The Link That Stopped Working When The Company Was Sold

An address that worked for years now returns nothing, or worse, returns something unrelated. The page may still exist somewhere. The link was never holding it.

A link is a location, not a copy

An address specifies a host, a path and a resource, and resolves by asking that host for that path at the moment you click.

Nothing in the mechanism guarantees the host still exists, still serves that path, or still means the same thing by it. The link is a set of instructions, not a container.

This is what made the web cheap to build and easy to extend, and the fragility is the same property viewed from the other side.

Migrations break addresses by default

Moving a site to a new platform changes how paths are constructed, and the new system has no knowledge of the old scheme unless somebody tells it.

Preserving the old addresses means building a redirect map covering every page, testing it and carrying it forward through subsequent migrations.

That work has a real cost and no visible benefit, since the pages it protects are old ones that generate little current traffic, so it is regularly skipped.

Acquisitions and closures remove whole hosts

When a company is bought, its site is typically folded into the acquirer's, and content judged non-essential is not carried across.

When a company closes, the domain lapses. It may then be registered by somebody else, at which point old links resolve to an entirely different site.

That is the worst failure mode, because it returns something rather than nothing, and nothing in the browser indicates that the destination changed hands.

Redirects decay in chains

Even where redirects are set up, they accumulate. One migration points old to new, the next points new to newer, and each hop depends on all the previous ones surviving.

Somebody eventually tidies up what looks like unused configuration, and a chain that had been quietly working for a decade stops at its first broken link.

Because nothing monitors inbound links from elsewhere, the breakage is invisible to the site owner and obvious only to the people who arrive.

Preservation is possible and is nobody's default

Archiving services capture and store copies, which is the only approach that actually survives the disappearance of the original host.

Coverage is patchy, since crawlers cannot reach material behind logins or paywalls, and pages built entirely by scripts often archive badly or not at all.

The practical position is that anything you expect to need again should be captured rather than bookmarked, because a bookmark is a promise made by somebody else.