RSS feed SEO: 14 of 26 sites have a feed, 7 say where it is
RSS feed SEO is a discovery problem, not a readership one. On 2026-08-20, 14 of 26 homepages served a working feed and only 7 declared it in the head — the freshest file in the panel was undeclared.
FIELD TEST · 2026-08-20 · 30 sites · single crawl plus 11 guessed paths
Sample and method: the same 30 sites used since 2026-08-15. Each homepage fetched once on 2026-08-20 with a browser user agent, the <head> parsed for <link rel="alternate">, then every discovered feed fetched and parsed. For the sites that declared none, 11 conventional paths were tried in a fixed order.
RSS feed SEO is not a question about whether people still read feeds. It is a question about whether a machine can find the file. Fourteen of 26 sites in this panel serve a working feed. Seven of them say so in the homepage head. The other seven only answer if you already know the URL.
How we measured it
Two passes, both fixed before the run.
- Parse the homepage
<head>for<link rel="alternate">with an RSS or Atom media type, then fetch the first one declared and count its<item>or<entry>elements. - For every site that declared nothing, try 11 paths in order:
/feed,/feed/,/rss,/rss.xml,/feed.xml,/atom.xml,/index.xml,/blog/rss.xml,/blog/feed,/blog/feed.xml,/blog/rss, stopping at the first that returns 200 with parseable XML and at least one item.
The usual four are excluded: stackoverflow.com and medium.com returned 403, www.canva.com and www.reddit.com returned 200 with a challenge or shell page. That leaves 26 homepages whose head we could actually read.
Seven sites declare a feed, and all seven feeds work
Declaration is rare and reliable. Every feed declared in a homepage head resolved on the first try: HTTP 200, valid XML, items present.
| Site | Items | Newest item |
|---|---|---|
| www.nytimes.com | 26 | 2026-08-20 04:41 UTC |
| www.theverge.com | 10 | 2026-08-19 23:52 |
| techcrunch.com | 20 | 2026-08-19 23:32 |
| www.wired.com | 50 | 2026-08-19 23:00 |
| news.ycombinator.com | 30 | 2026-08-19 17:32 |
| www.netlify.com | 30 | 2026-08-19 16:51 |
| developer.mozilla.org | 71 | 2026-06-15 |
Netlify declares three separate feeds — blog, changelog and knowledge base — which is the only site in the panel treating the head as a directory rather than a single pointer.
MDN is the one to look at twice. Its feed is declared, valid and 66 days out of date on the day we read it. Declaring a feed and maintaining it are two different commitments, and a consumer polling that URL has no way to distinguish "nothing published" from "nobody is watching this file".
Seven more have a feed and do not mention it
Of the 19 sites that declared nothing, seven answered a guess within the first few paths tried.
| Site | Path | Items | Newest |
|---|---|---|---|
| arstechnica.com | /feed | 20 | 2026-08-20 00:18 |
| webflow.com | /blog/rss.xml | 100 | 2026-08-19 13:29 |
| vercel.com | /blog/rss.xml | 1,486 | 2026-08-19 |
| nextjs.org | /feed.xml | 72 | 2026-08-18 16:00 |
| supabase.com | /rss.xml | 420 | 2026-08-18 |
| slack.com | /blog/rss | 20 | 2026-08-07 |
| discord.com | /blog/rss.xml | 100 | 2026-07-14 |
Ars Technica is the sharpest example of the whole crawl. Its feed was eight hours old when we read it, fresher than five of the seven declared feeds, and its homepage says nothing about it. The most current file in this panel is one you can only reach by typing /feed on a hunch.
Vercel's is the largest: 1,486 items in a single file, the entire archive rather than a recent window. That is a different kind of object from Netlify's 30-item file, and both are called a feed.
A feed nobody can find is a sitemap you forgot to declare.
What RSS feed SEO actually gets you
The concrete, documented use is not traffic from readers. It is that Google accepts a feed as a sitemap format. From the sitemap guide, read 2026-08-20 at developers.google.com: "RSS, mRSS, and Atom 1.0 sitemaps are similar in structure to XML sitemaps, however they are often the easiest to provide because CMSes automatically create them."
The page lists the trade-off in the same breath: a feed only covers recent items, so it complements a full sitemap rather than replacing one. Which is exactly the shape most of these files have — a rolling window of 10 to 100 recent entries, updated when something ships.
Twelve of 26 sites returned nothing feed-like on any of the 11 paths: stripe, linear, notion, figma, railway, framer, shopify, cloudflare, github, substack, bbc and wikipedia. Several of them publish constantly. They have simply moved discovery entirely onto sitemaps and their own distribution.
What this crawl cannot tell you
Eleven paths is a guess list, not a proof. A site could publish a feed at /blog/atom or behind a subdomain and land in our "no feed" column incorrectly. Absence in this data means absence from those 11 paths on that day, nothing more.
We did not test whether any AI engine or search crawler reads these feeds. Google documents accepting one as a sitemap; what any answer engine does with a feed URL is not something this measurement touches.
And we cannot separate a quiet publisher from an abandoned file. MDN's newest entry is dated 2026-06-15, which is either 66 days without a post or 66 days without anyone checking. One fetch of one URL does not distinguish those.
What to do about your own feed
Three checks that take five minutes.
- Fetch your own feed URL in a private window and confirm it returns 200 with parseable XML. Seven of seven declared feeds in this panel did; that is the easy half.
- Put
<link rel="alternate" type="application/rss+xml" href="…">in the head of the page it belongs to. Half the working feeds here are invisible without it. - Do not declare a feed you have stopped filling. A 66-day-old newest item is a signal, and not the one you wanted to send.
The other discovery line — the Sitemap: directive in robots.txt — is measured in the robots.txt sitemap survey, and how much the dates inside those files are worth is in the sitemap lastmod survey. When you want an engine to know about a change now rather than at the next crawl, the active route is IndexNow setup. Keeping all three lines pointing at the same set of URLs is what QueryWin is being built to do.
Common questions
How did you measure this?
One homepage request per site on 2026-08-20, head parsed for feed autodiscovery links, every discovered feed fetched and parsed. Sites declaring none got 11 conventional paths tried in a fixed order, first hit wins.
Does anyone still use RSS feeds?
Fourteen of 26 sites here maintain one, and six of those had published within 24 hours of the crawl. Whether humans subscribe is a different question, and this measurement says nothing about it.
Can an RSS feed replace my sitemap?
Google accepts RSS and Atom as sitemap formats but notes a feed carries recent items, so the two do different jobs. Use the feed for freshness and keep the sitemap for coverage.
Where should the feed link go?
In the <head> of the page the feed describes, with an explicit media type. Netlify puts three there with distinct titles, which is the clearest implementation in the panel.
