Add a watermark

A watermark fails in one of two directions, and both are common. Too faint and nobody notices it, so a confidential document circulates looking ordinary. Too strong and the document becomes tiring to read, which is how watermarks end up being stripped out by whoever has to work with the file.

Opacity around 15-25% works for most documents. Light grey at low opacity is the least intrusive; red carries more urgency and needs to be fainter to stay comfortable. Test at the size the document will actually be read — a mark that looks subtle on a full screen can be overwhelming on a phone.

A single large diagonal mark across the centre is the strongest signal and the most obtrusive. A corner mark is unobtrusive and easy to crop away. Tiling covers the page in repeated text, which is the hardest to remove or crop around and the most disruptive to read — appropriate for genuinely sensitive material, excessive for a draft.

A watermark deters casual misuse and establishes provenance. It makes it obvious that a document is a draft, or confidential, or was issued to a particular person — and that last one is the most effective use of all, because a recipient who knows their own name is on every page behaves differently.

It is not access control. Anyone who can open the document can read every word of it, watermark or no. It is also not permanent in an absolute sense: because the mark is drawn into the page, removing it means editing or re-rendering the page, which is more effort than deleting an annotation but well within reach of anyone determined.

If the document must not be readable by the wrong people, encrypt it with Protect PDF. A watermark and a password address different problems, and combining them is reasonable — the watermark tells the reader what they are holding, the password decides who gets to hold it.

Adding several PDFs at once applies the same watermark to all of them, which is what you want for marking a folder of documents as confidential or a set of drafts before circulation.

The more interesting use is the opposite: watermarking the same document once per recipient, each with that recipient's name. That means one pass per name, but it turns a leaked document into a traceable one, and the knowledge that it is traceable is what actually changes behaviour. It is the standard practice for circulating sensitive material to a known list.

The mark is drawn on top of existing content, so it sits over the text and images rather than behind them — that is why opacity matters. Everything underneath is unaffected: text stays selectable and searchable, images keep their resolution, and the document's structure, links and form fields are unchanged.

The watermark text itself is drawn as text, so it is part of the page's content stream. File size grows only marginally.

Applying a watermark to a document that is already watermarked stacks the second on top of the first rather than replacing it. There is no un-watermark operation, so keep your unmarked original — once you have only the stamped copy, the mark is part of the pages.