The strictest retailer quietly sets the standard for all of them.

You upload the same file to three retailers. One accepts it immediately, one accepts it with warnings, and one rejects it with a message referencing a specification you have never read. Nothing changed between uploads. What differs is how strictly each store checks.

Here is what each one actually validates, where they disagree, and how to build a file that satisfies all of them at once.

Each retailer checks different things

The stores are not applying one shared standard with varying enthusiasm. They have genuinely different architectures, and that produces different failure modes.

The Amazon row is the one worth dwelling on. KDP does not display your EPUB, it converts it. That conversion is forgiving of messy input, which sounds like an advantage and is actually the trap: unusual markup is silently reinterpreted rather than rejected, so problems surface as odd rendering on a Kindle rather than as an error you can act on.

Apple is the opposite. It validates against the EPUB specification and rejects files that do not conform. Authors experience this as Apple being difficult. In practice Apple is telling you your file is malformed, and Amazon was papering over it.

Build for the strictest, upload everywhere

This is the whole strategy, and it removes the need to maintain per-retailer ebook files.

A file that passes Apple's validator will pass everywhere else, because everyone else is more permissive. A file built to Amazon's tolerance may fail Apple, and worse, may render differently across devices in ways nobody flagged at upload.

The practical consequence is that formatting to the loosest platform you plan to use is a false economy that surfaces later, usually the week you were planning to announce a date. If you are going wide, build to the strictest specification once.

The specific things that fail

Rejections cluster into a small number of causes, and the error messages rarely name them clearly.

Why the same book looks different on each device

Even a file that passes everywhere will not look identical everywhere, and authors are often surprised by this in a way that reads as a fault when it is the intended behaviour.

A Kindle applies its own typeface and spacing conventions during conversion. Apple Books renders closer to what you specified. A phone shows perhaps a third of the words a tablet does at the same setting. A reader who prefers large type sees roughly half as much per screen as one who prefers small. Dark mode inverts everything, which is why an image with a white background suddenly looks like a lit panel in the middle of a page.

None of that is breakage. It is the format doing what it was designed to do, which is to let the reader decide. The practical implication for you is that the things worth controlling are structural rather than visual: that chapters begin cleanly, that navigation works, that images scale sensibly, and that nothing depends on a specific line break. Anything that only works at one type size will eventually be seen at another.

Where print differs from ebook

For ebooks, one well-built EPUB genuinely does serve every store. For print, that is not true, and assuming otherwise costs a week.

Print files must be built per printer, because paper stock differs and therefore spine width differs for an identical page count. Bleed handling and barcode placement also differ. The interior PDF is often reusable across printers; the cover generally is not, and the spine calculation and dimensions are worth reading before assuming one file covers both.

There is also a distribution difference that shapes the decision. IngramSpark reaches bookshops and libraries through the wholesale channel, which Amazon's print service does not. Authors who want physical retail availability generally need both, and therefore maintain two print cover files that must be kept in step whenever the interior changes.

Going wide versus going exclusive

There is a decision sitting underneath all of this that determines how much of it you need to care about.

Amazon offers an exclusivity arrangement for ebooks, which brings subscription readership and some promotional tools in exchange for not selling the ebook anywhere else. Taking it means one store, one validator, and none of the cross-platform concerns on this page. Declining it means going wide, which means every validator applies.

This is a commercial decision rather than a technical one, and it is genuinely contested among authors. The relevant point here is narrower: format as though you are going wide, whichever you choose. Building to the strictest specification costs nothing extra and leaves the door open. Building to Amazon's tolerance and later deciding to go wide means rebuilding, usually at the point where you are already committed to a date.

The exclusivity arrangement also applies only to ebooks. Print is never exclusive, so print distribution is a separate decision and the IngramSpark question stands regardless of what you choose for the ebook.

What happens after you upload

Acceptance is not publication, and the gap between them catches out authors planning a launch date.

Each retailer runs its own review before the book becomes buyable, and the wait varies from hours to several days depending on the store and how busy it is. New accounts are typically reviewed more slowly than established ones. Print requires an additional step, because the printer generates a proof from your files that you should approve before the title goes live, and that proof takes physical time to produce and ship if you order it on paper.

The consequence for planning is that hitting a specific publication date means uploading meaningfully in advance, not on the day. Retailers also let you set a future release date, which is the better approach: upload early, let the review clear quietly, and have the book appear on a date you chose rather than whenever processing happened to finish.

Checking before you upload

Most of this can be verified in advance, which is considerably less stressful than discovering it in a rejection email.

1. Validate the EPUB. Free validators exist and report exactly what the retailer will object to.
2. Open it on a phone, not only a desktop reader. Phones are where most ebook reading happens.
3. Set the type to the largest size and check nothing overlaps or strands.
4. Tap every table of contents entry and confirm each lands on the right chapter opening.
5. Switch to dark mode. Images with white backgrounds announce themselves immediately.
6. For print, check the page count used for your spine matches the interior you are uploading in that same session.
7. Order a physical proof before announcing a publication date.

That first step resolves most disputes before they happen. A validator tells you what is wrong in specific terms, whereas a rejection email often tells you only that something is.

The practical takeaway

One well-built EPUB for every store, print interiors that travel between printers, and print covers built per printer. That is the whole architecture, and it is worth setting up once rather than discovering piece by piece.

If any of this sounds like more administration than you want, that is a reasonable conclusion. Our formatting service delivers files built to upload to all of them, in 7 days from a final manuscript, and the full formatting guide explains what each tier covers. See what your tier costs before deciding anything.