Most oversized PDFs aren’t big because the document is long. They’re big because the scan settings were wrong. A single-page certificate scanned at 600 DPI in color can easily come out at 8-10MB — for a page that, scanned correctly, would be under 300KB. Get the settings right once, at scan time, and you skip the whole problem instead of fixing it after the fact.
Here’s what actually matters, in order.
DPI: 150 is usually enough, 300+ is for a different job
DPI (dots per inch) controls how much detail the scan captures. More DPI means a sharper image and a much bigger file — the relationship isn’t linear, it’s closer to quadratic, because you’re adding pixels in both directions.
For a portal upload, an admission form, or a scanned certificate that someone will view on a screen, 150 DPI is the right target. Printed text, signatures, and stamps stay fully legible at 150 DPI — you can zoom in on a signature and still tell it’s a real signature, not a smudge.
300 DPI and above is overkill for this use case. That resolution exists for print reproduction and archival scanning, where someone might enlarge the document or run heavy OCR on faint text. A government portal checking that you uploaded your marksheet doesn’t need print-quality detail — it needs to read the numbers. Scanning at 300 DPI when 150 would do roughly quadruples the file size for no practical benefit.
If your scanner or app only offers presets like “Draft / Normal / Best” instead of a DPI number, Normal or Draft maps to roughly 150-200 DPI, which is the setting to use. Save “Best” for documents you’re archiving permanently or ones with genuinely tiny print you need to zoom into heavily.
Grayscale beats color for black-ink documents
If the original document is black text on white paper — which describes most certificates, forms, affidavits, and ID photocopies — scan in grayscale, not color. A color scan stores three channels of data per pixel; grayscale stores one. For a document with no color information to begin with, color mode is pure overhead: it triples the data for an image that looks identical to the eye.
The one exception is a color photo ID — a passport, a driver’s license, a photo page — where the photo itself carries color information that matters. Scan those in color, but don’t reach for 300+ DPI; 150-200 DPI in color is still far smaller than a high-DPI color scan and plenty sharp for an upload.
A third option exists, and it’s even smaller. If your scanner or app offers something beyond Color and Grayscale — usually labeled “Black & White,” “B&W,” “Text,” or “Line Art” — that’s true 1-bit black-and-white: every pixel gets pushed to pure black or pure white, with no gray in between. It compresses harder than grayscale because there’s less information per pixel to store, and it’s the smallest file you’ll get out of a scan. Use it for documents that really are just black ink on plain white paper. Don’t use it for anything with a light gray stamp, a pencil signature, a faint watermark, or a photocopy that’s already a bit washed out — 1-bit mode has no gray to represent those with, so they either vanish or turn into a solid black smudge. Grayscale is the safer default; reach for true black-and-white only after you’ve checked the result still looks right.
If you’ve already scanned something in color and don’t want to redo it, Grayscale PDF converts the existing pages after the fact. It outputs grayscale, not 1-bit black-and-white, and it usually shrinks the file as a side effect — but if hitting a specific size is the actual goal, Compress PDF is the tool built for that.
Crop to the document, not the scanner bed
Every extra bit of background in the scan — the gray border of the flatbed glass, a strip of the desk next to a phone photo — adds pixels that compress badly (mixed textures and shadows don’t compress as cleanly as flat white paper) and add nothing useful. Crop tight to the document edges before you save or export. Most scanning apps do this automatically; on a flatbed scanner, check the preview and drag the selection box in before you scan, rather than scanning the whole bed and cropping later in an editor.
Use a scanning app, not just the camera
A plain photo from your phone’s camera app works, but a dedicated scanning app (Google Drive’s built-in scanner, Adobe Scan, Microsoft Lens, or your phone’s native “Scan Document” feature) does three things a photo doesn’t: it auto-detects and crops to the document edges, it corrects the perspective so a slightly angled shot looks flat and square, and it applies a contrast filter that pushes the background toward pure white and the text toward pure black — which compresses dramatically better than a photo’s natural gradient of off-white page and soft shadow. If you’re scanning with a phone at all, use the scanning mode over the camera mode.
Most phone scanning apps don’t show you a DPI number at all — they capture at a fixed high resolution and then let you pick a color mode from a filter or palette icon, usually somewhere on the edit screen after you capture the page. That filter picker is where the color-mode advice above actually applies on a phone: choose Grayscale (or B&W, with the caveats above) for a plain document, and save Photo or Color for pages where color information genuinely matters.
What size limit are you actually scanning for?
Portal limits vary a lot, and they change without much notice, so don’t treat any number here as fixed — check the specific portal before you scan a large batch. But as a general sense of scale:
- Government application and e-filing portals commonly cap uploads somewhere in the 5-15MB per file range, though some allow less and a few allow more.
- College, scholarship, and job application portals tend to run stricter — often 2-5MB per document.
- Email has its own ceiling: most providers cap a single message around 20-25MB total, and that’s shared across every file you attach, not per file.
Given how much these vary, the useful habit isn’t “hit the limit” — it’s “beat the limit with room to spare.” Aim to land around 60-70% of whatever the stated cap is. That way a slightly-off memory of the number, or one extra page you didn’t account for, doesn’t send you back to rescan or recompress at the last minute.
If you’ve already scanned the batch too large
Sometimes you don’t get the choice — someone hands you a USB drive of scans done on an office photocopier at whatever settings it defaults to, or you’re working from documents someone else scanned years ago. If you’re looking at a stack of oversized PDFs right now and need them under a portal’s size limit today, Compress PDF fixes that after the fact: pick a target size or a preset, and it re-encodes each page down to fit, entirely in your browser.
But that’s the fix for scans that already exist. For the next batch, or before you scan a large set of documents for the first time, set grayscale and 150 DPI before you start, crop tight, and use a scanning app — and you won’t need the fix at all.