Open a bookmark you saved four years ago. There is a decent chance you get a 404, a redirect to somebody’s homepage, or a page that has been quietly rewritten into something you did not save.
Your bookmark is fine. It is exactly where you left it, spelled correctly, pointing at the address you chose. The thing it points at is gone.
That is link rot, and it is the failure mode nobody plans for because the tool never reports it. A dead bookmark looks identical to a live one right up until you click it.
- A bookmark is a pointer, not a copy. You saved an address. You never saved the page.
- 38% of pages that existed in 2013 were gone by 2023, according to Pew. The rate is worse for the small sites you bookmarked precisely because they were good.
- Backups do not help. Restoring an export gives you the same dead links back, in the same folders.
- The archive is a lottery. The Wayback Machine has some pages and not others, and almost nothing behind a login.
- The only real fix is keeping the text at the moment you save, on a machine you control.
Why it happens more than you expect
Sites get redesigned and the URLs change. Companies get acquired and the blog goes. A writer moves from Medium to their own domain and does not redirect. A publication puts its archive behind a subscription. Someone deletes a post because they changed their mind about it.
None of those are unusual events. They are the normal life cycle of a web page, and any of them turns your bookmark into a link to nothing.
The Pew Research Center study on disappearing web content put a number on it. 38% of the pages that existed in 2013 were no longer reachable in 2023. A quarter of all pages collected between 2013 and 2023 had gone by the end of it.
The uncomfortable part is which pages go. Big sites with money and legal obligations keep their archives. The pages that disappear are personal blogs, small forums, one-person newsletters and company sites that got acquired. Those are disproportionately the ones you bookmarked, because you bookmarked the thing that was genuinely useful rather than the thing everybody had already written about.
Why your backup does not save you
This is the part that catches people out, and it catches them out at the worst moment.
Exporting your bookmarks feels like insurance. Chrome writes an HTML file, you put it somewhere safe, and you stop worrying. Then something goes wrong, you restore the file, and you get every one of your links back exactly as they were.
Including the dead ones.
A bookmark export contains titles and addresses. That is the entire format, and it has been the entire format since Netscape invented it in 1994. There is no page content in there, because your browser never kept any. A backup protects you from losing the list. It cannot protect the pages, because the pages were never yours.
If you want the full version of that argument, how to back up your bookmarks properly goes through what an export does and does not cover.
The archive is not the safety net people think it is
The usual advice is to check the Wayback Machine, and it is worth trying. The Internet Archive has crawled an enormous amount of the web and has saved many pages you will want back.
It has not saved all of them. It works from crawls, so a page has to have been crawled before it died, and coverage is thinnest for exactly the sites most likely to disappear. It also respects blocks, cannot see anything behind a login, and generally does not hold paywalled articles.
Google’s cache used to be a second chance. Google retired it in 2024, so that option is gone.
So the archive is a reasonable thing to try after the fact and a bad thing to rely on in advance. It is a lottery ticket, not a backup.
What actually works
Stop saving the address. Save the page.
The distinction sounds pedantic and it is the whole thing. A copy made at the moment you save contains what you actually read. A copy made later, or fetched later, can only ever contain what still exists later, which is the part in question.
Archivists have a word for this and it is not “export”. It is capture, and the timing is the entire point.
There are a few ways to get it:
- Save the page as a file. Chrome can save a full page to disk, and it works. What you get is a folder of HTML and assets per page, unsearchable and unsorted, which is why almost nobody keeps it up past the first week.
- Print to PDF. Fine for one important document. Hopeless as a system, because a folder of 400 PDFs is not something you will ever search.
- Use a tool that captures on save. Read-later apps like Instapaper and Readwise Reader store a copy of the article text server-side, which is genuinely good protection against the page dying. What they cannot reach is anything behind your login, because they fetch the URL from their own servers rather than reading your screen.
- Capture in the browser. An extension that reads the tab already open in front of you gets whatever you can see, including subscription articles and internal company pages.
Where ContextBolt fits, and where it does not
I built ContextBolt, so read this part knowing the bias.
It is a Chrome extension. Click the icon on any page and press Save this page, and it keeps the title, the author, the site and up to 10,000 characters of the article text, stripping the navigation, adverts, comments and cookie banners first. That copy lives on your machine. When the page 404s in three years, your copy still has the words in it.
Because it reads the tab already open rather than fetching the URL, it works on pages that server-side tools cannot reach: articles on subscriptions you pay for, documentation behind a company login, anything you had to sign in to see.
Two honest limits. It has no reader view, so it saves and finds articles rather than presenting them nicely to read. And there is no import, so it protects what you save from today, not the 3,000 links already sitting in your bookmarks bar.
That second one matters for this topic specifically. If your existing bookmarks are already rotting, nothing here retroactively saves them. The realistic move is to open the twenty you would genuinely be annoyed to lose and save those properly, which takes about ten minutes, and let the rest go.
The one habit worth building
Whatever tool you land on, the change is the same and it is a change in timing rather than in effort.
You are already deciding a page is worth keeping. That decision is the moment you have the page in front of you, fully loaded, past whatever login or paywall it sits behind. Copy it then.
Everything else is a plan to get it back later from someone who has no obligation to keep it.