Why Is My PDF So Large?

By Nathan Hays · Updated July 31, 2026

A one-page agreement has no business being 40 MB, yet there it is, bouncing back from the email server. PDFs balloon for a handful of predictable reasons, and almost all of them come down to one thing: images. Here's what's actually inside your file, and how to shrink it.

First, work out what kind of PDF you have

Two quick checks narrow the problem down before you touch a single setting. Open the file and try to select a sentence with your cursor. If the words highlight as words, you have a text-based PDF, and the weight is coming from images placed on the pages or from the extras further down this list. If the cursor just drags a rectangle across the page and selects nothing, the page is a picture. You have a scan, however crisp it looks.

Then do the arithmetic: divide the file size by the number of pages. Text documents land far below a megabyte a page, often far below. Scans and photo-heavy pages land at or above it. The larger that per-page number, the more image data is sitting in the file, and image data is the only kind there is ever much of to reclaim.

1. It's a scan, so every page is a photograph

This is the big one. A scanner doesn't save words; it saves a picture of the page. At 600 DPI in color, a single letter-size page is about a 30-megapixel image. Even compressed, that's megabytes per page; a 10-page scan can easily pass 50 MB. Text-based PDFs (exported from Word, for example) store letters as fonts and vectors instead, which is why they're often 100× smaller.

It gets worse than the raw pixel count suggests, because image compression cannot tell your letters apart from the paper they sit on. Every speck of texture, every shadow in the gutter of a bound page, every slightly gray patch where the scanner lid did not quite close is detail that has to be stored along with the words. That is why a scan of a blank-looking page is rarely small, and why scanning in "document" or "text" mode, which flattens the background toward clean white, does more for file size than most people expect.

2. The scan resolution and color depth are overkill

For reading on screen or ordinary record-keeping, 150–200 DPI is plenty, and 300 DPI is archive quality. Color triples the data of grayscale for documents that are just black ink on white paper. If you control the scanner, dropping 600 DPI color to 200 DPI grayscale can cut file size by 90% before any compression at all.

3. Uncompressed or oversized embedded images

Slide decks and reports exported to PDF often embed photos at their full original resolution: a 12-megapixel phone photo squeezed into a quarter-page figure still carries all 12 megapixels. Some export pipelines also store images losslessly (PNG-style) when a JPEG would look identical at a fraction of the size.

The trap here is that shrinking a picture on the page does nothing to the data behind it. Dragging a photo's corner in a document makes the box smaller; the image keeps every pixel it had. This is the single most common reason a report with a handful of illustrations turns out to be surprisingly heavy, and it is worth fixing at the source: resize the images before you place them, and export rather than print when you save the document out.

4. Embedded fonts (the small stuff)

Every font used in the document is usually embedded in the file, occasionally including full multi-megabyte font families (CJK fonts are the classic offender). Fonts, metadata, thumbnails, and leftover edit history each add a little; they only matter once the images are under control.

Most exporters embed only a subset, the handful of characters the document actually uses, which is why fonts are usually a rounding error. The exceptions are documents that embed complete families, and scripts with very large character sets where even a subset is substantial. If your file is a few hundred kilobytes and you cannot find any images in it, fonts are the likely answer, and there is not much to be done beyond using fewer of them.

5. The file was merged from other heavy files

Merging adds sizes together: combine five 8 MB scans and you get a 40 MB PDF. If you've just merged several PDFs, compress the result once at the end rather than each piece beforehand.

6. Repeated resources and leftover revisions

Two quieter causes worth knowing about. The same logo, letterhead, or scanned signature can end up embedded once per page instead of once per document, so a 60-page report quietly carries 60 copies of the same picture. And the PDF format allows incremental saving, where each save appends the changes to the end of the file instead of rewriting it, so a document that has been edited and re-saved many times can be carrying the sediment of every earlier version. Exporting or saving a fresh copy from a PDF application usually collapses that history into one clean file.

How PDF compression works, in general

Every compressor, ours included, is choosing from a short menu of moves. Knowing the menu explains why two tools can give wildly different answers for the same file.

• Downsampling images. A 600 DPI scan of a page that will only ever be read on a screen is thrown away down to something nearer 150 DPI. Nearly all real savings come from this one move, and it cannot be undone.
• Re-encoding images. The same picture stored with a lossier JPEG setting: smaller, slightly softer, also permanent.
• Subsetting or dropping fonts. Keep only the characters the document actually uses. Modest savings with no visible cost, though dropping a font entirely can change how the page lays out.
• Removing duplication and dead weight. Collapsing repeated images into one, discarding old revisions, page thumbnails, unused objects, and metadata. Nothing on the page changes.
• Repacking the page instructions. The drawing commands behind each page can be stored more tightly with ordinary lossless compression. Small, reliable, invisible.

The first two are lossy: picture data is discarded and you are not getting it back. The last three are lossless: the file gets tidier, the page is untouched. This is why a tool that only performs the lossless moves can never do much for a scan. A scan is nothing but image data, so the only way to shrink it meaningfully is to store less of that image.

What the Compress PDF tool here does, exactly

Ours takes the simple, aggressive route, and it is only fair to describe it precisely so you can judge whether it suits your document.

It opens your PDF and draws each page as it would appear on screen, capping the width of that drawing at roughly 1000 pixels on the Smallest setting, 1400 on Balanced, and 2000 on High quality. Each drawing is saved as a JPEG, at a lower encoding quality for Smallest and a higher one for High quality. It then builds a brand new PDF in which every page is that single image, held at the original page's dimensions in points so the document still prints at its proper size. A white background goes down first, because JPEG has no transparency.

The consequences follow directly from that description. Text becomes part of the picture, so it can no longer be selected, copied, or searched. Links, bookmarks, and form fields do not make it into the new file. Every page is rebuilt this way, including pages that were pure text and did not need it. In exchange, scans and image-heavy documents get dramatically smaller, and the tool shows you the old size, the new size, and the percentage saved before you download anything. If the result is not actually smaller, it says so plainly instead of quietly handing you a worse file.

So: reach for Compress PDF when the document is a scan, a photo-heavy report, or anything where what matters is how the pages look. Reach for your PDF application's own reduce-file-size export when the document has to stay searchable or interactive.

What to expect, by document type

When a PDF cannot get much smaller

Sometimes the answer is that the file is the size it needs to be, and knowing when to stop saves an afternoon.

• It is already efficient. Text and vector artwork are stored as instructions, not pixels, and instructions are cheap. Hundreds of pages of text can weigh less than one full-page color photograph.
• Its images were squeezed hard already. Re-encoding an image that has been through a lossy pass before mostly adds smudge for very little size.
• The content genuinely is that big. A high-resolution photographic portfolio or an archival scan is large because it is supposed to be. Compressing it defeats its purpose.
• The requirement is the problem. If a portal demands 5 MB and honest compression lands at 12, split the document and send it in parts, or send a link instead of an attachment.

When the tool tells you the result is no smaller, believe it and keep your original.

How to shrink it

1. Compress the PDF. The Compress PDF tool re-encodes each page as an image at screen-suited quality, shows the savings up front, and keeps your file on your device. Scans routinely drop 70–90%. Text becomes part of the page image, so keep your original if you need selectable text.
2. Re-scan smarter if you can: 150–200 DPI, grayscale, "document" mode.
3. Split out what you need. Emailing 3 pages of a 60-page file? Extract just those pages instead of sending the monster.
4. Export, don't print-to-PDF, from Office apps; proper export keeps text as text.
5. Resize images before you place them, not after. Scaling a photo down on the page leaves every original pixel in the file.
6. Save a fresh copy of a document you have edited many times. An export writes one clean file instead of a stack of appended revisions.
7. Don't count on zipping. Most of a large PDF is already-compressed image data, and a ZIP archive can do very little with that. Expect a few percent, not a few times smaller.
8. Send a link when the file is legitimately huge. A shared folder beats an attachment that bounces.

Frequently asked questions