The Quick Checklist for Prepping a PDF Before You Split or Compress
You've been there. You upload a hundred-page PDF to a compression tool, wait two minutes, download the result — and the bookmarks are gone. Or you split a legal contract by page range and discover the signature page landed in the wrong chunk. Or a font that looked fine in preview now renders as boxes in the output.
These aren't random bad luck. They're almost always the result of skipping a few basic checks before you hit "process." Think of this as your pre-flight checklist — the stuff you run through before you hand your PDF over to any tool that's going to slice, compress, reorder, or otherwise touch its internals.
It takes maybe five minutes. It saves you from the kind of headaches that show up after you've already sent the file.
Before Anything Else: Know What Kind of PDF You're Working With
Not all PDFs are the same. A scanned document is fundamentally different from a text-native export, and what you can safely do to each one is different too.
- Text-native PDF: Generated directly from Word, InDesign, Google Docs, etc. Has actual selectable text. Generally handles compression and splitting well.
- Scanned/image-only PDF: Each page is literally a rasterized image. Compressing aggressively can make text unreadable. Splitting is fine, but be careful with quality settings.
- Mixed PDF: Some pages are text-native, some are scanned images (common with older contracts where someone re-scanned a signed page). Most dangerous to compress blindly.
How to check: Open the file, try to select text on a representative page. If you can highlight individual words, it's at least partially text-native. If you get a whole page selected as a single block or nothing selects at all, you're dealing with an image PDF.
The Pre-Split Checklist
☐ 1. Document the structure first
Open the bookmarks/outline panel in your PDF viewer before you do anything. Write down — or screenshot — how the document is organized. Which bookmark corresponds to which page range? This takes sixty seconds and it's your reference if something goes sideways.
Adobe Acrobat, PDF-XChange, and most free viewers show bookmarks in a sidebar panel. If your viewer doesn't, open the file in a browser — Chrome renders PDF bookmarks in a left-side panel natively.
☐ 2. Check for logical page numbers vs. physical page numbers
This one trips people up constantly. Many PDFs use logical page numbering — the document might show "Page iii" for its third physical page, then restart at "1" for the main body. When you split "pages 1–30," some tools count physical pages (what they see in sequence) and some honor the document's embedded page labels.
Always confirm: is page "1" in the split tool the same as what page "1" means to a human reader? Open the file, navigate to what the document calls page 1, and check what number your viewer's page counter shows at the bottom. If there's a mismatch, account for it in your page ranges.
☐ 3. Check for cross-references and internal links
PDFs can contain hyperlinks that jump to other pages within the same document. If you split a document and a link on page 4 points to page 87, and you've put those pages in separate output files, that link now points nowhere.
Not every PDF has these, but reports, manuals, and technical documents often do. Scan through quickly. If you see "see page X" or "refer to section Y on page Z" styled as clickable links, note where they land before splitting.
☐ 4. Flatten form fields if the PDF has them
If your PDF contains interactive form fields (fillable text boxes, checkboxes, dropdowns) that have already been filled in, some split tools will drop or corrupt that data. The safe move: flatten the form first. Flattening bakes the filled-in values into the page as static content, so there's nothing for a split tool to mishandle.
In Acrobat: Tools → Print Production → Flattener Preview → flatten. Many free tools also have a "flatten" option. Don't skip this on signed or completed forms.
☐ 5. Confirm no pages have rotation metadata quirks
Some PDFs have pages that look right in your viewer but are technically stored rotated — the viewer corrects it on display. Certain split tools preserve this metadata, others strip it. Result: your landscape-oriented diagram ends up portrait in the output, cropped or distorted.
Scroll through all the pages in your viewer. If any page looks correctly oriented, switch to a second viewer (even just a browser) to double-check it renders the same way. A mismatch flags a rotation metadata issue you'll want to normalize before splitting.
The Pre-Compress Checklist
☐ 6. Set a minimum acceptable file size target, not just "as small as possible"
Compression is a trade-off. Every tool that offers "maximum compression" is making decisions on your behalf — usually by downsampling images and stripping metadata. Before you compress, decide: what's the actual target size, and what's the minimum quality you can live with?
For email attachments: 10MB or under usually works. For print-ready documents: don't compress images below 150 DPI. For archival: consider that maximum compression can strip XMP metadata that includes copyright and authorship information.
Set your floor before you start, not after.
☐ 7. Check embedded fonts — are they subsetted?
Most modern PDF exporters embed fonts in "subsetted" form — only the characters actually used in the document, not the full font file. This is good for file size but it means aggressive compression tools that re-process font streams can occasionally corrupt character rendering.
In Acrobat, you can check this under File → Properties → Fonts tab. Any font listed as "(Embedded Subset)" is fine. If a font is listed as "Not Embedded" — stop. Compression won't fix this and could make it worse. You need to go back to the source and re-export with fonts embedded.
☐ 8. Watch for transparency and layering
Transparency effects (drop shadows, semi-transparent overlays, blending modes) are stored in PDFs as a "transparency group." Some older compression pipelines flatten transparency in unexpected ways — particularly when converting color spaces or downsampling nearby images. The result can be text that renders with a grey box behind it, or colors that shift slightly.
If your PDF was exported from InDesign or Illustrator and has complex graphics, open the compressed output on a second screen before sending it anywhere. Look for any page with overlapping elements.
☐ 9. Test one page before processing the whole document
For long documents, split off a representative page — ideally one with text, one with an image, and one with mixed content — and run your compression settings on that mini-sample first. Download it, open it, zoom in to 200%. Does the text look sharp? Does the image hold up? Does anything look blurry that was crisp before?
This takes two minutes and saves you from discovering that your compression settings were too aggressive after you've already delivered the file.
☐ 10. Preserve or export your bookmarks separately if they matter
Some compression tools — especially free web-based ones — silently strip the document outline (bookmarks) in the process. If your PDF has a table of contents built from bookmarks and those bookmarks matter to your audience, either:
- Use a tool that explicitly states it preserves document structure, or
- Export a list of your bookmarks and their target pages before compressing, so you can verify they're intact in the output.
Tools like PDF-XChange and Acrobat let you export the bookmark tree. Even a simple screenshot of the panel is better than nothing.
The Final Verification Step (Don't Skip This)
☐ 11. Open the output in a different viewer than the one you'll use to process it
Your processing tool and your usual PDF viewer often share the same rendering engine. Open the output file in something different — if you used an online tool, open the result in both Chrome and a standalone desktop viewer. Rendering differences surface problems that a single viewer might paper over.
☐ 12. Spot-check page count and page order
It sounds paranoid, but: check that your output has the right number of pages and that they're in the right order. Jump to the middle of a split file and make sure the content matches what you expected there. Jump to the last page. This takes thirty seconds and catches the one-in-fifty cases where a tool silently dropped a page or duplicated one.
A Note on Tool Choice
The checklist above assumes your tool might do something wrong. Most don't — most of the time. But browser-based free tools in particular are processing your PDF through a conversion pipeline you have no visibility into, and some of them re-encode the entire document rather than just adjusting compression parameters. That's when you lose bookmarks, flatten transparencies you didn't ask to flatten, or end up with a file that's technically smaller but visually degraded in ways that matter.
If you're working with documents that have legal, archival, or print-production significance, use a tool where you can see what it's actually doing to the file. The checklist doesn't change — but your confidence in the output improves significantly when you know the tool is respecting your document's structure rather than rebuilding it from scratch.
Run through this once. You'll probably never need most of these checks on any given file. But the one time you do — when a client asks why their bookmarks disappeared, or why page 47 of a 200-page manual is now sideways — you'll be glad you looked first.