Polymarket Outranked Reuters in Google News Because Bet Pages Look Like News
Polymarket's prediction-market pages surfaced in Google News' For You feed, home page, and Top Stories results between January and April 2026, sometimes ranking above Reuters and the Financial Times on stories like Strait of Hormuz transit. Google called the placement "in error" and stopped surfacing the site in News. The eligibility mechanism that let it through has not changed.
The reporting is now thoroughly documented. Futurism's investigation in early April caught a "Bitcoin Up or Down – 5 Minutes" market sitting at the top of a Google News For You feed, with markets like "Will ChatGPT Be the Number One Free App in the US Apple Store by April 10?" appearing as headline-shaped content blocks. Google spokesperson Ned Adriance told Gizmodo the site "briefly appeared in Google News in error." The complaints, per PPC Land's deeper write-up, started in late March and trace back as far as January.
The "error" framing is doing a lot of work. The ranker behaved exactly the way the documentation says it should.
Bet pages have the structural fingerprint of news
Polymarket markets hit every signal Google's news ranker checks for. Frequent updates. News-style headlines. Fresh timestamps. Topical relevance to whatever is moving on Twitter that hour. From an indexing-system perspective, "Will ChatGPT Be the Number One Free App in the US Apple Store by April 10?" is structurally indistinguishable from a Reuters headline about the same event.
The current Google News documentation is explicit. There is no application gate. Google considers all web content for inclusion automatically, Publisher Center is optional, and eligibility is decided by the ranker, which weighs freshness, frequency, topical relevance, and structural markers. A Polymarket market page hits four out of four.
Add the news sitemap rule that articles older than two days should be removed from the news sitemap or stripped of news metadata, and you have a system that explicitly rewards content with constantly-shifting URLs. Polymarket spins up new bets every few minutes. Calling that an algorithmic accident misses what actually happened: the ranker doing its job on inputs nobody bothered to filter.
The "automatic eligibility" model is the real story here
This only became possible after Google quietly retired the manual approval process for Google News. The old Publisher Center workflow involved an editorial review step. The current system replaces that with automatically-generated publication pages and algorithmic eligibility for any site that meets content policies.
That is great if you run a small news site that used to wait three months for approval. It is less great when your competitor is a betting platform that publishes tens of thousands of headline-shaped pages a day. The same reasoning that lowered the bar for legitimate independent newsrooms also lowered it for anyone who can mass-produce news-shaped URLs.
I think most SEO teams have not internalized this shift. There is still a folk belief that Google News inclusion is a vetted privilege, like a press credential. It is not anymore, at least not in the technical sense that matters to the ranker. The bar is structural, not editorial, and Polymarket is the proof of concept.
We covered an adjacent failure mode last week, where Google's AI Mode rewrote publisher titles wholesale until Lily Ray flagged it. Same pattern. Editorial trust is no longer encoded into the surface. What is encoded is structure.
What this exposes about Top Stories' real bar
A few practical notes from sitting with this for an hour.
The News structured data spec does not require NewsArticle markup to be eligible for Top Stories, but it is the cleanest signal you can send. If your news content does not declare its Article subtype, you are competing with sites that do. We publish NewsArticle plus BreadcrumbList JSON-LD on every news post for this reason. Your CMS may not, and the default WordPress installs I have looked at recently still do not.
News sitemap hygiene matters more than it used to. Articles past the two-day window should drop out of the news sitemap automatically. If yours does not, the sitemap reads as stale, and the ranker quietly demotes your fresh URLs. Audit it this week.
Internal linking from the homepage and section pages is part of the eligibility signal. The Google News docs reference articles being linked to from "a prominent position." A buried article in week two is not a Top Stories candidate, even when every other signal is right.
And do not assume the Polymarket fix has tightened anything else. Google removed Polymarket from News explicitly. The underlying ranker did not change. From what I have seen, anyone with a network of fast-updating, headline-shaped pages can run the same play under a different domain by next quarter. The countermeasure, if there is one, will probably be a domain-level allowlist for borderline categories like prediction markets and price tickers, which means more editorial gatekeeping returning quietly through the back door.
The accuracy claim is also weaker than it looks
A small detour worth taking before the action item, because it changes the framing.
A Vanderbilt University study by Joshua Clinton and TzuFeng Huang analyzed roughly 2,500 prediction markets covering $2.5 billion in volume during the final five weeks of the 2024 US election. Polymarket scored 67% accuracy. Kalshi hit 78%. PredictIt, with its $850-per-contract cap, hit 93%. The narrative that prediction markets are an oracle layer for journalism has been quietly downgraded in the academic literature.
Polymarket also concentrates profit dramatically. Per platform data cited by PPC Land, only 0.04% of Polymarket traders captured nearly 70% of profits. Calling the price signal there "wisdom of crowds" is a stretch. A small group of insiders set most of those prices, and Google News then surfaced them as if they were reporting.
If Google's news ranker carried any editorial weight on source credibility, none of this would matter. The fact that it did not is the part I would chew on if I were on a publisher SEO team this week.
The 15-minute audit before someone parks in your news slot
If you have any news content on your site, here is what I would actually do before the weekend.
- Pull your news sitemap. Confirm that articles older than 48 hours have either been removed or had their
<news:news>metadata stripped. Most CMS plugins handle this. Some do not, and the failure mode is silent. - Spot-check your top 20 news URLs in Search Console under Performance, then Search Appearance. If they are not appearing in the News appearance bucket, they are not eligible. Missing NewsArticle structured data is the most likely cause.
- Count the news articles linked from your homepage and category pages. If it is under ten, you are sending a "this is not a news site" signal even when every individual article is correct.
- Run a search for your category's top trending topic and screenshot what is in the Top Stories carousel. If a non-news source is sitting in slot one or two, you have just found a competitor gaming the same signals Polymarket gamed.
The slot is decided by signals, not by whether you "deserve" to be there. From what I have seen, that calculation is not getting reversed any time soon. Probably the most useful response is to assume the carousel will keep getting weirder, and to make sure your structural signals are clean enough that you survive the weirdness.
Notice Me Senpai Editorial
