Google Discover 2027 Playbook: What Headlines, Aspect Ratios, and E-E-A-T Signals Are Driving the Highest CTRs

Nearly nine months after Google’s February 2026 Discover-focused core update, the feed has settled into a clearer set of rules. The update, which rolled out from 5–27 February 2026 (initially for English-language users in the United States, then more broadly), explicitly targeted better relevance and quality: more locally relevant content, less sensationalism and clickbait, and stronger preference for original material from sites demonstrating topic expertise.

For publishers and content teams operating in Arabic and English feeds, the practical question is no longer “Will we appear?” but “What combinations of headline, image, and trust signals are still earning the highest click-through rates after the quality filter tightened?”

The Current Performance Reality

Discover remains a major distribution channel even as overall Google referral traffic to many publishers has declined. Chartbeat’s 2026 Publisher Playbook data showed Google Discover referrals falling 34.3% year-over-year from July 2025 to July 2026 (widening from a 6.6% decline the prior year). At the same time, Discover has often represented a larger share of Google-driven pageviews than classic Search for news and content sites in the network.

Typical article-level CTR in Discover sits in the 4–6% range. High-performing pieces regularly reach 8–12% or higher when the headline and image work together effectively. Because Discover is almost entirely mobile and interest-driven (users are not searching), the card itself—image plus headline—does nearly all of the work.

Traffic is also more volatile than Search. Individual articles often peak within 24–72 hours and then drop sharply, partly because of Google’s deployed staleness-filtering systems that remove content once it has delivered most of its predicted audience or when fresher coverage appears.

Headlines That Are Winning Right Now

Google’s documentation is unambiguous: use titles and headlines that “capture the essence of the content.” It specifically warns against clickbait, misleading or exaggerated details in previews, and sensationalism that manipulates curiosity, titillation, or outrage. The February 2026 update accelerated demotion of the older curiosity-gap style.

What still performs:

• Specific, concrete claims or numbers (“What 3.4 million Discover headlines actually showed about quote-led titles”)
• Clear benefit or insight without over-promising
• Timely framing when the story is genuinely current
• Accurate representation of what the reader will find on the page

Optimal length is commonly observed in the 40–70 character range so the full headline remains readable on mobile cards without truncation. Longer headlines get cut; very short ones fail to create enough pull.

High-CTR patterns observed across English feeds after the update tend to be direct rather than teaser-driven: data-backed observations, clear “what changed and why it matters” framing, and specific outcomes. Pure “you won’t believe” or “this one trick” styles have lost ground. The same principles apply in Arabic feeds—native phrasing that matches how readers actually speak and search outperforms stiff translations or generic MSA when the audience is primarily consuming colloquial content.

Aspect Ratios and Image Performance

Images carry disproportionate weight in Discover. Google’s official recommendations (current documentation) remain:

• At least 1,200 pixels wide
• More than 300,000 total pixels
• 16:9 aspect ratio
• Large previews enabled via max-image-preview:large (or AMP)
• Specified via og:image or schema.org markup
• Relevant, high-quality, and not a logo or heavily text-based graphic

A 1280×720 image (921,600 pixels) clears the threshold comfortably. Many high-performing publishers go larger (1600×900 or 1920×1080) for sharper rendering on high-density mobile screens while keeping the 16:9 ratio.

In practice, cards with large, relevant 16:9 images consistently outperform those with smaller or poorly cropped visuals. For Arabic content, the same technical rules apply, but composition matters more: keep key subjects and any on-image text (used sparingly) readable in right-to-left scanning patterns, and avoid Western visual clichés that feel mismatched to Gulf, Levantine, or North African audiences.

Avoid text-heavy graphics in the primary image slot—Google explicitly discourages them because the headline already appears below the image. Original photography or clean, high-quality illustrations of the actual subject outperform stock or generic visuals.

E-E-A-T Signals That Are Currently Rewarded

The February 2026 update visibly favored sites with stronger experience, expertise, authoritativeness, and trustworthiness signals, especially when those signals were topic-specific rather than site-wide. Google has long linked Discover eligibility to the same helpful, reliable, people-first standards used in Search.

Observable winning patterns include:

• Clear bylines with real author names and linked bios or credentials
• Visible publication and update dates
• Inline citations or links to primary sources
• Transparent “About” and author pages
• Original reporting, data, or firsthand experience rather than rewritten aggregation

Transparency requirements are particularly explicit for news content: clear dates, bylines, information about authors and the publisher, and contact details. After the update, sites that strengthened these signals generally retained or gained visibility while thinner or more sensational content lost ground. In Arabic-language verticals (finance, health, real estate, consumer products), local expertise and culturally accurate sourcing appear to carry additional weight because of the update’s emphasis on locally relevant content.

Arabic and English Feeds: What Differs

The core ranking and display rules are language-agnostic, but practical performance differs:

• Headlines: English feeds respond well to concise, data-led or insight-led phrasing. Arabic feeds often perform better with natural dialectal or lightly localized phrasing (Khaleeji, Egyptian, Levantine, etc.) that matches the primary audience, rather than pure Modern Standard Arabic or direct translation. Test both.
• Images: The 16:9, 1200 px+ specification is identical. Visual cultural codes (clothing, settings, models, color preferences) affect engagement; a high-quality image that feels locally relevant consistently lifts CTR.
• Local relevance: The February update’s push for more locally relevant content means Arabic-language publishers based in or focused on specific MENA markets have an advantage in those feeds compared with purely translated global content.
• Measurement: Use the Discover Performance report in Google Search Console (impressions, clicks, CTR by page, country, and date). Segment by language or country where possible. Note that a separate generative AI features report currently shows impressions only, not clicks.

Practical 2027 Checklist

1. Every article has a relevant 16:9 image ≥1200 px wide with max-image-preview:large and proper og:image/schema markup.
2. Headline accurately captures the content’s essence in 40–70 characters, specific rather than sensational.
3. Author, date, and sourcing signals are visible on the page.
4. Content is timely or offers unique insight; update evergreen pieces when they regain topicality.
5. Page experience is strong on mobile (Core Web Vitals, readable layout).
6. Track CTR in the Discover report; aim for ≥4–6% as a baseline and investigate any consistent under-performers.
7. For bilingual sites, produce and test language-specific headlines and images rather than relying on translation alone.

Discover in late 2026 and heading into 2027 rewards the same fundamentals the February update reinforced: accurate, specific headlines; large, relevant 16:9 images; and visible E-E-A-T. The volume of traffic may be more constrained and more competitive than in earlier years, but the cards that still earn the highest CTRs in both Arabic and English feeds are those that respect the reader’s time and the platform’s quality bar.