Skip to content
All posts

PDF vs EPUB: when each one wins

PDF is paginated, EPUB is reflowable. The choice between them isn't about quality, it's about whether the document needs to look the same on every device.

MDMedia DeskFiles & Media EditorPublished May 14, 20266 min readbeginner

# The fundamental difference

A PDF describes pages — fixed-size canvases with text positioned at exact coordinates, the same on every device. A EPUB describes a document — reflowable HTML that adapts to the screen, font size, and orientation. Both formats can hold the same content. The choice between them is about how that content is consumed.

If the answer is "printed, or treated as a print-equivalent", use PDF. If the answer is "read on a screen, especially a phone or e-reader", use EPUB. Almost everything else follows from that.

# When PDF wins

Print artifacts. Forms, invoices, contracts, certificates, posters, manuals with specific page references ("see page 47"). The whole point is that the document looks the same when printed by anyone, on any printer, anywhere. PDF was designed for this in 1993 and nothing has displaced it.

Fixed-layout documents. Children's picture books, comic books, art catalogues, technical drawings, anything where the visual composition of the page is the content. EPUB has a "fixed-layout" variant (EPUB FXL) for this, but it's poorly supported and most readers fall back to PDF anyway.

Long-term archival. PDF/A is an ISO standard (19005) specifically for archival, embedding fonts and prohibiting external dependencies. Government archives, libraries, and legal systems use PDF/A because a file opened in 50 years should look identical to today.

Searchable, indexable text on the open web. Google indexes PDFs. The text inside a PDF can rank in search results. EPUB sits inside a ZIP container and is invisible to crawlers — you'd index the metadata, but not the body.

Documents where the page count matters. Academic papers, legal filings ("the seventh paragraph on page 12"), homework assignments. Reflowable formats can't reference page numbers reliably because the page is a runtime property of the reader, not the document.

# When EPUB wins

Reading on a phone. A 300-page PDF on a 6-inch screen is unreadable — either the text is tiny (whole page shown) or you scroll horizontally on every line (zoomed in). EPUB reflows to the screen width, with the user choosing the font size. It's the difference between a book and a scan of a book.

Reading on a dedicated e-reader. Kindle, Kobo, Boox, ReMarkable. These devices have grey-scale screens optimised for reflowable text. EPUB renders cleanly with adjustable font, line height, and margins. PDF on an e-reader works but feels wrong — the device is doing what your phone does, pretending to be a sheet of paper.

Accessibility. EPUB carries semantic structure: this is a heading, this is a paragraph, this is a footnote, this is alt text on an image. Screen readers can navigate that structure. Modern PDFs can be tagged for accessibility too, but most aren't, and EPUB makes accessible publishing the default rather than the exception.

Adjustable typography. The reader picks font, font size, line height, page margins, day/night mode. Visually impaired users can pump font size to 36pt and the document still reflows beautifully. PDF locks all of that.

Active links and dynamic content. EPUB supports live HTML, SVG, MathML, and (with EPUB 3) audio and video. PDF supports links but is much more locked down. Educational textbooks, interactive cookbooks, language-learning materials are EPUB-natural.

# When the conversion is lossy

Both directions of the PDF → EPUB and EPUB → PDF conversion exist on AnytimeConvert, but with different fidelity expectations.

PDF → EPUB is the harder direction. PDF doesn't carry the structure (paragraphs, headings, reading order) that EPUB needs — converters have to reconstruct it from visual cues like font size and position. Tagged PDFs convert cleanly; untagged PDFs (most of them, especially older ones and academic papers with multi-column layouts) come out crooked. Plan on manual cleanup unless the PDF was specifically authored for reflow.

EPUB → PDF is mechanical but destructive in a different way. The reflowable EPUB has to be baked into fixed pages at some chosen size. You lose the reader's freedom to pick font and font-size; you gain a printable artefact. The conversion always works; the result is just a less flexible document than the source.

# The practical guide


Document type                       Format    Why
─────────────────────────────────   ──────    ────────────────────────────────────
Novel, memoir, narrative book        EPUB      Reflow on phones and e-readers
Textbook (with diagrams, problems)   EPUB FXL  Structure + fixed image placement
Academic paper                       PDF       Page references matter
Government form                      PDF       Print-exact, fillable fields
Comic book / graphic novel           PDF       Visual layout is the content
Cookbook (modern, illustrated)       EPUB      Searchable, reflowable, accessible
Cookbook (vintage, photographic)     PDF       Visual layout is the content
Magazine                             PDF or EPUB FXL  Depends on whether it's an archive copy
Restaurant menu                      PDF       Will be printed
Sheet music                          PDF       Spatial layout is the music
Encyclopedia, reference book         EPUB      Search across volume, reflow on phone
Manual you'll actually print         PDF
Manual you'll search on a tablet     EPUB
Legal document                       PDF       Page references, certified copy

# A worked decision

You wrote a novel. You're publishing it as an ebook on Amazon, Apple Books, and Kobo, plus offering a "printable PDF" for sentimental readers.

  • EPUB: the primary deliverable. Generated from your Markdown source via Pandoc. One file, validated with epubcheck, uploaded to all three storefronts. Reflows on every device, accessible, searchable.
  • PDF: a secondary, derived deliverable. Generated from the same Markdown to A5 page size at 11pt serif with ~0.75-inch margins. Looks fine printed at home; readable enough on tablet. The PDF is a print-substitute, not an ebook.

Don't try to use the PDF as your primary ebook. The phone-readers will hate it.

# Bottom line

PDF is the right answer when the document needs to look the same everywhere — forms, contracts, archives, anything paginated. EPUB is the right answer when the content needs to be readable everywhere — books, articles, anything text-heavy. Most people overuse PDF because it's the format that opens by default; switching to EPUB for the documents that earn it is one of the cheapest readability upgrades in publishing.

Common questions

Frequently asked.

Can I read EPUB on a Kindle?

Yes — Amazon's Send-to-Kindle service accepts EPUB directly as of 2022. You email or upload the file, Amazon converts it to its internal format on your device. Older Kindles without recent firmware may need a manual conversion via Calibre to MOBI/AZW3 first.

Why does PDF look so much worse on a phone?

Because PDF is paginated. The page was laid out at print size (US Letter or A4) and the phone has to either show the whole page (tiny text) or zoom in and force horizontal scrolling. Neither works well. EPUB reflows to fit the screen.

Which format does Google index better for SEO?

PDF gets indexed; EPUB doesn't, by default. Google can extract text from PDF and show it in results. EPUB sits inside a ZIP container and isn't crawled. For documents you want findable on the open web, PDF wins. For documents that exist inside a reader's library, EPUB wins.

Nowe posty, raz w tygodniu.

Praktyczne przewodniki dla programistów. Bez spamu. Anuluj subskrypcję w dowolnym momencie.

Tools mentioned

Pick up where the post leaves off.

Keep reading

More from the field notes.