Print on demand prints one copy when someone orders it, at a unit cost that never changes. Offset lithography prints a whole run at once, at a unit cost that falls as the run grows because the setup is amortised across it.
That is the real difference, and it is the only part of this comparison that is properly sourced. The break-even quantity everyone quotes is not.
The Number Nobody Publishes
You will be told offset becomes cheaper at 500 copies, or 1,000, or 2,500. I went looking for the source of those figures.
Neither Amazon KDP, IngramSpark, Lightning Source, Lulu nor Blurb publishes a crossover quantity. Neither does PRINTING United Alliance, whose economics reports are member-gated, nor the Book Manufacturers' Institute. No trade body sets a minimum order quantity for offset either; it is negotiated per printer and per job.
Every circulating figure traces to self-publishing blogs and printer marketing. So this page will not give you one, and you should be suspicious of any that does.
What you can do instead is compute your own, because the POD side of the equation is published exactly.
Work Out Your Own Crossover
Amazon publishes its printing cost as a formula: *"Fixed cost + (page count per page cost) = printing cost"**. On Amazon.com, black ink, regular trim:
| Product | Fixed | Per page |
|---|---|---|
| Paperback, 24 to 110 pages | $2.30 | none |
| Paperback, 110 to 828 pages | $1.00 | $0.012 |
| Standard colour, 72 to 600 pages | $1.00 | $0.0255 |
| Premium colour, 42 to 828 pages | $1.00 | $0.065 |
| Hardcover, 110 to 550 pages | $5.65 | $0.012 |
So a 300-page black ink paperback costs $1.00 + (300 × $0.012) = $4.60 per copy, at any quantity.
Then get an actual offset quote for the run length you are considering. The comparison is that simple, and it is specific to your book rather than to an averaged blog figure. IngramSpark publishes no equivalent formula, exposing only a calculator, so a like-for-like comparison against it is not possible from published sources.
The structural point is what matters. The POD formula contains no quantity term at all. Unit cost is flat forever. Offset's falls with run length. That is why a crossover exists; where it sits depends on your page count, trim, colour, and the quote you can negotiate.
What POD Genuinely Cannot Do
Here the sourcing is solid, and these constraints matter more than the arithmetic for some books.
No spot colour. IngramSpark's File Creation Guide is explicit:
"SPOT COLORS: Please convert all spot colors (PMS / PANTONE) with/without transparencies to CMYK, (even in supporting .eps files). These may produce unexpected color results when printing. Any dissatisfaction with color shift will be the publisher's responsibility to correct."
And separately: "Please do not include Spot colors or ICC profiles in your file as these can produce unexpected results during processing."
If your cover design depends on a specific Pantone, POD cannot reproduce it and the risk is explicitly yours.
A tighter ink limit. IngramSpark caps total ink coverage at 240 percent CMYK and recommends a rich black of 60 cyan, 40 magenta, 40 yellow, 100 black. Sheetfed offset commonly runs higher.
A closed, short paper list. IngramSpark offers four interior stocks. KDP offers cream, white and groundwood, with glossy or matte lamination on an 80-pound cover stock. Offset gives you whatever the printer can source. No platform states "offset offers more stocks", but the closed lists speak for themselves.
A closed binding list. Across all four POD platforms the published options are perfect bound, case laminate, Digital Cloth, jacketed, coil and saddle stitch. Sewn binding, foil stamping, blind deboss and emboss, head and tail bands, ribbon markers, edge staining and spot UV appear on none of them. That is an argument from four independent closed enumerations rather than a quoted denial, but it is a strong one.
The printer can override your paper. A clause worth knowing: "TITLES WITH HEAVY INK COVERAGE... In the event LS encounters this issue during printing, LS reserves the right to change the printing of the interior to a heavier weight paper and adjust the cover to compensate for the new spine width."
What POD Does That Offset Cannot
No inventory and no capital. The effective minimum order is one copy. Nothing is printed before it is sold, so there is no unsold stock and no storage.
No risk on a first book. This is the decisive point for most self publishers, and it does not appear in a unit cost comparison at all. A 1,000-copy offset run that does not sell is a loss plus a storage problem. POD's worst case is that nobody orders.
Corrections are cheap. A typo found after publication is a re-upload. In offset it is a reprint or 1,000 copies carrying the error.
Distribution is built in. IngramSpark reaches the trade catalogue and KDP reaches Amazon, without you shipping anything.
Choosing
POD for a first book, an unproven title, a backlist, anything where you cannot confidently forecast demand, and anything you want in bookshops without warehousing it.
Offset for a confirmed print run you can sell, a book whose physical production is part of its value (special bindings, foil, uncoated stock, precise colour), and art or photography books where reproduction quality is the product.
Both is common and sensible: an offset run for a launch or a special edition, POD for the long tail.
What Deckle Exports
Deckle's print PDF is built for POD's published requirements: bleed on three edges only, a gutter enforced against KDP's page-count bands, and fonts embedded.
Three things to change if you are going to offset:
- Ask the printer for their spec first. Offset printers have their own templates, bleed and margin requirements, and they vary far more than the POD platforms do.
- PDF/X. Deckle exports a standard PDF rather than PDF/X-1a, which most offset printers will want. Convert in a preflight tool.
- Colour. Deckle produces interiors, not covers, so the spot colour question belongs to your cover designer either way.
For POD's own requirements, see KDP print file requirements and IngramSpark file creation requirements.
Common Mistakes
- Trusting a quoted break-even figure. No qualifying source publishes one.
- Comparing unit cost alone. It ignores unsold stock, storage and capital, which is where offset runs actually fail.
- Designing a cover in Pantone for POD. Convert to CMYK, and the colour shift risk is yours.
- Expecting foil or sewn binding from POD. No platform lists them.
- Sending a POD-spec PDF to an offset printer. Different templates and usually PDF/X.
- Ordering a run to save money on a first book. The saving is per unit; the risk is the whole run.
FAQ
Q: At what quantity does offset get cheaper? A: No platform or trade body publishes a figure. Compute it from KDP's published formula and a real offset quote for your book.
Q: Is print on demand lower quality? A: For black ink text on standard stock the difference is small. The real gaps are spot colour, special bindings and paper choice.
Q: What is the minimum order for offset? A: Not published by any authoritative source. It is negotiated per printer.
Q: Can I use both? A: Yes, and many publishers do: offset for a launch run, POD for the long tail.
Q: Does POD cost the same at 10 copies as at 1,000? A: Per copy, yes. The published formula has no quantity term.