Google Discover SEO in 2026, 11 fixes for publishers (headlines, images, timing, and topic focus)

Google Discover can feel like a fickle newsagent. One day your story sits in the front window, the next it’s gone. That’s normal. Still, google discover seo isn’t guesswork in 2026, because patterns show up fast when you track headlines, images, cadence, and topic choices.

In February 2026, many publishers saw Discover volatility after a Discover-only core update. Some sites gained steady reach, while others lost it overnight. The difference often came down to the basics: clear titles, honest visuals, consistent publishing, and depth on a small set of themes.

This article gives 11 fixes you can apply this week, plus a method to diagnose drops without jumping at shadows.

What changed in Google Discover in February 2026 (and what stayed the same)

The February 2026 Discover Core Update began rolling out on 5 Feb 2026, and it targeted Discover feeds rather than traditional Google Search. Several industry round-ups reported stronger emphasis on local relevance, expertise by topic, and less tolerance for sensational framing, for example a practical summary of the February 2026 update.

Two things are worth holding in your head at the same time:

First, Discover is still personalised. It doesn’t rank pages by a typed query. It predicts what a person might want next, based on interests, recent behaviour, and location signals. So your best article can underperform if it doesn’t match the user’s “now”.

Second, Discover remains picky about presentation. Your headline and hero image are your shopfront. If either feels misleading, the click might happen once, but trust drops quickly after that. Over time, that costs distribution.

Google also continues to stress “increase the likelihood” rather than promise outcomes, in its “Get on Discover” guidance. That tone matters, because Discover isn’t a tap you can turn on. You can, however, make your work easier to understand, easier to trust, and easier to match to an audience.

Treat Discover like a front page: one clear promise, backed by the first paragraph and the image.

11 fixes publishers can apply for Google Discover SEO (headlines, images, timing, focus)

Below are 11 fixes, grouped by what usually moves the needle fastest.

1. Write “what it is” headlines before “why it matters”
   Lead with the entity and action. Save the opinion for the subhead or dek. Clarity beats cleverness when users scroll quickly.

2. Reduce ambiguity, keep one main idea
   If the headline can mean two things, Discover often under-delivers. Remove stray clauses, stacked jokes, and vague pronouns like “this” or “it”.

3. Use entity-first phrasing to help relevance
   Put the recognisable subject up front (company, person, location, product, event). It makes matching to interests easier, and it reads cleaner.

4. A/B test headlines without breaking trust
   Test variations that change angle, not facts. Try:
   Version A: “What happened”
   Version B: “What changed for readers”
   Keep the same promise as the opening paragraph, or you’ll train poor engagement.

Here are quick rewrites you can use as a pattern:

5. Use large, original images, not “stock-looking” fillers
   Discover favours strong visuals, but “strong” means believable. Use real photos, charts you built, or event images with clear context. Google’s guidance has long pointed publishers towards large images (at least 1,200 px wide) and correct preview settings (for example max-image-preview:large).

6. Choose crops that render well in mobile cards
   Practical crops that usually hold up: 1.91:1 for wide cards, and 4:3 for safer framing. Keep the subject centred, avoid tiny text overlays, and check how it looks on a small screen.

Example image do and don’t (descriptions you can hand to editors):

• Do: “Real photo of the CEO on stage, logo visible, neutral lighting, matches the story.”
• Don’t: “Generic handshake stock photo for a layoffs story.”
• Do: “Original chart showing the change, with a clean legend and large labels.”
• Don’t: “Meme-style image that implies a claim your article doesn’t support.”

7. Make your image selection consistent across metadata
   If your page, Open Graph image, and structured data point to different images, platforms can pick the wrong one. Use one primary hero consistently, and verify with social preview tools. Also, avoid mismatched images that hint at a different story.

8. Match cadence to content type, don’t “burst then silence”
   Discover likes freshness, yet it also rewards reliable sources. If you publish 25 posts on Tuesday then nothing for ten days, you look like a pop-up stall. Aim for a steady drumbeat. For news, that can be daily. For evergreen, it might be three strong pieces a week, plus planned refreshes.

9. Refresh with rules, not vibes
   Update when you have something real to add (new numbers, a confirmed outcome, a corrected detail). Keep the same URL for true updates, so engagement signals stay consolidated. Publish a new URL when the angle becomes a new story. Avoid mass “republishing” that only changes a date, because it can backfire when users notice the thin change.

10. Build topic clusters with recurring coverage arcs
    Pick 3 to 6 core themes where you can publish repeatedly with depth (for example, “UK personal finance”, “AI policy”, “cybersecurity incidents”, “Premier League tactics”). Then run arcs: explainers, updates, case studies, and interviews. That steady body of work helps Google read you as a consistent source for that topic.

11. Stop thin category sprawl, strengthen a smaller set
    A broad site can still win, but thin sections bleed trust. If “Health” has five short rewrites, while “Tech” has 80 strong originals, trim Health or invest properly. In 2026, expertise looks more topic-based than site-wide, as many analyses have noted, including this overview of Discover’s topical signals.

Diagnosing Discover drops: seasonality, eligibility, or quality

When Discover traffic falls, the first risk is misdiagnosis. Don’t assume a penalty. Use a simple triage.

Start with Google Search Console’s Discover report. Look for whether the drop is across the whole site, one topic, or one template. Then separate these three buckets:

Seasonality and demand shifts
Some topics surge then cool (football transfer windows, budget day, a viral product). If impressions fell but CTR stayed steady, interest may have moved on.

Eligibility and presentation issues
If impressions collapse across many topics, check for technical blockers: indexing problems, accidental noindex, paywall changes, image preview settings, or heavy interstitials. Also re-check policy compliance. Discover can be stricter about clickbait-style packaging than your editorial team realises.

Content quality and trust signals
If impressions hold but CTR drops, it’s often headline and image mismatch. If CTR holds but impressions fall in one topic, your coverage might have gone thin, repetitive, or too similar to everyone else.

For context on Google’s updated stance, see coverage of revised Discover guidelines. Keep one principle front and centre: measure changes as experiments. Run controlled headline tests, log image swaps, and track impact by topic cluster, not by “site traffic” alone.

Conclusion

Discover rewards publishers who keep their promises. Tight headlines, honest images, steady timing, and deep topic arcs give the algorithm fewer reasons to hesitate. None of these fixes guarantees distribution, because the feed depends on each user’s interests. Still, a disciplined approach to google discover seo turns volatility into something you can read, test, and improve. Which one fix will you ship today: the headline rewrite process, or the image standard that stops mismatches?

Please follow and like us: