Zipping it won't help: which files actually compress and which don't

"Just zip it" is the standard advice when a file is too big. Half the time it works spectacularly and half the time it saves you two percent, and which half you land in is entirely predictable before you try.

The one rule

Compression works by finding repetition and describing it more briefly. A file full of patterns — repeated words, runs of identical pixels, structured rows — has a lot of redundancy to remove. A file that has already been compressed has had that redundancy removed, and there is nothing left for a zip to find.

So the question is never "will zip work" but "has this format already been compressed". Almost every media format has been. Almost no plain-data format has.

Compresses well — often by 70–95%

• Text of any kind: .txt, .csv, .json, .xml, .html, source code, log files, SQL dumps. These are the best case; a 200 MB log file routinely becomes 10 MB.
• Uncompressed images: .bmp, most .tiff, camera RAW to a lesser degree.
• Uncompressed audio: .wav, .aiff.
• Some databases and disk images with a lot of empty space in them.
• Many files that resemble each other — but only with the right tool, see below.

Does not compress — expect 0–3%

• Photos and video: .jpg, .png, .webp, .heic, .mp4, .mov, .mkv, .avi. All of these are compressed formats already. PNG is lossless but still compressed — a common misconception is that "lossless" means "uncompressed".
• Audio: .mp3, .aac, .m4a, .ogg, .flac. FLAC is lossless and still compressed.
• Office documents since 2007: .docx, .xlsx, .pptx. This one surprises people — those files are zip archives. Rename one to .zip and you can open it. Zipping a .docx compresses an already-compressed container.
• PDFs that are mostly images — a scanned document is a stack of JPEGs in a wrapper.
• Anything already in an archive: .zip, .7z, .rar, .gz, .apk, .jar, .epub.

The exception worth knowing: many similar files

A plain zip compresses each file separately. If you have 500 photos that share nothing, that is correct behaviour and there is nothing to gain. But if you have 500 near-identical files — daily exports, versions of a document, a folder of similar spreadsheets — the redundancy is between the files, and a zip cannot see it.

A solid archive can: 7-Zip's .7z format and tar.gz compress the whole set as one continuous stream, so repetition across files is found too. On a folder of similar text exports the difference is frequently 5–10× better than zip. The trade-off is that extracting one file means decompressing everything before it, which is why solid archives are a bad choice for something you will dip into repeatedly.

What to do with the files that will not compress

1. Video: re-encode it, do not archive it. Lowering the resolution or bitrate is the only thing that shrinks a video meaningfully — see how to shrink a video file.
2. Photos: re-export at a sane size. A 12 MP photo at full quality is often 6 MB and looks identical at 2 MB. Batch export at 80–85% JPEG quality is nearly always invisible.
3. Scanned PDFs: run them through a PDF compressor that downsamples the embedded images. 300 DPI is right for print; 150 DPI is fine for reading on screen and is a quarter of the size.
4. Everything else: stop fighting the size and change the channel. If the file is 800 MB because it is 800 MB of video, no archiving trick will get it into an email — see email attachment size limits.

More guides