Google Discover's February 2026 Update: What the Data Actually Shows
The first data analysis of Google's February 2026 Discover update is in. Fewer publishers, more topics. Here's what changed and what content teams should do about it.
Google released a Discover-focused core update in early February 2026, the first update specifically targeting its Discover feed rather than Search. For a few weeks, publishers were guessing at what changed. Now the first proper data analysis is in.
NewzDash, a news SEO tracking platform, ran their DiscoverPulse panel across millions of US users, comparing two windows: the week before the update (January 25-31) and the week after (February 8-14). They analysed the top 1,000 domains, top 1,000 articles, and state-level feeds in California and New York.
The findings confirm what Google said it was targeting, with a few surprises along the way.
The Core Finding: More Topics, Fewer Publishers
Here is the number that stood out. Unique publishers getting Discover distribution actually shrank in the US (172 to 158 domains) and in California (187 to 177 domains). At the same time, the number of unique content categories being served grew: US went from 163 to 173 distinct categories, California from 162 to 170.
More topics. Fewer publishers getting them.
The pattern is clear: Discover is widening its topical scope while becoming more selective about which sites it trusts to cover each topic. If you are producing content across twenty categories without genuine depth in any of them, the direction this update is heading is not in your favour.
Google's own framing of the update backs this up. They described the goal as surfacing "more in-depth, original, timely content from sites with demonstrated topic expertise, evaluated topic-by-topic." Not best overall site. Best on a specific topic, evaluated topic by topic.
Local Content Got a Real Lift
This is the clearest signal in the data, and the most directly actionable.
In California, the number of locally relevant articles appearing in the top 100 Discover placements jumped from 10 to 16, a 60% increase. Publishers like SFGate, LA Times, Sacramento Bee, and ABC7 were appearing in California feeds but not the national feed at all during the same period. New York showed the same pattern: NY-local domain articles appeared roughly five times more often in the New York feed than in the California feed, and vice versa.
That is sub-national personalisation being applied at meaningful scale.
The international publisher story is the flip side. The Guardian lost 11% of its Discover visibility in the post-update window. Reuters dropped 20%. The Independent fell 57%. Non-US publishers serving US Discover feeds took a measurable hit.
This update is currently rolling out to English-language US users. Google has confirmed plans to expand it to all countries and languages in the months ahead. When it arrives here, the same logic applies: publishers writing genuinely local content, estate-specific, culturally grounded, regulation-aware, will have a structural advantage that global generalists cannot easily replicate. An article about HDB renovation permit timelines or CHAS clinic subsidies is not something a New York content farm can produce with any credibility.
Clickbait Got Named (And Punished)
Google also updated its "Get on Discover" documentation alongside the core update, and the language change is significant. The word "clickbait" now appears explicitly in the guidelines for the first time. Previously, the guidance said "avoid tactics to artificially inflate engagement" without naming the tactic. "Sensationalism tactics" is now named directly too, replacing vaguer phrasing about "morbid curiosity, titillation, or outrage."
Words matter when Google puts them in official documentation.
The NewzDash data found specific examples of the algorithm acting on this. Yahoo's presence in the US top 1,000 Discover results dropped from 11 to 6 articles, and Yahoo went from having multiple items in the top 100 to zero in the post-update window. Autoevolution, a site that had been running near-identical "dramatic reveal" headlines across its content, dropped from 5 articles in the top 1,000 to zero. A high-ranking Geediting listicle, the "psychology says..." format you've seen a thousand times, fell from roughly position 14 to position 153 in a single measurement window.
NewzDash's conclusion: engaging headlines are not penalised. Some brands with genuine topic authority still rank well with punchy framing. What appears penalised is templated curiosity-gap patterns at scale, the same formula applied across dozens of articles by a site with no real expertise underneath.
This is a relevant warning for content teams using AI tools to produce Discover-targeted articles in volume. If the output follows a recognisable template and lacks original reporting or genuine specialist insight, the February update is a signal worth heeding.
X.com Is Now a Discover Surface
X.com posts from institutional accounts went from 3 items to 13 in the US top 100 Discover placements between the pre- and post-update windows. In New York, the jump was from 2 to 14.
NewzDash had been tracking X.com's growing Discover presence since November 2025. The February update appears to have accelerated the trend. Most top-performing X posts in Discover came from established media brands: NYT, news outlets, political figures with large followings. Google appears to be treating posts from authoritative accounts as timely, original content, exactly what the update says it wants to surface more of.
There is a complication worth flagging. NewzDash found a directional data point suggesting NYT appeared less often as a direct domain in the post-update window, while an NYT X post took the number-one Discover slot. That pattern is consistent with some traffic routing through X rather than directly to the publisher's site, an extra click step that will reduce some conversion for the publisher, even if the story still "wins" distribution.
For brands considering X as a Discover channel: the distribution opportunity is real, but if your goal is website traffic, a funnel of Discover to X to your site is leakier than Discover directly to your site. Worth measuring before committing to the strategy.
Page Experience Is Now in Discover Guidelines
One more documentation change worth noting: Google added "provide an overall great page experience" to the Discover guidelines, with a link to its Page Experience documentation. That recommendation was not in the previous version.
Page experience, covering Core Web Vitals, mobile usability, HTTPS, and no intrusive interstitials, has been part of Search guidance since 2020. Adding it to Discover-specific guidance now signals that a slow, poorly loading page carries a Discover penalty as well as a Search one. If your site has never been through a Core Web Vitals audit, this is a reasonable time to do one.
What to Actually Do
The update is US English only for now, but the expansion is announced and the strategic direction is clear.
Build depth on fewer topics. The consolidation pattern rewards sites with genuine topic authority. Pick the areas where you have real expertise and produce more thorough coverage of those areas. A local F&B business that writes substantive articles about hawker culture, food licensing, and kitchen operations will hold up better in Discover than a generic SME blog that touches twenty topics at 400 words each.
Lean into what only you can write. Local content surged in the post-update data because local publishers have a permanent advantage in their own geography. Write about Singapore-specific regulations, schemes, market conditions, and cultural context: not as a keyword play, but because it signals genuine local relevance that overseas content teams cannot credibly produce. When the update hits this market, that content already being indexed and trusted will be an advantage.
Review your title patterns. Not to make headlines bland, but to check whether your content library uses the same recognisable template across many articles. The sites that lost most visibly were running one formula at scale. Variation and genuine substance underneath each headline are what the algorithm now appears to distinguish.
Check your X presence as a Discover signal. If you have an institutional brand account on X posting timely, substantive updates, you may be building Discover surface area you are not currently measuring. Pull your Search Console Discover report and look at what is appearing there. You might find X-sourced content driving referrals alongside your direct site content.
Fix slow pages. Now that page experience is explicitly in Discover guidelines, a Core Web Vitals audit is no longer just an SEO housekeeping item. If your Discover traffic is underperforming relative to your content quality, page speed is a plausible culprit worth ruling out.
The direction Google is heading in Discover, local expertise, topical depth, no templated shortcuts, is the same content strategy that holds up across algorithm changes rather than chasing each one. If you are publishing useful, specific content on topics you genuinely know, and doing it consistently, this update is a tailwind.
Magnified works with SMEs to build content strategies that earn distribution over time, not just optimise for the current playbook. Talk to us about your content plan.