How to Calculate Cost Per Label

The invoice that didn't match the quote

An estimator at a six-press shop pulls up a job that shipped last week — 8,000 labels, four colours, a die that's been in the cabinet for two years — and the actual cost per label is almost 30% higher than what went out on the quote. Nothing on the job was unusual. The press ran at a normal speed, the waste was normal, the customer didn't change anything mid-run. The difference was buried in the spreadsheet: the plate cost cell was still referencing a job from eighteen months ago, and nobody had touched it since. The quote wasn't wrong because the estimator guessed badly. It was wrong because "cost per label" isn't one number you type in — it's five or six numbers added together, and if one of them is stale, the total is stale too, and nothing on the surface tells you which one broke.

By the end of this walkthrough you'll be able to build a cost-per-label figure from its actual components, by hand, and know exactly which input to check first when a number looks off.

What "cost per label" actually adds up in the box

Every per-label number on a flexo quote is the sum of five things divided across a run length:

• Material — substrate cost, priced by area, not by "per label"
• Ink — coverage-dependent, usually a smaller line item than people assume on simple jobs
• Plate cost — a fixed tooling cost, amortised (spread) across however many labels the plate will print
• Die cost — the same amortisation logic as plate, on its own schedule since dies last many runs
• Press time and makeready waste — the hourly rate times the hours the job actually occupies the press, including the labels burned getting the press into register before good product starts

Cost per thousand (CPT) is the more common unit converters actually quote in, and cost per label is just that number divided by 1,000. If you want the mechanics of building CPT specifically — including how markup and margin sit on top of it — that's covered in how to calculate cost per thousand for labels. This article stays one level down: the raw inputs that feed either number.

How to calculate cost per label from material cost by area

Material is priced by area because a die-cut label isn't sold by the piece from the supplier — the supplier sells rolls of substrate, and the converter's yield off that roll depends on label size, gap, and how many labels sit across the web. The U.S. label industry's standard unit for this is MSI — thousand square inches. A substrate has a cost per MSI; a label has an area in square inches; multiply and you get material cost per label before waste.

The formula in its simplest form:

(label width × label height ÷ 1,000) × cost per MSI = material cost per label

A 2" × 3" label is 6 square inches. At a hypothetical $0.02/MSI substrate cost, that's (6 ÷ 1,000) × $0.02 = $0.00012 per label in raw material — before matrix waste, before gap between labels, before roll-change waste. Those additions matter more than the base number, which is part of why a flat "$X per label for material" rule of thumb breaks down as soon as label geometry changes. The full mechanics — yield, gap, matrix waste percentage — are worked through in cost per MSI for labels explained.

Plate and die amortisation: spreading tooling cost across the run

Plates and dies are fixed costs. A plate set costs the same to produce whether the job runs 2,000 labels or 50,000 — so the per-label impact of that fixed cost falls as run length rises. This is the single biggest reason two shops quoting the identical job at different volumes can land on very different per-label numbers, and it's also the mechanical reason a flexo-to-digital crossover point exists at all: digital print carries no plate cost to amortise, so at low volumes the fixed flexo tooling cost dominates the comparison, and at high volumes it nearly disappears into the total. Where exactly that crossover falls depends on colour count, label geometry, and the shop's actual configured plate cost — it isn't a fixed number that transfers between shops.

Die cost amortises the same way but on its own schedule, since a die typically survives many jobs, not just one run. Some shops fully amortise a die on the first job that uses it; others spread it across an estimated lifetime run volume. Either approach is defensible — the mistake is not having a documented rule and instead re-deciding it, inconsistently, quote by quote.

Press time, makeready waste, and the hourly rate

This is the layer most estimators get closest to right and still lose the most margin on, because it has two moving parts that behave differently: run time (labels ÷ press speed for the colour count and substrate) and makeready waste (labels consumed getting the press into registration, colour, and pressure before saleable product starts).

Run speed isn't flat — a job with more colour stations or a more demanding substrate runs slower than a simple one-colour label, and that speed curve is exactly what determines press-time cost. Makeready waste is a separate number entirely: it's lost labels, charged at material-plus-ink cost, that never generate revenue but do consume press time. A shop that quotes makeready waste as a rough percentage guess instead of tracking it per job is guessing at one of the largest variable costs in the whole estimate.

Multiply press hours (run time plus makeready time) by the shop's loaded hourly rate — labor, overhead, and press cost rolled into one figure — and divide across the run to get the press-time-and-waste contribution to cost per label.

A worked example: how to calculate cost per label start to finish

Take a hypothetical 4-colour job, 5,000 labels, 2" × 3" size, on a press running at a hypothetical $120/hour loaded rate:

• Material: 6 sq in ÷ 1,000 × $0.02/MSI = $0.00012/label → $0.60 for 5,000
• Ink: roughly $0.30 for the run at this coverage (a smaller line item on a simple 4-colour job)
• Plate cost: $600 flat, amortised over 5,000 labels = $0.12/label → $600 for the run
• Die cost: $400 flat, amortised over an assumed 20,000-label die life, charged at the 5,000-label share = $100 for this run
• Press time and makeready: 1.5 hours total at $120/hour = $180

Total run cost: $0.60 + $0.30 + $600 + $100 + $180 = $880.90 for 5,000 labels, or $0.176 per label before markup.

Run the identical job at 20,000 labels instead of 5,000, and the plate cost drops to $0.03/label, the die-cost share drops proportionally, and press time per label falls because makeready is a fixed hit spread across four times the volume. The per-label number moves meaningfully — not because anything about the job changed, but because the fixed-cost layers amortise differently. This is the whole reason "what does a label cost" has no single answer without a run length attached to it.

For the full flexo build-up in one place — including where markup and overhead sit relative to these five layers — see how to estimate a flexo label job.

Where this math breaks down in a spreadsheet

The arithmetic above isn't hard. What's hard is doing it the same way, with current inputs, on the fortieth quote of the week — and a workbook has no way of telling you which cell went stale.

Every one of the five layers above lives in its own cell in most shops' Excel files: a plate-cost lookup, a die-amortisation assumption, a press-speed table, a makeready-waste guess. None of them are wrong in isolation. They go wrong quietly, one at a time, as jobs change and nobody goes back to update the reference cell — which is exactly what happened in the invoice discrepancy this article opened with. A calculation engine that treats each layer as a live input rather than a static cell is described in more detail in how a label pricing calculator actually works, and the full estimating process — from press-speed curve through final markup — is laid out start to finish in the complete guide to label estimating.

If you'd rather work through your own numbers on paper before trusting a tool with them, the Label Job Estimating Worksheet walks through each of these five layers with blank fields for your shop's real plate cost, die cost, hourly rate, and substrate pricing — the same structure used above, built for your actual jobs instead of a hypothetical one.