A practical file conversion checklist before you upload anything

A conversion goes wrong in one of three places: the format you picked, the copy you overwrote, or the thing inside the file that you never opened before sending it. None of the three is a hard technical problem. All three stay invisible until the file has already landed in the inbox of the person who asked for it, which is exactly when fixing it costs the most. The checklist below is the order I work in, and the order matters more than the tools.

Let the receiving system pick the format, not your habit

Format is not a preference. The upload field, the print shop, or the school portal has already decided, and your job is to read that decision before you convert anything. Open the form and note three things: the accepted extensions, the maximum size, and whether it wants one file or several. A field that accepts PDF only will reject a perfectly good DOCX no matter how clean the layout is. A print order that asks for 300 dpi at final trim size will reject an image exported at screen resolution even though the picture looks sharp on your monitor.

When nothing is specified, choose by what must survive. Pick PDF when the layout has to look identical on a device you will never see, because that is the only thing PDF is genuinely good at. Pick a document format when the recipient has to edit or comment. For photographs pick JPEG, for flat graphics and screenshots with text pick PNG, and convert phone photos out of HEIC before sending them to anyone whose device you have not confirmed, since a file that will not open is worse than a file that is slightly larger.

Keep one untouched original in the same folder

Before the first conversion, duplicate the source and add a suffix such as orig to the copy name. Keep it beside the working file rather than in a separate archive folder, because a backup you have to go looking for is a backup you will not use when you are in a hurry. This is not paranoia about disk space. It is because most conversions throw something away permanently and you usually discover which thing three steps later.

The losses are predictable. Re-encoding a JPEG that was already compressed adds new artifacts to the artifacts already there, and there is no way back. Turning a layered document into PDF flattens the structure, so the text you wanted to reflow is now fixed in place. Compressing a scan hard enough to fit a size limit can smear the digits in an account number until nobody can read them, and once the original is gone your only option is to scan the paper again.

Open the output before you trust the file size

A conversion that finished is not a conversion that worked. Open the result and check the page count against the source, because tools that hit a memory limit sometimes write a truncated file without complaining. Scroll to any page with a table or an unusual typeface and look for font substitution, which shows up as changed line breaks, a heading that now wraps onto two lines, or a page that has quietly become two pages.

For anything scanned, try to select a line of text. If the selection highlights nothing, the file is an image wrapped in a PDF container, which is fine for a human reader and useless for any system that expects to search or extract from it. Also check page rotation on scans fed through a sheet feeder, since a single upside-down page in a twelve-page contract is the kind of thing a reviewer sends straight back.

Check what travels along with the content

Files carry more than what you see on the page. Document metadata usually holds the author name and often the name of whoever created the template, phone photos can carry the coordinates of where they were taken, and comments or tracked changes sometimes survive an export into the final PDF as visible annotations. Before a file leaves your device, look at the document properties panel and at any comment sidebar, and strip what does not belong to the recipient.

Filenames leak too. A file called salary-negotiation-draft renamed to nothing before upload still shows the old name in some export paths, so rename deliberately rather than relying on the tool to forget. If the file is going to a public form or a shared drive, assume that everyone with the link reads the filename first and the contents second.

Before you commit

• Confirm the final format before uploading
• Keep an untouched original copy
• Check privacy and file-size limits

Run the four passes in order and the whole thing takes a couple of minutes: confirm the target format from the form, duplicate the original, open the output and read it, then clear the metadata and fix the name. The point is not perfection. The point is that every failure mode listed here is silent, so the only defence is looking, once, before the file stops being yours.