Why Is My PDF So Big? (And How to Actually Shrink It)

You scanned a lease with your phone. Eight pages. The file is 9.4 MB, your outgoing mail caps at 10, and the agency's inbox rejected it anyway. Meanwhile a 40-page contract from your lawyer arrives as a 300 KB attachment without breaking a sweat.

The difference isn't luck. PDF size follows directly from what's inside the file, and once you know how to read the symptoms, you can usually predict the fix before you try anything.

The 30-second diagnosis

Divide file size by page count. That single number tells you most of the story:

• 50-200 KB per page: a normal text document. There's not much fat to trim here.
• 300-800 KB per page: text with some images, or heavy font embedding. Moderate savings available.
• 1 MB or more per page: each page is essentially a photograph. This is where compression works miracles.

Second check: try selecting text with your mouse. If you can highlight individual words, it's a real text PDF. If your cursor just draws a box, every page is an image — almost certainly a scan — and the size problem is an image resolution problem.

Where the megabytes actually hide

Phone scanning apps are the biggest offender I see. They save each page as a full-resolution photo from a 12-megapixel camera. That's a 4000 by 3000 pixel image to represent a sheet of paper that will be viewed at roughly 1000 pixels wide. You're carrying twelve times more image data than anyone will ever look at, per page.

PowerPoint and Keynote exports come second. Every slide background, every pasted screenshot travels at its original dimensions, and a 30-slide deck quietly becomes 25 MB.

Fonts are the sneaky third source. A PDF that embeds four full font families can carry 1-2 MB of typeface data before a single word is on the page. Well-made PDFs embed only the characters actually used (called subsetting); quick-and-dirty export tools sometimes embed everything.

The dpi math that explains everything

Image data grows with the square of resolution. That's the whole game.

A page scanned at 600 dpi contains four times the data of the same page at 300 dpi, and sixteen times the data of 150 dpi. For a document that will be read on a screen, 150 dpi is genuinely enough — screens themselves render around 100-200 dpi. Print needs 300. Nothing needs 600 except archival work and lawyers who bill by the megabyte.

This is why compression tools talk in dpi targets. Resampling a scan from 600 to 150 dpi doesn't shave a bit off the edges — it removes fifteen-sixteenths of the image data.

How to shrink it, ranked by impact

1. Run it through a compressor and pick the right level. You can compress the PDF online in a few seconds — choose the balanced setting for email, or the strongest one when size is all that matters. Be realistic about the outcome, though: in testing, a 4.7 MB image-heavy file dropped to 70 KB, a 98% cut. But a lean text PDF might only lose 10-15%, because there's simply less to remove.

2. If it's a scan you control, rescan at lower dpi. Changing your scanner or scanning app from 600 to 150 dpi beats any after-the-fact tool. Most scanning apps hide this setting under quality options.

3. Resize images before they enter the document. If you're building the PDF yourself from Word or Docs, shrink photos to roughly the size they'll be displayed before pasting them. Half the width means a quarter of the data.

4. Don't "print to PDF" as a fix. It often backfires: the printer driver can rasterise clean text into images, producing a file that's bigger AND unsearchable. Print-to-PDF is a layout freezer, not a compressor.

5. Split as a last resort. When a hard upload limit stands and quality can't drop further, cutting the document into parts is the honest workaround.

The email limit trap

Here's the one that catches people who did everything right. Gmail advertises a 25 MB attachment limit, Outlook.com 20 MB, and plenty of corporate servers 10 MB. But attachments travel base64-encoded, which inflates them by about a third in transit. A 19 MB file becomes roughly 25 MB of encoded message and bounces off the very limit it seemed to fit.

Rule of thumb: aim for 70% of the stated cap. For a 10 MB corporate limit, that means a 7 MB file or smaller.

When nothing seems to work

Two situations produce stubborn files. The first is a PDF that's already been compressed once — compression doesn't stack, and a second pass mostly softens images further for single-digit gains. The second is a file whose images already sit at the target resolution; there's nothing left to resample, so the tool has nothing to work with.

In both cases the answer lives upstream: get a better original, or accept the size and split the file.