Why did a compression tool tell me there were no savings to be had?

Because the file was already efficiently structured and its bulk is in the images, not in redundant data. Lossless re-saving reclaims space from duplicated objects and unused resources; if a PDF has none of those, there is genuinely nothing to reclaim without re-encoding the images themselves and losing quality.

Is there a safe way to shrink a scanned PDF without wrecking the quality?

Re-scanning is the most reliable route — 200–300 DPI in greyscale is usually plenty for a text document and can be a fraction of the size of a 600 DPI colour scan. Failing that, a desktop tool that downsamples the embedded images lets you trade quality for size deliberately, which is the honest tradeoff any real size reduction on a scan involves.

Guide

Why Your PDF Is So Large — And What Actually Makes It Smaller

A three-page PDF has no business being 40 MB. And yet people run into exactly that constantly — a scanned form that won't attach to an email, a signed contract that takes a minute to open, a report that somehow weighs more than the photos on your phone. Meanwhile a 300-page ebook downloads instantly and takes up less space than a single one of those scanned pages.

The difference isn't page count and it isn't length. It's what the pages are actually made of.

What's actually inside a PDF

A PDF is a container holding a collection of objects: text runs with instructions about where to place them, font data, images, vector drawings, and assorted bookkeeping like metadata and page thumbnails. The file's size is essentially the sum of those objects, and in practice they are wildly unequal contributors.

Text is astonishingly cheap. A page of prose stored as actual text is a few kilobytes — the characters themselves, plus a reference to a font. Even embedding the full font usually costs a few hundred kilobytes once for the entire document, not per page. This is why a long, text-heavy PDF can be remarkably small: a 300-page book might be under 2 MB, because 300 pages of text simply doesn't amount to much data.

Images are the opposite. A single photograph of a page, scanned in colour at 600 DPI, can easily be 10–20 MB on its own. Put three of those in a document and you have your 40 MB file. The page count is irrelevant — what matters is that each page is a large photograph rather than a few kilobytes of text.

The usual culprit: your PDF is a stack of photographs

If your file came from a scanner, a phone scanning app, a photocopier with a "scan to email" button, or a photo you converted, then it almost certainly contains no text at all in the technical sense. Each page is one big image. The words are visible to you because you can see them in the picture, not because the file knows they're words. (If you're unsure which kind you have, the five-second test in this guide will tell you.)

Three settings drive how enormous those page images get, and they multiply together:

Resolution. 600 DPI has four times the pixels of 300 DPI, which has four times the pixels of 150 DPI. Scanning a text document at 600 DPI is usually pure waste.
Colour depth. A full-colour scan of a black-and-white page carries three colour channels it doesn't need. Greyscale is roughly a third the data; true black-and-white is a fraction again.
Image encoding. Whether pages are stored as lossless PNG-style data or lossy JPEG-style data, and at what quality, changes the result several times over.

What genuinely reduces file size

There are really only a few levers, and they are not equally powerful.

Re-encoding the images (large effect, costs quality). Downsampling a 600 DPI scan to 200 DPI, converting colour to greyscale, or re-compressing at a lower JPEG quality will genuinely take a 40 MB file down to a few megabytes. This is what "PDF compressor" services are doing when they report dramatic savings. It is a real reduction and it is also a real loss — the discarded detail is gone. For a text document that's usually an easy trade; for a photograph or a document with fine print, less so.

Lossless structural cleanup (small effect, costs nothing). A PDF that's been edited repeatedly, or assembled by joining other files, often accumulates duplicated objects, orphaned resources, embedded fonts nobody uses any more, and stale metadata. Re-writing the file cleanly discards all of that without touching a single pixel. Savings are typically in the low single-digit percentages, occasionally much more on a badly-built file, and sometimes exactly zero.

Font subsetting (small, situational). Embedding a complete font when the document uses twelve characters of it is wasteful. Most modern PDF producers already subset fonts automatically, so there's usually nothing left to win here.

Splitting the document (not compression, often the actual answer). If the real problem is an email attachment limit rather than disk space, you don't need a smaller file so much as smaller files. Extracting the pages the recipient actually needs is frequently faster and better than degrading the whole document to squeeze under a threshold.

What HerePDF's compress tool does — and doesn't

Being direct about this, because it's the single most misunderstood thing on this site: our compress tool performs the second lever only. It re-serialises your PDF — rewriting it cleanly, dropping duplicated objects and unused resources — and it is completely lossless. Your images are bit-for-bit identical afterwards. Text stays sharp. Nothing is degraded.

The direct consequence is that it cannot meaningfully shrink a scanned document, because a scan's size lives entirely in image data the tool deliberately doesn't touch. On a 40 MB stack of scanned pages you should expect roughly 40 MB back. When the tool reports "no savings found," that is not a failure — it's an accurate report that your file was already efficiently structured and there was nothing safe left to remove.

The reason it works this way is the same reason everything here works this way: the tool runs entirely in your browser, and re-encoding every image in a large scan locally would be slow and would mean silently degrading your document. We'd rather do a modest thing honestly than claim a dramatic number and quietly cost you quality. If you genuinely need a large scan made small, a desktop tool with explicit downsampling controls — where you choose the quality tradeoff deliberately — is the right instrument.

Practical advice, in order

Work out whether your file is text-based or scanned. Everything follows from that.
If it's text-based and still large, try lossless re-saving — a bloated, heavily-edited file can genuinely shed weight this way.
If it's a scan and you control the scanner, re-scan at 200–300 DPI in greyscale. This is by far the biggest, cheapest win available and it beats any after-the-fact compression.
If you can't re-scan and you truly need it smaller, use a tool that downsamples images, and accept that you're trading quality for size.
If the goal is just getting past an attachment limit, consider splitting out the pages that matter instead.

Scanned vs Text-Based PDFs: How to Tell, and Why It Changes Everything — the distinction underneath most of this. Or try the Compress PDF and Split PDF tools, both of which run entirely on your device.