In 1995 and 1996, if you wanted to find more websites about a topic you cared about — Star Trek, knitting, military history, your favorite obscure indie band — you didn’t search for them. There was no Google. AltaVista existed but had limited reach. What you did was follow the little navigation widget at the bottom of a page that said something like “Previous | Random | Next | List” and click your way around a webring.

A webring was a circular linking arrangement among independently-run sites that shared a topic. Each member site agreed to display the navigation widget, which would send visitors forward to the next site in the ring, backward to the previous one, or to a randomly selected member. Click “next” enough times and you’d return to where you started.

It was the original content discovery system. And then it disappeared, almost entirely, in the space of about five years.

How they actually worked

The original WebRing was built by a teenager named Sage Weil in 1995. He was a high school student in Florida at the time, and the system he built ran on a CGI script hosted on a server he managed himself. Site owners would register their site with the central WebRing service, get assigned a numeric position in the ring, and paste a snippet of HTML into their pages that linked back to the central script for navigation.

When a visitor clicked “next” on Site A, the script would look up Site A’s position, find the next member in sequence, and redirect the visitor there. The clever bit was that the central script handled all the routing logic, so individual site owners didn’t need to coordinate with each other. They just needed to maintain the navigation widget and not pull their site offline without notice.

By 1997, WebRing.org listed thousands of rings spanning tens of thousands of member sites. Topics ranged from the high-minded — academic research collaborations, professional associations — to the deeply niche. There was a ring for sites about specific X-Files episodes. There was a ring for fan sites about a single Welsh actor whose name I no longer remember. There was an “Unofficial Geocities Star Wars” ring with eighty-seven members at its peak.

The Geocities relationship

Webrings and Geocities were symbiotic. Geocities organized homepages by topical neighborhoods (Area51 for sci-fi, Athens for academic content, and so on), but the neighborhood structure was loose and often confusing. Webrings provided the actual semantic organization that the neighborhood metaphor only gestured at.

A Geocities user with a Star Trek tribute page would join the appropriate Trek webrings, which gave them inbound traffic from other Trek fans. The traffic was real and meaningful. Personal sites that were members of active rings could see hundreds of visits per day routed entirely through ring navigation, which was a lot of attention for an amateur webmaster’s hobby project in 1998.

The Yahoo! acquisition and the slow decline

Yahoo! bought WebRing in 1999 as part of its broader strategy to acquire community-oriented web properties. (Yahoo! also bought Geocities the same year, and Broadcast.com, and several other now-vanished services.) The acquisition itself wasn’t catastrophic, but the integration was clumsy. Yahoo! restructured the WebRing service repeatedly, changed the user interface in ways that broke existing rings, and generally treated it as a low-priority asset.

In 2001, Yahoo! sold WebRing back to a private operator. By that point, the search engine wars had been settled in Google’s favor, and the rationale for webrings was disappearing. If you wanted to find sites about Star Trek, you didn’t follow a ring anymore. You searched Google. The query returned more results than any ring contained, ranked by relevance, in less time than it took to click through three sites.

Webrings limped on through the 2000s as a nostalgic curiosity. WebRing.org still exists as of 2026, though the active ring count is a tiny fraction of what it once was.

What we actually lost

Here’s the thing that’s worth thinking carefully about. Search engines are excellent at retrieval. You ask a specific question, you get a specific answer. What search engines are bad at — and what algorithmic feeds turned out to be even worse at — is open-ended discovery within a community.

When you clicked “random” on a Star Trek webring in 1998, you got a site you wouldn’t have found through search because you didn’t know to search for it. You got someone’s idiosyncratic project — their fan fiction, their model spaceship gallery, their detailed argument about which Captain was best — and you got it because they were part of a community of people who had agreed to point at each other.

That mutual agreement to direct attention to peers, without algorithmic mediation, has almost no equivalent on today’s web. Newsletter recommendations are the closest thing, maybe. Some Mastodon and Bluesky communities have informal versions of it. But the structural form of the webring — the explicit, voluntary, topic-bounded mutual link — has not been recreated at scale.

There have been attempts. A small revival of the webring concept happened around 2020-2022, particularly within the indie web and personal site communities. Sites like hotlinewebring.club and various themed rings on neocities.org demonstrate that the underlying mechanism still works. The scale is tiny compared to the 1998 peak, but the rings that exist are often genuinely good entry points into communities of people making interesting things.

For internet historians, webrings are worth remembering because they represent a road not taken. The web could have remained a network of small interconnected communities pointing to each other deliberately. Instead, it became a network of large platforms pointing users at content algorithmically. The webring is a small artifact of the first version. Worth knowing about, even if it can’t quite be brought back.