What's actually inside a 40 MB .pptx?

Someone sends you a 40 MB deck. Twelve slides. Nothing on them looks like a
high-resolution photo. So where did 40 MB come from?

The answer is one unzip away, and it tells you a lot about how Office formats
actually work.

A .pptx is a ZIP full of XML

OOXML — .docx, .xlsx, .pptx — is a ZIP archive. Rename a copy to .zip
and open it:

[Content_Types].xml
_rels/.rels
ppt/
  presentation.xml
  slides/slide1.xml … slide12.xml
  slideLayouts/
  slideMasters/
  media/          <- here
  embeddings/
  fonts/

The XML describing twelve slides is measured in kilobytes. The whole deck's
structure — text, positions, colours, transitions — is tiny.

Practically all of the weight sits in ppt/media/. Sort that folder by size
and the culprit is usually obvious within seconds: a video someone dropped in
"just for reference", or one 8 MB JPEG that got reused as a background.

That has a very practical consequence: if you want to know why a file is big,
you don't need to guess. You look inside.

Cropping an image does not shrink the file

This is the one that surprises people.

Crop a 4000×3000 photo down to a neat circle on your slide. The slide now shows
a circle. The file still stores the full 4000×3000 image, because cropping is a
display operation — it changes which pixels are visible, not which pixels are
saved.

PowerPoint does give you a way to actually discard them: select the picture,
Picture Format → Compress Pictures, and tick Delete cropped areas of
pictures. That checkbox, not the crop itself, is what removes the bytes.

Most "my deck is still huge" complaints are this.

Compress Pictures is lossy and permanent

The same dialog re-encodes the images at a lower resolution. Worth knowing before
you click it:

• It is not undoable. Undo does not restore the original quality.
• Keep an uncompressed copy if there's any chance you'll need it later.
• 150 ppi is usually indistinguishable on a projector, and it's a big cut from the 220+ ppi that phones and stock photos ship with.

Video is a different story — and platform-dependent

File → Info → Compress Media re-encodes embedded video at Full HD 1080p,
HD 720p or Standard 480p. It's usually the single largest win on any deck with
video in it.

One catch: it exists in PowerPoint for Windows only. The Mac version has no
equivalent. On Mac you have to compress the video externally before inserting
it — which is a genuinely annoying asymmetry if your team is split across
platforms.

The quiet contributors

Two more things that add up:

Embedded fonts. Settings → Save → Embed fonts in the file. Fine if the
deck depends on a font your audience may not have; pure overhead otherwise.

Unused slide masters and layouts. Stock templates ship with a dozen layouts
across multiple masters. View → Slide Master, right-click, delete the ones no
slide uses. Each one carries its own background and theme assets.

And a smaller one: Save As under a new name. Incremental saves accumulate
editing history; a fresh save drops it.

Reading the file in the browser

The nice thing about all of this is that it's just a ZIP. Which means a browser
can read it too — File → ArrayBuffer, parse the central directory, list the
entries, sum the sizes. No server, no upload, no library.

That's what I ended up building: a page that takes a .pptx and shows you which
files inside it are responsible for the size,
broken down by percentage. The file never leaves the device — it's read with the
File API and nothing is transmitted.

The whole trick is that there is no trick. The format is open, the container is
a ZIP, and the answer to "why is this file 40 MB" has been sitting in
ppt/media/ the entire time.

If you want the step-by-step version instead of the format explanation, that's
here: How to compress a PowerPoint file without ruining the images.