Niblah

Verified
SEO

Google Discover optimization in 2026 (practical playbook)

Albin Hot
By Albin Hot
February 12, 20267 min read

Google Discover is not traditional SEO.

Nobody types a query. People scroll. Google predicts what they’ll care about next.

That changes everything: you don’t “rank” in Discover with keywords — you earn distribution with *relevance, freshness, trust and packaging*.

In 2026, Discover can still be one of the fastest ways to get big spikes in traffic, but it’s also volatile. The goal isn’t to chase spikes. The goal is to build a site that Discover *wants* to recommend, consistently.

What is Google Discover (and where does it show)?

Google Discover is a personalised feed inside:

  • the **Google app** (mobile)
  • **Chrome on mobile** (new tab / feed surfaces)
  • It shows articles and videos based on a user’s interests, behaviour, location context and engagement history.

    Key difference vs Search:

  • **Search** = user asks a question.
  • **Discover** = Google *suggests* content the user didn’t ask for (yet).
  • That means Discover is closer to a social feed than a search result page.

    How Discover “chooses” content (a simple mental model)

    Discover is best understood as 4 filters:

    1) Eligibility: can Google safely recommend this page?

    Before anything else, your content must be:

  • **indexable** (not blocked by robots / noindex)
  • **mobile friendly**
  • **fast and stable**
  • compliant with Google’s content policies (misleading / unsafe / spammy content gets filtered out)
  • 2) Interest fit: does this match a user profile?

    Discover is built around *topics and entities*.

    If your site has a track record around a clear theme (e.g. SEO experiments, Shopify growth, link building), Google is more likely to test your new posts with people who already engage with those topics.

    3) Timing: is this relevant *right now*?

    Discover is extremely sensitive to:

  • trends (seasonality, industry news)
  • “new angle” updates
  • follow-up pieces to hot topics
  • Evergreen can work — but evergreen that is **freshly updated** or tied to a timely hook usually performs better.

    4) Performance signals: do people actually tap and engage?

    Once Google tests your article, Discover looks at behavioural signals such as:

  • **CTR** (headline + image)
  • quick back / pogo behaviour
  • engagement (scroll depth, time, repeat exposure)
  • If it performs, distribution expands. If it doesn’t, it dies quickly.

    The biggest lever in 2026: topical authority (not volume)

    Discover rewards *focus*.

    A site that publishes 3 strong articles per week in one tight theme often outperforms a site that publishes 15 mixed posts across unrelated topics.

    Action steps:

  • Pick **3–5 core topics** your site will be known for.
  • Build **clusters** (a pillar guide + supporting posts that internally link back).
  • Create recurring formats (e.g. “SEO case study”, “Tool teardown”, “Checklist”).
  • If you’re currently publishing broad “everything marketing” content, consider trimming your scope.

    Write for scroll: headlines that earn taps without clickbait

    Discover users don’t have search intent in hand — your headline must create it.

    Rules of thumb:

  • Be **specific** (avoid generic titles like “SEO tips”)
  • Lead with **a clear outcome** (what will the reader get?)
  • Add a **time context** only when it’s real (don’t slap “2026” on everything)
  • No bait-and-switch (clickbait kills long-term Discover performance)
  • Headline formulas that work

  • **Outcome + mechanism**
  • - “Google Discover traffic: the 12 signals you can control in 2026”

  • **Mistakes + prevention**
  • - “7 Discover mistakes that quietly suppress your reach”

  • **Framework + proof**
  • - “The Discover checklist we use to launch posts that spike”

    Headline mistakes that fail

  • “You won’t believe what Google did…”
  • “This ONE trick…”
  • Titles that promise news but deliver a generic guide
  • Images & thumbnails: the real CTR multiplier

    In Discover, your thumbnail often matters more than your text.

    Minimum requirements (practical)

  • Use **original, high-quality images** where possible
  • Aim for **large images** (1200px+ width is a safe target)
  • Prefer **16:9** compositions that still look good when cropped
  • Avoid heavy text overlays or stock-photo vibes
  • Make your imagery consistent (brand without spam)

  • Use a consistent colour style or framing
  • Keep brand elements subtle
  • Match the image to the story angle (misleading images reduce trust)
  • Technical checklist for images

  • Ensure Google is allowed to show large previews (e.g. max-image-preview:large)
  • Provide proper Open Graph / Twitter card meta tags
  • Don’t block critical images via robots or auth
  • Technical foundation: make Discover-friendly pages boringly solid

    Discover is not “technical SEO”, but the same fundamentals apply.

    1) Indexing & canonical hygiene

    Confirm that:

  • the URL is **indexable**
  • canonical points to the correct version
  • you don’t accidentally publish duplicates across parameters
  • 2) Speed and mobile UX

    You don’t need AMP. You do need:

  • fast initial load (LCP)
  • no layout jumps (CLS)
  • readable typography on mobile
  • no intrusive popups that block content
  • 3) Structured data (helpful, not magic)

    Structured data won’t “force” Discover, but it helps Google understand your page.

    Add Article/NewsArticle/BlogPosting JSON-LD with:

  • headline
  • author
  • datePublished / dateModified
  • image
  • publisher
  • Example (adapt to your CMS):

    {
      "@context": "https://schema.org",
      "@type": "BlogPosting",
      "headline": "Google Discover optimization in 2026: a practical playbook",
      "author": {
        "@type": "Person",
        "name": "Albin Hot"
      },
      "datePublished": "2026-02-12",
      "dateModified": "2026-02-12",
      "image": [
        "https://www.niblah.com/og/google-discover-2026.png"
      ],
      "publisher": {
        "@type": "Organization",
        "name": "Niblah",
        "logo": {
          "@type": "ImageObject",
          "url": "https://www.niblah.com/logo.png"
        }
      }
    }

    E‑E‑A‑T for Discover: practical trust signals

    Discover leans heavily into trust.

    If your site looks anonymous, thin or “made for clicks”, your ceiling is low.

    Concrete upgrades:

  • **Real author pages** (bio, credentials, links)
  • a clear **About** page and **Contact** page
  • editorial standards (how you research, update, correct)
  • cite sources when you make claims
  • show real examples (screenshots, experiments, data)
  • Tip: if you publish opinion pieces, make the author’s experience explicit (“I ran this test on 30 pages…”).

    Content formats that perform well in Discover

    Discover rewards content that feels *native to a feed*:

  • strong opinions (backed by experience)
  • original data / mini studies
  • practical checklists
  • “what changed” updates
  • visual explainers (graphics, short videos)
  • Avoid:

  • generic rewritten content
  • thin affiliate pages
  • content that is obviously produced at scale without depth
  • How often should you publish?

    Consistency matters more than frequency.

    Pick a cadence you can sustain:

  • 2–3 high-quality posts per week is enough for most sites
  • Then build momentum with:

  • series (part 1/2/3)
  • scheduled updates to evergreen posts
  • rapid follow-ups when a topic trends
  • Monitoring: how to learn from Discover spikes

    Discover traffic is spiky by nature. Your job is to turn spikes into patterns.

    In Google Search Console:

  • open **Discover** (Performance)
  • review by **page**
  • compare winners vs non-winners:
  • - title structure

    - image quality

    - topic freshness

    - author trust

    What to test (safely)

  • change the hero image on underperforming posts
  • tighten the intro (first 2–3 sentences)
  • update headline to be more specific
  • Do not:

  • rewrite the entire article daily
  • constantly change the URL
  • Troubleshooting: why you might not show up

    Common reasons:

  • your site is too broad (no clear topical identity)
  • your images are too small / generic
  • the content is not indexable
  • your page looks untrustworthy (no author, no brand, no proof)
  • you publish too irregularly to build momentum
  • Copy/paste checklist (2026)

    Before you publish:

  • [ ] Topic fits your site’s core authority
  • [ ] Headline is specific and not clickbait
  • [ ] Large, high-quality image (1200px+)
  • [ ] Page is indexable + correct canonical
  • [ ] Strong author signals (bio, About, Contact)
  • [ ] Fast mobile experience
  • [ ] Article schema + OG tags
  • After you publish:

  • [ ] Check Discover report in GSC (when available)
  • [ ] Compare CTR vs your own benchmarks
  • [ ] Iterate headline/image on posts that almost hit
  • Summary

    To win in Google Discover in 2026, think like a publisher.

  • Build a narrow topical identity.
  • Package content for scroll (headline + image).
  • Keep the technical foundation clean.
  • Prove trust with real authorship and real evidence.
  • Measure winners and repeat the pattern.
  • If you want help building a Discover-ready content machine, reach out via [contact](/contact).

    Albin Hot

    Albin Hot

    Albin Hot is Senior SEO Specialist bij Niblah, een toonaangevend marketing platform voor zoekmachines, AI en meer. Hij werkt al meer dan 5 jaar in SEO en specialiseert zich in omzetgedreven strategieën in nauwe samenwerking met multidisciplinaire teams.

    Connect op LinkedIn