What does Google say about SEO? /
Quick SEO Quiz

Test your SEO knowledge in 3 questions

Less than 30 seconds. Find out how much you really know about Google search.

🕒 ~30s 🎯 3 questions 📚 SEO Google

Official statement

Do not use structured data tags to mark images that are not visible to the user on the page. This violates the general structured data guidelines.
53:35
🎥 Source video

Extracted from a Google Search Central video

⏱ 1h13 💬 EN 📅 22/04/2021 ✂ 29 statements
Watch on YouTube (53:35) →
Other statements from this video 28
  1. 4:42 Le nombre de pages en noindex impacte-t-il vraiment le classement SEO ?
  2. 4:42 Trop de pages en noindex pénalisent-elles vraiment le classement ?
  3. 6:02 Les pages 404 dans votre arborescence tuent-elles vraiment votre crawl budget ?
  4. 6:02 Les pages 404 dans la structure d'un site nuisent-elles vraiment au crawl ?
  5. 7:55 Faut-il vraiment s'inquiéter d'avoir plusieurs sites avec du contenu similaire ?
  6. 7:55 Peut-on cibler les mêmes requêtes avec plusieurs sites sans risquer de pénalité ?
  7. 12:27 Faut-il vraiment vérifier les Webmaster Guidelines avant chaque optimisation SEO ?
  8. 16:16 La conformité technique garantit-elle vraiment un bon SEO ?
  9. 19:58 Pourquoi une redirection HTTPS vers HTTP peut-elle paralyser votre indexation ?
  10. 19:58 Faut-il vraiment supprimer tous les paramètres URL de vos pages ?
  11. 19:58 Faut-il vraiment déclarer une balise canonical sur toutes vos pages ?
  12. 19:58 Pourquoi une redirection HTTPS vers HTTP paralyse-t-elle la canonicalisation ?
  13. 21:07 Faut-il vraiment abandonner les paramètres d'URL pour des structures « significatives » ?
  14. 21:25 Faut-il vraiment mettre une balise canonical sur TOUTES vos pages, même les principales ?
  15. 22:22 Google peine-t-il vraiment à distinguer sous-domaine et domaine principal ?
  16. 25:27 Faut-il vraiment séparer sous-domaines et domaine principal pour que Google les distingue ?
  17. 26:26 La réputation locale suffit-elle à déclencher le référencement géolocalisé ?
  18. 29:56 Contenu mobile ≠ desktop : pourquoi Google pénalise-t-il encore cette pratique après le Mobile-First Index ?
  19. 29:57 Peut-on vraiment négliger la version desktop avec le mobile-first indexing ?
  20. 43:04 L'API d'indexation garantit-elle vraiment une indexation immédiate de vos pages ?
  21. 43:06 La soumission d'URL dans Search Console accélère-t-elle vraiment l'indexation ?
  22. 44:54 Pourquoi Google refuse-t-il systématiquement de détailler ses algorithmes de classement ?
  23. 46:46 Faut-il vraiment choisir entre ciblage géographique et hreflang pour son référencement international ?
  24. 46:46 Ciblage géographique vs hreflang : faut-il vraiment choisir entre les deux ?
  25. 53:14 Faut-il vraiment afficher toutes les images marquées en données structurées sur vos pages ?
  26. 64:03 Faut-il vraiment normaliser les slashs finaux dans vos URLs ?
  27. 66:30 Faut-il vraiment ignorer les erreurs non résolues dans Search Console ?
  28. 66:36 Faut-il s'inquiéter des erreurs 5xx résolues qui persistent dans Search Console ?
📅
Official statement from (5 years ago)
TL;DR

Google states that marking non-visible images on the page with structured data violates its general guidelines. This practice, often used to artificially optimize display in rich results, is considered deceptive. Specifically, only images that are actually displayed to the user should be marked — marking others exposes you to manual or algorithmic penalties.

What you need to understand

What is an invisible image in structured data? <\/h3>

We refer to a hidden image when a webmaster adds a structured data tag (schema.org ImageObject, for example) for an image that never appears in the user’s viewport. This includes images with display:none<\/strong>, absolutely positioned off-screen, or loaded only for Google.<\/p>

This technique has long been used to inject visuals optimized for rich snippets<\/strong> (recipes, products, articles) without affecting the actual layout. The problem? Google sees it as manipulation of search results.<\/p>

Why does Google ban this practice? <\/h3>

The logic is simple: structured data must reflect what the user sees<\/strong>. If Google shows an image in its results that the visitor will never see when landing on the page, it’s a form of misrepresentation.<\/p>

Google’s general guidelines for structured data require that the marked content is directly accessible and visible<\/strong>. Marking a hidden image violates this fundamental principle — and risks manual action or partial de-indexing of rich results.<\/p>

Which images are affected by this rule? <\/h3>

All images marked via schema.org that are not visually rendered for the user<\/strong>. This includes images hidden by CSS, lazy-load images that never display, invisible sprites, or visuals in 1×1 pixel.<\/p>

On the other hand, an image that uses lazy-load but ends up displaying<\/strong> on scroll remains compliant. The criterion is actual visual accessibility, not the technical timing of loading.<\/p>

  • Images with display:none<\/strong> or visibility:hidden<\/strong> are prohibited in structured data<\/li>
  • Images positioned off-screen via CSS (left:-9999px<\/strong>) violate the guidelines<\/li>
  • Visuals loaded only for Googlebot without user rendering are prohibited<\/li>
  • Lazy-load images that display on scroll remain compliant<\/li>
  • Only actually visible<\/strong> images in the DOM and viewport can be marked<\/li><\/ul>

SEO Expert opinion

Is this rule really new or just a reminder? <\/h3>

Let’s be honest: this directive has been circulating in the guidelines for years. What’s changed is that Google is publicly reaffirming it<\/strong> — likely a sign of increasing abuse in certain verticals (e-commerce, recipes, articles).<\/p>

On the ground, many sites continue to mark hidden images without visible penalties. [To be verified]<\/strong>: Does Google systematically detect this manipulation, or only during manual audits? The consistency between automatic detection and sanction remains unclear.<\/p>

What nuances should be applied to this prohibition? <\/h3>

The real issue is the definition of “visible”<\/strong>. Is a lazy-load image that only loads when the user scrolls compliant? Yes, as long as it eventually displays. A CSS background image that only appears on mobile? Technically gray — but if it’s rendered somewhere, it’s acceptable.<\/p>

The blurry area mainly concerns alternative images<\/strong>: for example, a site offering a desktop version and a mobile version with two different visuals, both marked. If one of the two never displays for a given user, is it a violation? Google does not make this clear.<\/p>

In what situations is this rule circumvented in practice? <\/h3>

Some SEOs inject 1×1 pixel transparent images<\/strong> at the top of the page, technically visible but imperceptible. Others use carousels or tabs with pre-loaded images in the DOM but not displayed by default.<\/p>

These gray tactics do not strictly violate the rule — the image is technically rendered — but they betray its spirit. If Google detects a pattern of manipulation (image size vs. actual visibility), there’s no telling it won’t impose manual penalties. The risk exists, even if the probability remains low given the current state of detection.<\/strong><\/p>

Attention:<\/strong> Sites with high traffic or subject to frequent manual audits (health, finance, YMYL) are at greater risk. A manual action on structured data can wipe out all rich snippets at once.<\/div>

Practical impact and recommendations

What should you do to remain compliant? <\/h3>

Start with an audit of marked images<\/strong>: collect all image URLs from your JSON-LD, Microdata, or RDFa, then verify that they actually appear in the user’s viewport. A simple script can cross-reference the structured data with the visible DOM.<\/p>

Next, eliminate any images with display:none<\/strong>, off-screen CSS, or invisible sprites. If you use different images for desktop/mobile, ensure that the structured data markup corresponds to the version actually served — or adopt a single responsive image.<\/p>

What errors should be prioritized to avoid? <\/h3>

Never mark an image that exists only for Google. Some CMS or SEO plugins automatically generate default images<\/strong> (logo, placeholder) even when no image is displayed on the page — disable this logic.<\/p>

Avoid marking images in carousels or tabs if only certain ones are displayed. Better to mark only the first visible one, or accept not having a rich snippet rather than risking a penalty.<\/p>

How can I verify that my site complies with this rule? <\/h3>

Use Google’s rich results testing tool for each type of strategic page. Verify that the images shown in the report actually appear on screen<\/strong> when visiting the page in normal browsing.<\/p>

A manual audit remains the most reliable: for each page with structured data, load it in incognito, inspect the marked images, and confirm that they are all visible without manipulating the DOM. If you detect discrepancies, correct them before Google does it for you.<\/p>

  • Audit all images marked in structured data and check their actual visibility<\/li>
  • Remove schema.org tags from images with display:none or off viewport<\/li>
  • Adapt structured data to desktop/mobile versions without duplicating hidden images<\/li>
  • Disable plugins or CMS logic that inject default non-displayed images<\/li>
  • Regularly test with Google's Rich Results tool to detect inconsistencies<\/li>
  • Prioritize a single responsive image over multiple marked versions<\/li><\/ul>
    Marking invisible images violates Google's guidelines and exposes you to penalties on rich results. Only images that are actually displayed to the user should be marked. A thorough audit and adaptation of templates are essential to remain compliant. These optimizations often touch on the technical structure of the site and fine-tuning of CMS settings — if the stakes warrant the investment, hiring a specialized SEO agency can expedite compliance and secure long-term implementation.<\/div>

❓ Frequently Asked Questions

Une image en lazy-load est-elle considérée comme invisible par Google ?
Non, tant qu'elle finit par s'afficher au scroll. Google distingue le chargement différé (conforme) de l'invisibilité permanente (non conforme).
Puis-je baliser une image qui apparaît uniquement sur mobile ?
Oui, si elle est réellement visible pour les utilisateurs mobiles. Assure-toi que la structured data correspond à la version servie, sinon utilise une seule image responsive.
Que risque mon site si je baise des images cachées ?
Perte des rich snippets, action manuelle sur les structured data, voire désindexation partielle. Le risque dépend de l'ampleur de la manipulation et de l'exposition du site.
Les images en arrière-plan CSS sont-elles conformes pour les structured data ?
Zone grise. Si elles sont rendues visuellement, techniquement oui — mais Google préfère les images HTML classiques. Évite de baliser un background-image non critique.
Comment vérifier rapidement si mes images balisées sont visibles ?
Utilise l'outil de test des résultats enrichis de Google, puis compare les URL d'images détectées avec ce qui s'affiche réellement en navigation incognito sur la page.

🎥 From the same video 28

Other SEO insights extracted from this same Google Search Central video · duration 1h13 · published on 22/04/2021

🎥 Watch the full video on YouTube →

💬 Comments (0)

Be the first to comment.

2000 characters remaining
🔔

Get real-time analysis of the latest Google SEO declarations

Be the first to know every time a new official Google statement drops — with full expert analysis.

No spam. Unsubscribe in one click.