How to Reduce PDF File Size by a Specific Percentage

PDF compression potential varies enormously by document type, and knowing the realistic ranges will calibrate your expectations before you start. Scanned paper documents — contracts, invoices, forms, medical records, handwritten notes photographed and saved as PDF — are the most compressible category. They're essentially collections of image files, and JPEG compression can often reduce them by 60–80% with text remaining fully readable. A 10 MB scanned contract can reasonably become 2–4 MB. Image-heavy design PDFs (portfolios, catalogs, presentations with many photos) typically compress by 40–70% depending on the original image resolution and how aggressively they were compressed when exported. Digital text-only PDFs (created directly from word processors or spreadsheet software with minimal imagery) compress by only 10–30% because there's relatively little image data to compress. Highly technical PDFs like CAD drawings exported as PDF, schematics, or vector illustrations may compress by only 5–20% — vectors are already compact, and the file structure overhead becomes a limiting factor. If someone asks you to reduce a scanned document by 70%, that's usually achievable. If they ask you to reduce a text-only PDF by 70%, that is likely not achievable without destroying readability.

Here is a practical iterative workflow for hitting a specific file size target with your PDF.

1. 1Step 1: Calculate your target. If you need a 70% reduction, multiply the original file size by 0.30 to find your target. Example: 15 MB × 0.30 = 4.5 MB target. If you need a maximum size (e.g., 5 MB), that's your target directly.
2. 2Step 2: Start with medium compression in LazyPDF and download the result. Check the file size. If you've already hit your target, stop — don't over-compress unnecessarily.
3. 3Step 3: If the medium-compressed file is still over your target, apply another round of compression (high level) to the medium-compressed file. Check the resulting size again.
4. 4Step 4: Open the compressed file and check quality. For scanned documents, verify that the text is sharp and readable at normal zoom (100%). For image documents, check that key visuals are acceptable.
5. 5Step 5: If you cannot hit the target through compression while maintaining acceptable quality, consider alternative strategies: split the document into smaller parts using LazyPDF's Split tool, host it in cloud storage and share a link instead of attaching it, or negotiate with the receiving system about the size limit.

Users often notice that compressing two seemingly similar PDFs at the 'medium' setting produces very different reduction percentages — one goes from 8 MB to 2 MB (75% reduction) while the other goes from 8 MB to 6 MB (25% reduction). This inconsistency is not a bug or a miscalibration in the tool — it reflects the actual content composition of each file. PDF compression primarily targets embedded image streams. A PDF where 90% of the file size is image data (typical for scanned documents) has enormous compression headroom because nearly all the data is compressible. A PDF where images account for only 20% of the file size (a text-heavy report with a few diagrams) will see much smaller percentage reductions because most of the file's data — fonts, structure, text streams — has already been efficiently encoded and has little redundancy to remove. Additionally, PDFs that were exported from applications with heavy compression settings already applied (many online form generators, email-converted PDFs, or previously compressed files) may show minimal improvement because the image data inside was already compressed. The tool has less to work with. This is why it's always better to start from the highest-quality source PDF available.

Different percentage reduction targets come with different quality implications, and knowing what to expect at each level helps you choose consciously rather than being surprised. A 30–50% reduction is generally achievable for most document types with no visible quality impact — this is the 'free lunch' range of compression where redundant data is removed without touching perceptible image quality. A 50–70% reduction for scanned documents and image-heavy PDFs is achievable with very minor quality reduction — slight softening of photographic images that is essentially invisible at normal screen viewing size. Text remains fully sharp because text in scanned documents is high-contrast black-on-white, which JPEG compression handles well. A 70–85% reduction pushes into territory where image quality is noticeably reduced on close inspection — gradients may show banding, photographs lose fine detail, and small text in scanned documents may become slightly less crisp. This range is acceptable for archival documents where readability matters more than visual fidelity. Over 85% reduction typically produces clearly degraded visual quality and is only appropriate for documents where file size is absolutely paramount and quality is secondary.

Sometimes the math doesn't work out. You need a PDF under 500 KB but the source document is a high-quality technical drawing that can only compress to 2 MB without becoming unreadable. When pure compression can't get you to your target, several practical alternatives exist. First, consider splitting the document. LazyPDF's Split tool lets you divide a multi-page document into smaller segments, each of which may individually meet the size limit. This works well for submission systems where multiple uploads are accepted. Second, if you're uploading to a web form or content management system, check whether the platform accepts links rather than uploaded files — a OneDrive, Google Drive, or Dropbox link bypasses the upload size limit entirely. Third, for PDFs containing many high-resolution images, consider whether all pages need maximum resolution. Technical documents where only certain pages contain critical images may benefit from selective treatment — though this requires more advanced tools than browser-based compression. Fourth, if the size limit is a business or platform requirement, contacting the platform's support to discuss accommodations is sometimes productive, especially for professional submissions.