Discover’s share of publisher traffic from Google has nearly doubled in two years from 37% in 2023 to approximately 68% today. Traditional web search traffic to news publishers, meanwhile, has dropped from 51% to around 27% over the same period. Those two figures, published by NewzDash analysis of over 400 publishers, are what make the Google Discover core update of 2026 the most consequential algorithmic event for news publishers this year arguably more significant than any core search update Google has launched in recent memory. The update rolled out from February 5 to February 27, 2026. It is the first time Google has ever publicly designated a core update as targeting Discover specifically, and the first time the Discover feed algorithm has been formally decoupled from the broader web search ranking system. What changed, who was punished, and what you need to do right now are the three things this article covers.
Key numbers from the update
- Discover now drives 68% of publisher Google traffic up from 37% in 2023
- Unique domains in the US Discover top 1,000 dropped from 172 to 158 after the update an 8.1% tighter authority threshold
- The Independent lost 57% of its Discover traffic. The Sun lost 67%. Reuters dropped 20%
- California Discover feeds showed a 60% increase in locally sourced content post-update
- Articles with 1200px+ images and max-image-preview:large earn a 45% higher click-through rate in the Discover feed
What the Discover Core Update Actually Targets
Google confirmed three structural goals for the update through its Search Central Blog on February 5, 2026. First: showing users more locally relevant content from websites based in their own country, not international publishers writing for a foreign audience. Second: reducing sensational and clickbait content, with explicit updated guidance in Discover documentation about what constitutes prohibited headline tactics. Third: surfacing more in-depth, original, and timely content from publishers that have demonstrated expertise in a given subject area and here is the important detail evaluated on a topic-by-topic basis rather than at a domain level.
That third point carries significant strategic weight. Google’s systems now evaluate expertise per topic, not per domain. A local news site with a dedicated, consistently updated gardening section can have established gardening expertise in Discover’s view, even if it covers other topics. A general interest site that published one viral article about gardening will not. What this means in practice is that you don’t need to be a single-topic publication to win in Discover. But you do need to publish consistently and with genuine depth in every subject area where you want sustained placement. Sporadic coverage of trending topics to chase short-term traffic no longer earns lasting Discover visibility.
How Google Now Detects Clickbait The Headline-Content Alignment Classifier
The most technically significant change embedded in this update is a new headline-content alignment classifier deployed specifically within the Discover feed ranking system. It compares the implied promise of a headline against the actual substance of the article body. When the gap between the two exceeds a defined threshold, a ranking demotion is applied automatically no manual review required.
Google updated its official Discover documentation in conjunction with the update to define prohibited tactics with unusual specificity. The revised guidelines prohibit headlines and preview content that use “exaggerated or misleading details” to inflate appeal, content that “withholds crucial information required to understand what the content is about,” and any tactic that “manipulates appeal by catering to morbid curiosity, titillation, or outrage.” The language is more prescriptive than any previous Google quality guidance for Discover. More importantly, it arrived with a technical classifier rather than just an editorial guideline. This is algorithmic enforcement, not policy aspiration.
The specific content types being demoted are identifiable. Listicles that pad genuinely thin content with filler items. Articles using “shocking” or “you won’t believe” framing without delivering substantive content behind it. Stories that bury the main finding behind extensive preamble designed to maximise scroll depth rather than inform the reader. Previews that imply a revelation the article never actually delivers. If your editorial team has been writing for CTR optimisation rather than content satisfaction, this classifier is targeting exactly that approach and it will not be switched off.
Headline test apply before every publish
Read your headline. Read the first three paragraphs. Ask: does the article deliver what the headline promises without forcing the reader to scroll to the end to find out? If the answer is no if the headline implies a finding, a list, or an exclusive that requires extensive reading to reach the classifier is likely to flag it. Rewrite the headline to match what the article delivers, or restructure the content to front-load the promised substance.
Image Requirements Post-Update The 1200px Standard
Images have always mattered in Discover. The update has made them significantly more consequential, and the performance data is specific enough to act on immediately. Articles with images of at least 1200 pixels wide, combined with the max-image-preview:large meta tag in the page header, earn a 45% higher click-through rate in the Discover feed compared to articles with smaller or absent images. That differential is not marginal. It is the difference between a card that stops a scroll and one that gets passed.
Google’s official Discover optimisation documentation states that to enable large image previews you must either use a Creative Commons licence or set the robots meta tag to max-image-preview:large. Images should accurately represent the story, be taken from within the article content rather than decorative, and avoid generic stock imagery that could apply to any story on the same subject. Discover’s visual interface places the image at the centre of what a user encounters before they click. A generic stock photo of a laptop next to an article about AI regulation tells the algorithm nothing differentiating about your content. A specific, original image directly relevant to the story signals that the content is produced with editorial care.
Image fix three actions this week
Audit your 50 most recent articles and check all featured image dimensions flag anything below 1200px wide and replace it. Verify the max-image-preview:large meta tag is present in your site’s global page template. Then set minimum image dimension standards in your CMS publishing workflow so substandard images cannot be published to Discover-eligible content without a specific editorial override.
Local and Fresh Content The Clear Winners
The local relevance signal is the most measurable and most strategically actionable change from this update. Google’s systems now apply a stronger weighting to content from publishers based in the same country as the user viewing the feed a fundamental departure from Discover’s previous approach, which distributed content more globally regardless of publisher origin. NewzDash reported state-level feed personalisation in California and New York that was not previously present. California’s state-specific Discover feeds showed a 60% increase in locally sourced content in the post-update window. SFGate, Sacramento Bee, and EdSource began appearing more frequently in California state feeds while remaining absent from national top placements exactly the behaviour you’d expect from an algorithm weighting local origin over national reach.
For Indian publishers, this is the early signal to act on before the global rollout arrives. The update is currently US English-only, but the international expansion is confirmed. When it reaches India, publishers with domestic audiences, locally reported stories, and consistent regional coverage will be in the structurally advantaged position. Building that coverage depth now before the local relevance signal reaches your market means arriving at the rollout as an established authority rather than scrambling to build one after the fact.
Freshness compounds the local signal. Timely content stories published within hours of an event, analyses that follow breaking news quickly, content that addresses questions while they are actively being asked earns disproportionate Discover visibility under the updated algorithm. Quality alone is not sufficient; quality combined with timeliness is what the current system rewards. A well-researched article published three weeks after its subject became relevant will consistently underperform a timely, well-researched article on the same topic.
Which Publishers Gained and Which Lost
The post-update traffic data from NewzDash is specific enough to draw clear conclusions about what the algorithm is penalising and what it is rewarding.
On the losing side, international publishers writing for US audiences took the sharpest hits, driven primarily by the geographic relevance signal. The Independent dropped 57% in Discover traffic. The Sun fell 67%. Reuters was down 20%. The Guardian dropped 11%. These are not underperforming publications they are globally respected brands whose Discover visibility was penalised specifically because they are non-US publishers whose content was appearing heavily in US Discover feeds, a distribution pattern the update was designed to correct. Beyond geography, the publishers whose losses exceeded what geography alone explains tend to share specific content strategy characteristics: broad topical coverage without deep subject expertise, reliance on engagement-optimised headline writing, high article volume prioritised over content depth, and significant syndicated or aggregated content without original reporting.
On the winning side, US regional publishers with strong local coverage and concentrated topic authority gained meaningfully. Niche authority sites that maintain consistent, deep publishing in defined subject areas saw Discover visibility strengthen as competitor thin content was de-ranked around them. Sites with original reporting, first-hand expertise signals, and front-loaded content structure articles that answer the headline promise immediately rather than padding to page length maintained or grew their placements. DiscoverSnoop’s analysis also identified an unexpected distributional shift worth monitoring: X.com posts from institutional news accounts climbed from 3 to 13 items in the US Discover top 100 placements post-update. When Google surfaces a publisher’s tweet instead of their article, the publication retains brand visibility but loses site traffic entirely. This pattern has no precedent in Discover’s previous behaviour and deserves close attention as the global rollout expands.
“The February 2026 Discover update represents a maturation of Google’s content recommendation approach. The same quality signals that improve Discover visibility also enhance citation likelihood in AI-generated responses from ChatGPT, Perplexity, and Google’s own AI Overviews.” ALM Corp analysis, February 2026
How to Optimise Your News Site for Discover Post-Update
Recovery from a Discover traffic decline caused by this update requires content strategy changes, not technical SEO fixes. Google re-evaluates Discover eligibility on a rolling basis, which means substantive improvements to content quality can restore traffic within two to four weeks significantly faster than the months-long recovery timelines associated with core search updates. That speed of feedback is a genuine advantage. Use it.
The most important strategic shift is moving from domain-level authority thinking to topic-level authority building. Identify the three to five subjects where your publication genuinely produces better, more specific, more original content than competitors. Build an editorial calendar that maintains consistent, deep publishing in those areas as an ongoing commitment not just when a trending story provides a convenient hook. Google’s classifier evaluates the pattern of your coverage in a subject over time, not just individual articles. Fifteen well-researched pieces on a specific topic over six months builds the authority signal that one viral article on that topic will never generate.
Front-loading content structure is the second fundamental change. The update’s quality signals reward articles that deliver on their headline promise immediately. Restructure your articles so the primary finding, the key data point, or the main story development appears in the first 150 to 200 words not buried after three contextual paragraphs designed to build engagement time on page. This is not a dumbing-down of editorial standards. It is a respect for the reader’s time that simultaneously satisfies the headline-content alignment classifier. The deeper analysis, background context, and expert commentary can and should follow, but the substance must lead.
Systematic headline auditing, rather than reactive policy changes, is the third necessary adjustment. Pull your ten most-trafficked Discover articles from before the update and read each headline against the article body. Then pull your ten articles that lost the most Discover impressions post-update and do the same. The pattern of what the classifier is flagging in your specific content will become visible in that comparison. Use it to write a headline style guide with specific before-and-after examples not a generic “no clickbait” directive that different editors interpret differently in practice.
For Indian publishers specifically: The update is currently US English-only, but Google has confirmed a global rollout. When the local relevance signal reaches Indian market feeds, publishers that have been building consistent, locally reported, India-specific content will be positioned as established topical authorities. The brands that build Discover authority before that rollout arrives will pay far less to maintain it than those who scramble to build it after the fact. The preparation window is open now and it will not stay open indefinitely.
Post-Update Checklist for News Publishers
Complete this before your next editorial meeting
- Open Search Console and compare Discover performance for the two weeks before February 5 against the two weeks after February 27 these are your clean pre- and post-update windows
- Verify the max-image-preview:large meta tag is in your global page template not just on individual articles
- Audit your 20 highest-traffic Discover articles for featured image dimensions flag and replace anything below 1200px wide
- Run a headline audit on the articles that lost the most Discover impressions check each headline against the article body for promise-delivery alignment
- Define three to five subjects where your publication has genuine depth and build a six-month calendar that maintains consistent coverage in those areas
- Write a headline style guide with specific before-and-after rewrites not a vague “no clickbait” policy
- Review your lowest-performing Discover articles for content depth identify any under 600 words lacking original data, expert quotes, or first-hand reporting
- If you are a non-US publisher: accept the geographic correction in US feeds and redirect your Discover strategy toward your domestic audience, where the local relevance signal will strengthen your position when the global rollout arrives
- Monitor whether Google is surfacing your brand’s social posts instead of your articles in Discover if it is, prioritise improving article discoverability rather than expanding social output
- Set a 30-day performance review Discover re-evaluates on a rolling basis and substantive changes should produce measurable results within four weeks
What every publisher must understand about the Google Discover core update 2026
The Google Discover core update of 2026 is not a temporary volatility event that will self-correct as the algorithm settles. It is a permanent recalibration of how Google distributes content through the surface that now drives the majority of publisher traffic from Google properties. Geographic relevance will continue to reward publishers who have built for their domestic audience with locally specific, original reporting.
Topical authority will reward publishers who have committed to depth over breadth. And the headline-content alignment classifier will continue to penalise any publication that has built its Discover traffic on engagement-optimised framing rather than content satisfaction. The recovery timeline is fast two to four weeks if changes are substantive. The publishers gaining ground on your Discover placements right now are making those changes today. The gap between acting this week and acting next month is not neutral. It is measurable traffic that your competitors are taking in your absence.
