Saving a page as a document rather than a bookmark
Web pages are not permanent. Articles are edited after publication, pages are moved, sites are redesigned, and things you were relying on disappear entirely. A bookmark records where something was, not what it said, which is exactly the wrong half when the content is what mattered.
Converting to PDF captures the content as it was on the day you read it. That matters for anything you might need to point at later: terms and conditions you agreed to, a price you were quoted, a policy that governed a decision, a published statement.
It is also how you make a web page portable. A PDF can be emailed, filed alongside related documents, annotated, and read without a connection — none of which is true of a link.
Why the result does not always match the screen
Web pages are designed to flow into whatever width the reader's window happens to be. A PDF page is a fixed rectangle. Converting between the two means deciding where the content breaks, and there is no arrangement that satisfies every layout.
Elements positioned relative to the viewport rather than the document are the usual casualties: sticky headers that repeat, navigation bars that appear mid-page, floating buttons that land over the text. Content loaded by script after the initial page render may be missing entirely if it had not arrived when the conversion ran.
Pages that provide a print stylesheet convert far better, because the site has already described how it should look on paper. Many news and documentation sites do; most applications do not.
Getting a cleaner capture
Use the page's own reader mode or print preview first where one exists. Both strip navigation, advertising and sidebars, which is most of what makes a converted page look chaotic, and they usually produce something closer to a document than the page itself.
Scroll through the full page before converting so that anything loading lazily has actually loaded. Long articles with images further down are the common failure, and the resulting PDF has gaps where pictures should be.
Accept that some pages will not convert well no matter what you do. Highly interactive applications, anything behind a login that re-renders on scroll, and pages built almost entirely from script are poor candidates, and for those a screenshot of what you need is often the more honest record.
Archiving pages you may need to prove later
There is a category of web page worth saving specifically because it might be disputed: the terms you accepted, the price you were shown, the delivery date you were promised, the specification of something you ordered, a published policy that governed a decision affecting you.
For these, capture more than the content. A PDF that includes the page address and the date of capture is far more useful than one showing only the text, because the question later is not just what it said but where and when it said it. Most conversion routes include a header or footer with both, and it is worth leaving on.
Be realistic about what such a record proves. A self-made capture demonstrates what you saw and is not independent evidence, since anyone could produce one. For anything genuinely contentious, a third-party archiving service that timestamps captures independently carries considerably more weight — but your own copy is still far better than a dead link and a recollection.
It is worth building the habit around a specific trigger: capture the page at the moment you act on it, not later. The terms as they stood when you clicked accept, the price as displayed when you ordered, the specification you relied on when you chose. Pages change quietly and without notice, and the version that mattered is the one that existed at the moment of the decision, which is precisely the version nobody thinks to keep.