Unveiling Google Discover: Impact on Publishers & 42M Cards
Insights into Discover Feeds: Unveiling Google's Internal Architecture
Metehan Yesilyurt's comprehensive SDK analysis shed light on the pipeline names within the Google Discover platform. By examining months of real Discover feeds from hundreds of devices, we delved into the essence of each pipeline — assessing volume, reach, timing, and the dominant publishers. Let's dive into what 42 million feed cards have unveiled about the internal workings of Discover.
Our Methodology
Over a span of three months from December 2025 to February 2026, we meticulously scrutinized genuine Discover feeds from a diverse array of devices. The outcome of this extensive analysis was the examination of 42 million feed cards. Each card was intricately linked to the specific pipeline responsible for its selection.
While some pipeline names were previously identifiable from the SDK, our data brought a new dimension by elucidating the practical function of each pipeline. It unveiled the quantity of content selected, the devices it reached, its operational speed, and the publishers it favored — providing a comprehensive view of Discover's inner mechanisms.
For every pipeline, we calculated four key metrics:
- Reach — the percentage of devices viewing each URL per day
- Speed — the median age of articles when they appear
- Exclusivity — the proportion of URLs unique to that particular pipeline
- Volume — the share of the total feed attributed to each pipeline
For a visual exploration of all 20 pipelines, check out the interactive explorer: Open the interactive explorer →
The Complexity of Discover: A Multilayered Approach
Contrary to the common belief that Discover relies on a singular recommendation algorithm, our data portrays a sophisticated system comprising six distinctive functional layers. Each layer operates with unique logic, speed, and target audience, collectively shaping the dynamic landscape of Discover.
The six layers encompass:
- Core editorial — content, moonstone, aura, paginationpanoptic, and relatedcontentruby, each contributing distinct attributes to the content selection process.
- News urgency — featuring mustntmiss and newsstoriesheadlines with a special focus on timely and impactful news content.
- Trends — delving into the realm of trending topics with pipelines like deeptrendsfable and deeptrends driving the trend detection efforts.
- Local/geo — emphasizing geographically targeted stories and events through pipelines like geotargetingstories and webkicklocalstories.
- Social/video — highlighting the influence of social media and video content, especially through the YouTube-centric pipelines like creatorcontent and neoncluster.
- Commercial — focusing on commercial aspects with pipelines such as shoppinginspiration and feedads, maximizing reach through promotional content.
The Evolving Trends in Discover Feed (2026 Insights)
The YouTube Amplification Cascade
A standout feature of the English Discover feed, the YouTube cascade comprises three interconnected pipelines that amplify content at each stage:
| Stage | Pipeline | Content Mix | Reach | Timing |
|---|---|---|---|---|
| 1. Intake | creatorcontent | 72% YouTube, 23% x.com | 6.7% | T₀ |
| 2. Filter | freshvideos | 94% YouTube | 7.1% | T₀ + 15h |
| 3. Broadcast | neoncluster | 100% YouTube | 13.0% | T₀ + 23h |
The video cascade intensifies with each stage, narrowing down to pure video content and expanding its reach progressively. This evolution has propelled the YouTube-centric content to resonate with a larger audience, marking a significant growth trajectory in the Discover landscape.
In contrast, the French Discover feed lacks this distinct cascade, showcasing a unique trend in content presentation and distribution.
AI Overviews Integration in Discover Feed (Exclusive to English)
The incorporation of AI-generated summary cards, termed AI Overviews, is a recent addition to the Discover platform, exclusive to the English feed.
- discover_ai_summary: Engaging 1.1% of EN volume, predominantly featuring AI Overview content from notable sources like Reuters, New York Times, and CNBC.
- mustntmiss: Showcasing a substantial 29% of AI Overviews content, marking a significant presence in the content selection algorithm.
- paginationpanoptic: Enriching the scroll infrastructure with 7.8% AI Overviews content, demonstrating a comprehensive integration of AI summaries into the feed.
- In French: The AI Overview integration is notably sparse, depicting a language-specific variance in content aggregation.
Exploring Feedads: A Pivotal Pipeline in English Discover
feedads emerges as a predominant force within the Discover ecosystem, captivating a remarkable 58.4% of English devices with its targeted advertisements. Notably, video advertising, especially via YouTube, constitutes a significant portion of these ads, with campaigns spanning a median of 57 days. The exclusivity of URLs within this pipeline is exceptionally high, indicating a distinctive advertising strategy for maximum impact.
Comparatively, the editorial pipeline with the highest reach in English Discover, neoncluster, reaches 13% of devices, illustrating the influential role of feedads in intensifying the platform's monetization efforts relative to other content-driven pipelines.
The Enigmatic EPL Exclusion
A surprising revelation from our analysis is the notable under-representation of Premier League content across more than seven English pipelines within Discover. Terms linked to the Premier League exhibit consistent marginalization in pipelines like aura, deeptrendsfable, geotargetingstories, and others.
This exclusion does not extend to other sports entities such as NFL, NBA, Olympics, rugby, and Formula 1, suggesting a specific bias against EPL content. The underlying reasons for this systematic exclusion remain speculative, possibly attributed to broadcasting rights and licensing constraints that sculpt editorial decisions within the content selection framework.
Insightful Publisher Profiles
Quality Press Representation (Guardian, BBC, New York Times)
The quality press presence in Discover spans across multiple pipelines, with Guardian exhibiting a diverse profile across 12 pipelines, showcasing expertise in various content genres. BBC occupies a significant presence in pipelines like astria and deeptrends. The incorporation of AI Overviews content has become a distinguishing feature, providing a competitive edge for quality press entities within the English feed.
Tech and Review Site Influence (TechRadar, Tom's Hardware)
Exploring the tech and review domain through shoppinginspiration reveals a robust reach and lifespan within the Discover environment. However, there's a notable trend of channeling tech content into specific silos, with an opportunity for broader editorial integration to tap into additional pipelines, thereby diversifying content exposure.
Video-Centric Publishers (YouTube Channels)
For video-first publishers, the prominence of YouTube channels in the Discover feed signifies a potent distribution channel. The amplification cascade ensures substantial reach for content genres like news, politics, science, and current affairs, offering an advantageous organic distribution platform that continues to expand within the English Discover landscape.
Stay tuned for detailed recommendations on various publisher categories including quality press, tech, video, local, lifestyle, finance, and pure players in our upcoming Substack series.
Delve Deeper into Discover
This article provides a condensed overview of the intricate workings within the Google Discover ecosystem. For an in-depth analysis encompassing all 20 pipelines, domain leaders, and typical content titles, explore:
- Interactive Explorer: Navigate through all 20 pipelines, compare metrics, and discover top domains and content titles
- EN Substack: Engage with weekly deep-dives with comprehensive data, charts, and tailored recommendations
- Full Reference: Visit 1492.vision for an in-depth pipeline-by-pipeline analysis with detailed articles.
The dynamic nature of the Discover system emphasizes the importance of continuous monitoring to capture evolving trends and opportunities. Our insights span from December 2025 to February 2026, reflecting the emergence and growth of trends like the video cascade and AI Overviews within the Discover ecosystem.
Data Source: 42 million Discover cards, December 2025 – February 2026. Analysis conducted by 1492.vision. Special thanks to Metehan Yesilyurt for the perceptive Google SDK analysis, enhancing our understanding of Discover's operational dynamics.
Explore over 200 online tools in areas such as SEO, development, text, image, PDF, CSV, and conversion/calculator tools on MATSEOTOOLS. Enhance your digital ventures with precision and efficiency!
Some Question