An RSS feed is a small XML file. A site publishes it — titles, summaries, links, dates — in a format any software can parse. No authentication required, no agreement between publisher and reader. You put the file on your server and anyone with a feed reader can subscribe. The people who designed RSS pushed the cost of participation about as low as a standard can push it, and that still wasn't low enough to produce the coverage they wanted.
Adoption did grow. A 2013 study summarized earlier estimates showing roughly 30% of consumer-media websites offering feeds in 2005, rising to 50% by 2008, with catalogs tracking millions of individual feeds. But even during that growth, the format's own history was signaling trouble.
Netscape published the original RSS in 1999 using RDF, a structured metadata vocabulary. Within months, Netscape itself published a simpler 0.91 that dropped RDF, borrowing elements from Dave Winer's own syndication format. Winer's UserLand adopted and extended that non-RDF line. An independent group then published RSS 1.0, returning to RDF with modular extensions. Winer's line continued separately as RSS 2.0, which kept the simpler structure and declared itself frozen. By 2005 the IETF had produced Atom, a clean-sheet alternative with formal governance. Three branches of a format whose entire premise was minimalism. Aggregator developers coped by building parsers that accepted all of them, which worked — but that normalization was itself the cost the format existed to eliminate, now paid by every reader, every tool, and every new developer deciding which branch deserved support.
The deeper retreat came later, and it had nothing to do with syntax. Twitter retired RSS support in June 2013, replacing open feed URLs with an authenticated API. Facebook's page RSS endpoints stopped returning data in 2015. That same period, Google shut down Reader, removing the most widely used consumption infrastructure. In 2018, Firefox removed built-in feed support, citing low usage relative to maintenance cost.
Publishers pulling feeds and browsers pulling readers are different decisions, but they rhyme. Each organization looked at what it cost to keep an open, unauthenticated channel alive and concluded the resources were better spent elsewhere. None of this was mysterious inside those companies. A platform large enough to control its own distribution has no reason to maintain a side door that lets third parties take the content without passing through its analytics, its advertising, or its ranking. The feed still worked. It had simply stopped being useful to the only people in a position to keep it running.
RSS survives in two places: where it costs almost nothing to produce, and where someone powerful insists on it. WordPress generates feeds automatically, and since WordPress powers roughly half the sites running a recognized CMS, that alone is a large technical footprint. Apple requires RSS 2.0 for podcast distribution. Some news organizations, The Guardian among them, still publish feeds because their readers use them.
Everywhere else — social platforms, algorithmic recommendation, messaging — the web's dominant content surfaces moved away from open feeds entirely. The network was always capable of carrying a machine-readable version of anything on it. What the arc of RSS suggests is that capability was never the constraint. Somebody still had to decide to produce a second output alongside the first, and then keep producing it, year after year, for an audience they could not see and did not control. Small as that ask was, it was never zero, and universal coverage is a threshold that only zero clears.