Introduction to optimising Alt tags
Alt tags are an often overlooked aspect of web search engine optimisation (SEO). However, they play a crucial role both for your site’s accessibility and for its optimisation for search engines (SEO). An Alt tag, or alternative text, is a brief description of an image, which is displayed when the image cannot be loaded on a web page. It is also used by assistive technologies, such as screen readers, to help visually impaired users understand the content of an image.
The importance of Alt tags for accessibility
In terms of accessibility, Alt tags are essential. They allow people who use screen readers and other assistive technologies to understand the content of your website. Without Alt tags, these users could miss important information and have a poorer user experience. Moreover, even those who do not use assistive technologies can benefit from Alt tags. For example, if an image does not load properly, the Alt tag can provide useful context.
The role of Alt tags in SEO
In addition to their vital role in accessibility, Alt tags also play an important role in search engine optimisation. Search engines, such as Google, cannot “see” images in the same way humans do, and therefore rely on Alt tags to understand their content. Well-optimised Alt text can help your site rank higher for relevant keywords and appear in image search results. Thus, not only do Alt tags improve your site’s accessibility, but they can also lead to increased organic search traffic.
Need a website?
Ask for a free quote!
html
Legacy of accessibility standards and the role of the attribute alt
When Tim Berners-Lee published the first proposal for the World Wide Web in 1989, he implicitly introduced the notion of a universal internet where content would be readable by any machine. Very early on, the W3C understood that part of the population cannot absorb visual information and included in the HTML 2.0 (1995) specification the attribute alt. From there, regulatory bodies such as the US Section 508
, the European Web Accessibility Directive and then the General Framework for Improving Accessibility (RGAA) in France, took up the idea and turned it into a legal obligation. Ignoring or incorrectly completing an alt tag therefore exposes you to fines or litigation, just like a shop without an access ramp.
Netflix’s experience in 2016 illustrates the importance of these standards. After a class action brought by the National Association of the Deaf, the platform had to overhaul its multimedia ecosystem to include subtitles and textual descriptions. Even though Netflix mostly deals with video, the case boosted the entire sector: the number of companies carrying out WCAG compliance audits doubled the following year, and with them the attention paid to attributes alt. In other words, the legislator has made “accessible SEO” an industry standard.
Direct impact on the user experience: from speech synthesis to offline browsing
On a well-designed site, the journey of a blind user differs little from that of a sighted user: they simply listen to the descriptions. A screen reader such as JAWS translates the HTML structure into audio signals; the synthetic voice jumps from heading to heading, reads links and, when it reaches an image, automatically switches to the attribute alt. If it is empty or irrelevant (“image1.jpg ), the experience becomes a labyrinth. User tests carried out by the University of Lyon show that 72 % of blind participants leave a page if they encounter more than two images without a textual alternative.
But the issue is not limited to disability. When 3G connections struggle, the browser may not display the image and instead show the text alt. In the archives of the New York Times website, viewing in aeroplane mode brings up descriptions such as “Front-page photograph of
the Apollo 11 launch, July 16 1969 . The user understands the historical context despite the absence of visuals. The same scenario occurs in countries where data costs are high; a simple tag alt becomes the main channel of visual information.
Correlation between visual description and semantic relevance
Since the Hummingbird algorithm (2013) and especially RankBrain (2015), Google has enhanced a page’s relevance using semantic signals. The attribute alt then plays a role equivalent to internal link anchors: it provides lexical context. If your product page mentions “waterproof Gore-Tex® hiking shoes” and the image displays a alt generic label like “shoes”, the algorithm loses an essential clue and the long-tail query “women’s goretex trekking boots” may slip through your fingers.
Conversely, Patagonia experimented in 2020 with rewording 18,000 tags alt using key phrases taken from their internal search tool. Result: +12 % organic traffic on Google Images and +5.8 % e-commerce conversions. Even more interesting: CTR on the classic universal search increased by 2 points, proof that a clear description of images also strengthens the overall semantic coherence of the HTML document.
Technical SEO: how tags alt influence indexing and ranking
Googlebot behaves like a browser without a graphical interface. It reads the markup, follows links, assesses the structure and records the text. Images, however, are handled by a separate processor that attempts to recognise shapes via computer vision. To reduce this costly step, Google uses tags alt as a shortcut. A Searchmetrics study (2022) reveals that 30 % of thumbnails displayed in the “Associated images” section of a SERP come from pages where the main keyword appears in the attribute alt.
The influence does not stop at the “Images” vertical. On the main SERP, the “Top Stories” block and rich results (recipes, products, how-to) include thumbnails. These thumbnails are weighted by the semantic relevance of the alt combined with the and the structured data. In short, a “vegan tiramisu” recipe that has an alt evocative (“creamy cuisine with plant-based mascarpone”) will be more likely to appear in a carousel than a simple “IMG_832.png”.
Another channel: Google Discover. Here, the algorithm favours visually engaging content linked to the user’s history. Tests conducted by Spanish media outlet El Confidencial showed that an update to their attributes alt to include the names of sports personalities doubled visibility on Discover during major events such as Roland-Garros.
Step-by-step methodology for auditing and optimising attributes alt
1. Inventory and categorisation
The audit starts by extracting all image URLs via a crawler (Screaming Frog, Sitebulb or a Python script using BeautifulSoup). Export the columns « image and « alt attribute . Sort them into three groups: « relevant , « missing , « non-descriptive . This simple taxonomy makes it possible to visualise the workload and assign priorities.
2. Alignment with the keyword map
Each unique page already has a target query. A supporting lexical field is then associated with the alt to avoid cannibalisation. On a “men’s running trainers” page, it is better to reserve the main keyword for the H2 title and use a synonym (“neutral running shoes for marathon”) in the tag alt.. This broadens the semantic spectrum while maintaining consistency.
3. Writing: respect the context-accuracy-conciseness triangle
This involves balancing three parameters:
• Context: indicate the action or intent (“athlete crossing the finish line of the Paris half-marathon”).
• Accuracy: mention key details (brand, colour, material).
• Conciseness: limit the description to 140 characters, beyond which some screen readers truncate the content.
4. Implementation and QA
After going live, test your changes with Lighthouse (the “Accessibility” tab) or the Web Developer extension. For large sites, automated tests via axe-core or Pa11y are essential. Run them in pre-production to avoid regressions: a CMS update can overwrite attributes alt if the media field is not locked.
5. Monitoring and A/B Testing
Then monitor the KPIs: Google Images impressions (Search Console > Performance > Filter “Images”), click-through rate, average positions and, on the business side, conversion rate on the impacted pages. An A/B protocol (Google Optimize, AB Tasty) makes it possible to validate that the variant enriched in alt actually gets more clicks without degrading loading speed (image weight or use of alternative images).
Case study: international e-commerce and organic traffic
Zalando, German fashion giant, decides in 2019 to overhaul its image library. The challenge lies in translation: each product has 20 to 40 photos, and the site covers 17 European markets. The SEO team starts by creating a linguistic model based on entities (colour, type of garment, use). It generates alt dynamic tags such as “burgundy red midi dress, short sleeves, summer collection”. These descriptions are then automatically translated by an NMT (Neural Machine Translation) engine trained on the fashion lexicon.
The results over 6 months: +18 % of sessions from Google Images, +7 % of new users, +3.2 % of revenue attributable to the organic channel. Notable point: performance is better in countries where competition in visual SEO is lower (Finland, Austria). Conclusion: optimising the alt generates a competitive advantage proportional to the level of market maturity.
Case study: public institution and RGAA compliance
The Bibliothèque nationale de France (BnF) holds more than 15 million iconographic documents. Its Gallica portal, launched in 1997, suffered from hundreds of thousands of images lacking text alternatives. In 2021, the BnF sets up a semi-automatic indexing project; a neural network (ResNet-152) extracts metadata while document specialists manually validate the suggestions which are then used to fill in the tags alt. Example: a First World War poster receives the description “Colour illustration: French poilu shaking hands with an American soldier, slogan La Fayette, here we are! .
In addition to legal compliance, Gallica notes a 25 % increase in visits from American academic institutions. The reason? English-speaking researchers can now find the resources via Google Images thanks to the multilingual descriptions inserted in the alt. The operation proves that accessibility can serve a dual purpose: inclusivity and international reach.
Common missteps and their consequences
It happens that developers leave the attribute alt="" on informative images. According to a WebAIM analysis (2023) covering one million home pages, 55 % contain at least one information-bearing image without a description. The opposite mistake also exists: over-describing a purely decorative element, which overloads the auditory experience and introduces semantic noise. The WCAG standard recommends the null attribute (alt="") for these cases.
Another pitfall: over-optimisation. Stuffing the attribute alt with keywords can trigger a quality filter (“Keyword stuffing ) and reduce visibility. BMW paid the price in 2006: Google temporarily delisted certain pages of its German site for abusive use of hidden keywords, including in the tags alt.
Beyond the attribute alt : longdesc, aria-label and JSON-LD schema
For complex diagrams (infographics, maps), a simple descriptive sentence is insufficient. The HTML standard provides (barely implemented) or, more commonly, an adjacent link leading to a detailed description. Example: the Air Quality Observatory publishes an interactive map every quarter; each image has an alt succinct (“ATMO Index, France, July ) and a “Full description” button linking to a text table — a practice required by the RGAA.
For functional icons (basket button, burger menu), we will choose aria-label rather than alt, because the icon is often decorative and scripted via CSS or SVG. Finally, structured markup (such as “ImageObject” in JSON-LD) makes it possible to add a caption, licence and creator. Google reads this schema to display a “Licence attributed” badge in Google Images, an advantage for photographers and image banks.
Future of the semantic image: generative AI, multimodal search and the Metaverse
In 2022, Microsoft implements anauto-captioning feature in Edge; the AI automatically generates a description when an alt tag is missing. Google, for its part, is refining its multimodal models such as MUM (Multitask Unified Model) capable of understanding and translating a visual query. In this context, providing a alt rich remains crucial: it feeds these AIs and increases the chances that they correctly match the image to the search intent.
Virtual worlds (Meta’s Horizon Worlds, the future Apple Vision Pro) will carry images in 3D. How do you describe an interactive object? The W3C’s work on Accessible 3D Web already proposes extending the logic of textual attributes to depth: a glTF file could include an “a11y” section listing elements readable by speech synthesis (“light wood chair, Scandinavian style, placed on a Berber rug ). Knowing how to write these descriptions will become a sought-after skill, at the crossroads of UX, SEO and 3D.
Thus, optimising alt is not just a line in a checklist; it is part of an overall evolution of the web towards universal accessibility and unified semantic search. Ignoring this attribute means depriving yourself of a bridge between visual content, search engine and users, both current and future.








