How to compress a PDF without losing quality (the honest guide)

The promise sounds great. "Compress your PDF without losing quality." Three clicks, file goes from 24 MB to 4 MB, problem solved.

Here's what nobody on a marketing page tells you: every PDF compressor in existence makes the same set of trade-offs. The good ones make those trade-offs intelligently and tell you what they're doing. The bad ones make them randomly and hide the consequences in fine print.

This piece walks through what's actually happening when a PDF gets compressed, what each kind of compression costs you, and how to pick the right mode for your specific document. By the end, you'll know exactly what to do with that 24 MB PDF — and what to never do.

Why PDFs get big in the first place

A PDF's file size comes from three places, in roughly this proportion for a typical business document:

• Embedded images — usually 70–95% of total size. A single full-page scan at 300 DPI is around 4–8 MB on its own.
• Embedded fonts — typically 1–3% of size for typical documents, but can spike to 30%+ for design-heavy files with many typefaces.
• Vector content + text + metadata — almost always under 1% of size combined.

This proportion matters because it tells you where the gains actually come from. If your PDF is 24 MB, ~20 MB of that is almost certainly images. Compression that doesn't touch images can't help you. Compression that aggressively touches them will help you a lot — and may hurt your document.

The three compression strategies, in plain English

Strategy 1: Lossless metadata cleanup

This is the safest, most boring compression. It removes things from the PDF that don't affect how the document looks: orphaned objects, duplicate resources, unused fonts, embedded JavaScript, form-field metadata, file properties, unused images that aren't referenced anywhere.

Expected file-size reduction: 5–30%. Visual quality impact: zero. Text searchability: preserved. Vector content: preserved.

If your PDF is mostly text with light formatting (reports, contracts, invoices), lossless cleanup may be the only compression you need. Modern tools — including our Compress PDF tool's Conservative mode — apply this automatically as the first step before any image processing.

Strategy 2: Image re-compression

This is where real gains start. The tool extracts each embedded image, re-encodes it at a lower JPEG quality or smaller dimensions, and re-embeds it. The image still displays in the same position, the document still looks visually similar, but each image is now smaller.

Expected file-size reduction: 40–80% when images dominate the file. Visual quality impact: mild to moderate, depending on how aggressive the re-encoding is.

The honest trade-off: at moderate settings, the document looks almost identical to the original at normal viewing zoom. At aggressive settings, you'll start to see compression artefacts in photos and gradients, and zooming in 200% reveals visible blocking.

Strategy 3: Rasterise the whole document

The nuclear option. The tool renders each page as a single image, throws away the original vector content, embeds the rendered image, and saves. Every page becomes a picture.

Expected file-size reduction: 30–95%, varying wildly by document type. Visual quality impact: noticeable. Text searchability: destroyed. Copy-paste from PDF: broken.

This is the strategy that explains why some "smart compression" tools shrink PDFs so dramatically and then your document looks fuzzy and you can't search it. They rasterised it without telling you. If a tool ever offers a single "compress" button with no quality slider and the output file is dramatically smaller than the input, this is almost certainly what happened.

How to pick the right strategy for your document

Use the table below as a quick reference. Match your document type to the right strategy first, then adjust quality from there.

• Contract, invoice, report (text-heavy): Strategy 1 (lossless). Already-text documents have very little to compress past metadata cleanup. If size is still a problem, the document probably has embedded images of signatures or logos — try Strategy 2 at the highest quality setting.
• Scanned document or photo album: Strategy 2 (image recompression) at medium quality. This is the sweet spot — visual quality stays good, file size drops dramatically, OCR still works.
• Design portfolio, marketing brochure: Strategy 2 at high quality. Don't sacrifice quality on documents where the visuals are the point. Accept a smaller file-size reduction in exchange for preserved fidelity.
• Email attachment hitting a size limit: Strategy 2 at medium quality, with Split PDF as a backup if you're still over the limit. Splitting into two emails is almost always better than rasterising.
• Document you'll never read again, just need to archive: Strategy 3 (rasterise) is acceptable. You're explicitly trading future-usability for storage savings.

The "quality setting" myth

Most compression tools expose a quality slider — usually three or four presets. Here's what those presets actually mean in JPEG terms, and what you'll see at each:

• "High quality" / 95%+: Visually indistinguishable from the original. Modest file-size reduction (often 20–40%).
• "Good" / 80–90%: Slight softness in fine detail, visible only on direct comparison. Strong file-size reduction (50–70%).
• "Web" or "Email" / 65–75%: Visible compression artefacts in gradients and photographs. Heavy file-size reduction (70–85%).
• "Smallest file" / under 65%: Visible artefacts everywhere. Reserved for documents you genuinely don't care about.

Pick the highest-quality setting that gets you under the size limit you actually need. There's no benefit to over-compressing past your real constraint. If a recipient's mailbox accepts 10 MB attachments, don't shrink to 2 MB — go from 24 MB to 9 MB and keep the extra fidelity.

What about the tools that promise "smart" or "AI-powered" compression?

Most of them apply the same three strategies above, with marketing language wrapped around them. "AI-powered" usually means "we adaptively choose between Strategy 2 and Strategy 3 per page based on heuristics." That's fine — but you give up control over the trade-offs, and you usually can't tell whether a given page got vector-compressed or rasterised.

If you want predictability — important documents, repeatable workflows, anything you'll process at scale — pick a tool that exposes the strategy and quality settings explicitly rather than hiding them behind a single "compress" button.

Putting it all together

Compression isn't magic, and the tools that pretend it is are either over-promising or quietly destroying your document. The honest version is much simpler:

1. Identify what's making your PDF big (almost always: images).
2. Pick the gentlest strategy that hits your target size (usually: lossless cleanup, then medium-quality image recompression if needed).
3. Avoid rasterising unless you specifically don't care about future usability.
4. Verify the result is acceptable at the zoom level you'll actually view it at.

Get those four steps right and you can shrink almost any PDF by 50–80% without anyone — including you — noticing.