CPM: ~$236/M

Same label, same die, same colours — roughly double the per-thousand cost at the shorter run, driven almost entirely by the fixed plate and die lines and by waste spread over fewer good labels. That's the label printing cost per thousand calculation doing exactly what it's supposed to do: telling you the truth about where fixed costs land.

Running this by hand in a spreadsheet works, but it's also exactly where a stale plate price, a wrong MSI rate, or a forgotten die amortisation quietly breaks a quote. FlexoCommand's flexo quoting engine builds this stack automatically — press-speed curve by colour count, per-colour plate cost, makeready waste, die amortisation, and MSI substrate pricing all roll into the loaded hourly rate on one screen, so the same job at two quantities shows its real CPM without anyone re-deriving the math from a margin note.

Why CPM alone doesn't tell you if flexo is the right machine

A correct flexo CPM only tells you what flexo costs — it doesn't tell you whether flexo is the right press for the job. A digital press prices a label completely differently: an LEP click charge is flat per impression regardless of ink coverage, while UV inkjet cost scales with how much ink actually lands on the label. Neither of those follows the plate-and-die amortisation curve above, which is exactly why a job that's cheap per thousand on flexo at 25,000 units can flip the other way at 2,000.

FlexoCommand's comparative quote runs the flexo CPM stack, the LEP click-charge model, and the UV inkjet coverage model side by side on the same job and flags where the crossover falls for that specific label — geometry, colours, and plate cost included, rather than a rule of thumb borrowed from a different shop's job mix.

If you want to run this calculation on your own jobs before touching software, start with the complete flexo label estimating guide and the step-by-step walkthrough for estimating a flexo job, or go straight to how to calculate cost per label for the per-unit version of this same math. Or skip the rebuild entirely — download the Label Job Estimating Worksheet and drop your own substrate, plate, die, and press-rate numbers into a template that already does the arithmetic above.