An infographic is an image that combines data, very short text and graphics to get something across in a few seconds. Put less elegantly: it is a PNG or JPG file with some numbers locked inside it.
The gap between those two sentences is everything you need to know to use infographics well today. A human reader looks at the image and understands. Google, ChatGPT and Perplexity read the HTML around it, and they do not get inside the image.
We looked at the eight organic results on the first page of Google UK for "infographics" (DataForSEO, Google UK, September 2026). Four are product pages for design tools: Adobe Express, Canva, Infogram and Visme. The others are Wikipedia, the Venngage guide, the University of Hull library guide and the showcase site Cool Infographics. Neither of the two guides deals with SEO or alt text in its body, and neither mentions screen readers; Venngage brings up accessibility only as a feature of its own design tool. To its credit, Venngage is also the one page that links its research claim to a source, an IEEE paper.
What an infographic contains
Wikipedia defines infographics as "graphic visual representations of information, data, or knowledge intended to present information quickly and clearly". The technique has a recognisable father: Otto Neurath (1882-1945), an Austrian sociologist who in the 1920s developed Isotype, a system of pictograms for presenting social data to people who could not read tables.
It has four parts, and if one is missing the result is decoration:
- A measured figure, not a catchy line. If you take the numbers out and the meaning stays the same, it was not an infographic.
- A visual hierarchy: the most important figure is the biggest, and you see it first.
- Text kept to a minimum. A caption, not a paragraph redrawn as graphics.
- The source, written inside the image and repeated in HTML underneath. This is the part that always gets skipped, and it is the part that makes the chart citable by someone else.
The nine types of infographic and the typical mistake with each
The most widely used taxonomy is Venngage's, which lists nine formats. It is a good list, but it stops at what each one is for. Here, next to each type, is the way it usually goes wrong, which is the useful information when you have to choose.
| Type | When to use it | Typical mistake |
|---|---|---|
| Statistical | One strong number people should remember | Putting twelve numbers in it so nobody remembers any |
| Informational | Explaining a concept without numerical data | It turns into a paragraph of text converted into an image |
| Timeline | A sequence with real dates | Approximate dates or intervals not drawn to scale |
| Process | Steps in a fixed order | Steps that actually run in parallel, drawn in a line |
| Geographic | Data that varies by area | Colouring a map by absolute values instead of per head of population |
| Comparison | Two or three alternatives on identical attributes | A rigged comparison: truncated axes or attributes picked to favour one side |
| Hierarchical | Levels or pyramids (org charts, priorities) | Implying a hierarchy where there is none |
| List | Unranked items to scan | Numbering them, and implying a ranking |
| CV | Applications where the design is the message | Sending it through an applicant tracking system, which cannot read it |
The last row deserves a note of its own. An infographic CV is an image: the systems that screen applications look for text, and they get nothing out of a PNG. The same mechanism that stops an infographic from bringing in traffic makes a graphic CV invisible to anyone filtering candidates automatically.
To a machine, the figure inside the image does not exist
Google explains how it works out what an image is about, and it does not say "I read it". In the Google Images documentation the sentence is: "Google extracts information about the subject matter of the image from the content of the page, including captions and image titles". In other words: from the text on the page, the captions, the titles. Not from the pixels.
On the page Google devotes to AI features the word order is even clearer. The advice is "supporting your textual content with high-quality images and videos, when applicable": images support the text. They do not replace it. And to appear in AI Overviews or AI Mode, Google writes, "a page must be indexed and eligible to be shown in Google Search with a snippet". A snippet is built from the text on the page. If your best number is inside a JPG, no snippet can contain it.
A check you can repeat in thirty seconds: the AI Overview Google UK shows for "infographics" cites seven sources (Wikipedia, Monash University, Visme, Canva, Infogram and two YouTube videos). None of the seven is an infographic. On a subject that is literally infographics, the system summarising the answer drew only on text and speech; the one image in the box is a thumbnail, not a source.
Why this matters more than it did a few years ago: the Pew Research Center tracked 900 US adults across 68,879 Google searches in March 2025, and in the 12,593 searches where an AI summary appeared, the share of visits with a click on a result fell to 8%, against 15% for searches without a summary. Only 1% of visits included a click on a link inside the summary. It is a US sample, desktop plus mobile, and should be read as an order of magnitude, but the direction is the same one we see on client projects, and we covered it when writing about zero-click searches.
In practice: if your content has only one chance of being cited, that chance goes through the text. Here is what actually reaches its destination depending on where you put the figure.
| Where you put the number | A human reads it | It is text in the HTML | It can end up in a snippet | An LLM can cite it from the page |
|---|---|---|---|---|
| HTML paragraph | Yes | Yes | Yes | Yes |
| HTML table (native table block) | Yes | Yes | Yes, including as a featured snippet | Yes |
| SVG with real inline text | Yes | Yes | Partly | Partly |
| PNG or JPG (the classic infographic) | Yes | No | No | No |
| Attached PDF | Yes, if they download it | Separate from the page | Only as a result of its own | Rarely |
| Image alt attribute | No (screen readers only) | Yes | No | Partly |
This is the rule we apply to every article: the numbers go in an HTML table, the image comes afterwards. In the piece on SEO for manufacturing, the figures from the Macropix case are in two text tables, not in a chart: the Italian keyword "monitor pubblicitario" (advertising display) climbing from position 88 to 2, and a 25.55% share of voice on the LED wall cluster in July 2026, ahead of Amazon at 18.08%. Anyone can copy and cite those rows. If they had been a screenshot, nobody could.
The 60,000 statistic does not exist: do not quote it
Sooner or later, reading about infographics, you come across the line "the brain processes images 60,000 times faster than text". It turns up in presentations, in agency articles and on software vendors' pages. There is no research behind it.
Jon Schwabish, who works on data visualisation for a living, traced its genealogy on his site PolicyViz in September 2015 and wrote: "The problem is that the number appears to be completely unfounded". The oldest trace he found is a 1997 3M sales document, Polishing Your Presentation, where the number appears on page 2 without any reference. In 2012 Alan Levine offered 60 dollars to anyone who could find the original study. Schwabish notes: "To date, the $60 remains unclaimed".
Two relatives that circulate on marketing sites share the same fate: "90% of the information the brain processes is visual" and "65% of people are visual learners". Neither leads back to a study you can consult, and the link, when there is one, usually goes to another marketing article that copied it from the one before.
The reason to remove these lines is not rigour in the abstract. An infographic exists to make people believe a figure, and here it comes with a false one. Anyone who checks a single line of your article and finds nothing behind it stops trusting the others too: it is the same mechanism Google describes when it talks about E-E-A-T. If you need a measured argument for why visuals help, verifiable numbers exist, they come from usability research and should be cited with name, year and link. We have collected the ones that hold up in SEO statistics.
Text inside an image is also a legal issue
The W3C's WCAG guidelines have a dedicated success criterion, 1.4.5 "Images of Text", at level AA. The text reads: "If the technologies being used can achieve the visual presentation, text is used to convey information rather than images of text", with two exceptions: an image the user can customise, and cases where that particular visual presentation is essential to the message. Logos fall under the exception. A table of numbers exported as a PNG does not.
In the UK that criterion already has legal backing. The Public Sector Bodies (Websites and Mobile Applications) (No. 2) Accessibility Regulations 2018 require public sector websites to meet WCAG 2.2 AA, and the same government guidance points out that "All UK service providers have a legal obligation to make reasonable adjustments under the Equality Act 2010 or the Disability Discrimination Act 1995 (in Northern Ireland)". Across the EU, the European Accessibility Act (Directive 2019/882) has applied since 28 June 2025 to services including e-commerce, and it also covers UK businesses selling those services to consumers in EU countries. It is no longer only a public sector matter.
How far we are from the goal is measured by WebAIM, which analyses a million home pages every year. In the February 2026 edition, 95.9% of the pages analysed had automatically detectable WCAG 2 failures, with an average of 56.1 errors per page. On images, 16.2% of all those on the home pages (10.8 per page on average) had no alt text, and another 10.8% of those with alt text had repetitive or questionable alternative text. Add the two together and more than one image in four tells someone who cannot see it nothing at all.
For an infographic, alternative text is the trickiest part, because a ten-word alt cannot hold what the image says. The workable solution is to duplicate: a short, descriptive alt text on the image, and the full data in HTML immediately below, as a table or a list. It serves screen reader users and it serves the crawler, and it is the same piece of work. We have written about it in more detail under web accessibility and image optimisation.
How to publish an infographic so that it works
Six steps, in the order they should be done. The first two decide the result, the other four protect it.
- Write the data in HTML first, then design. The table or paragraph with the numbers exists on the page, and the image illustrates it. If you reverse the order, you end up designing an image and then looking for text to justify it.
- One infographic, one claim. If the image says three things, the reader remembers none of them and nobody cites it for any one in particular.
- A descriptive file name. Google says so explicitly: "my-new-black-kitten.jpg" is better than "IMG00023.JPG", and generic names such as "image1.jpg" or "pic.gif" should be avoided. This applies to infographics as much as to photos.
- Alt text that describes, without stuffing in keywords. The Google Images documentation advises against "filling alt attributes with keywords (also known as keyword stuffing) as it results in a negative user experience and may cause your site to be seen as spam".
- A caption under the image, with the source and year. The caption is HTML text: it is the line that ties the image to the content of the page, and the one another site copies when it cites you.
- Clear embed code and licence. If you want the infographic to be republished, provide a ready-made code block with the backlink already in it. Without it, whoever reuses it saves the file and cuts you out.
A seventh step, optional but useful on content you want cited: mark up the data with structured data. Google asks for nothing special for AI features ("You don't need to create new machine readable files, AI text files, or markup to appear in these features", its documentation says), but the Dataset and ImageObject types remain the clearest way to declare what that number is and where it comes from.
When an infographic is worth it, and when it is not
For years the infographic was a link building tactic: you publish a chart, others embed it, they link to you. It worked because the chart was the only way to get hold of that data. Today a data page in text does that job better.
Ahrefs documented the case with irritating precision. On 18 August 2020 Joshua Hardwick described how a page of SEO statistics picked up 36 backlinks from 32 different domains from 515 emails sent (a 5.71% conversion rate and a 17.55% reply rate), reaching the top position for "SEO stats". The page was a list of numbers in text, organised by category. No infographic.
That does not settle the question in favour of text in every case. The infographic is still the right choice in three specific situations:
- When the shape is the information. A map, an org chart, a flow with branches: the spatial layout says something a list does not.
- When the channel only accepts images. On Instagram, on LinkedIn, in a slide, the PNG is the only format available. There the infographic does not compete with HTML, it has to replace it.
- When it accompanies text that already says everything. This is the most common case and the least exploited: the image acts as a visual anchor for a section the search engine reads anyway.
The case where it does not pay is just as easy to spot: when the infographic is the content. A page that is an image with two lines of introduction is, to a search engine, an almost empty page. If you are building a content marketing plan, this is the item to review first.
Which tools to make them with
Design editors with ready-made templates. Canva, Adobe Express, Visme, Venngage, Genially and Microsoft PowerPoint: you start from a template, change the text and colours, and export. Adobe Express, Canva and Visme hold three of the four tool slots on the UK first page for "infographics", and for an informational or process infographic they are perfectly fine. Genially adds an interactive version, which lives on a web page and so, unlike a PNG, contains real text.
Data visualisation tools. Datawrapper, Flourish and RawGraphs start from the spreadsheet rather than from a design template. They are the right choice when the content is a series of numbers rather than a concept: they produce charts that are correct by construction, with axes to scale, and in several cases they return an embeddable chart that stays readable as text.
On automatic generators that promise an infographic from a prompt (Infogram, the fourth tool on that UK first page, now describes itself as "AI powered from charts to infographics"), the verdict is provisional: they lay out quickly and readily get the scales wrong. They are fine for a draft, not for a number you publish. The same goes for AI-generated content in general: the problem is not the machine doing the drawing, it is the check that gets skipped.
If the point is to be found and cited by generative systems, the work is not on the image: it is on the structure of the content. We offer it as a service under the name AI SEO, and the results on real projects are collected in our projects.
Frequently asked questions
An infographic is a visual representation that combines data, minimal text and graphics to explain a subject in a few seconds. Wikipedia defines infographics as graphic visual representations of information, data or knowledge intended to present information quickly and clearly. The technique goes back to Otto Neurath's Isotype in the 1920s. Technically it remains an image file: the text it contains is not part of the page's HTML and search engines do not read it as text.
The nine most used formats are: statistical, informational, timeline, process, geographic, comparison, hierarchical, list and CV. The choice depends on the type of data: a sequence with real dates calls for a timeline, values that vary by area call for a map, two or three alternatives on the same attributes call for a comparison. If the data is a single number, the statistical infographic is the one people remember.
First write the data in HTML on the page, then design the image that illustrates it. Keep to one claim per infographic, give the file a descriptive name, write alt text that describes the image without piling up keywords, add a caption underneath with the source and year, and provide embed code with a backlink if you want it to be republished.
Automatic generators that work from a prompt lay out quickly and often get chart scales wrong, so they are fine for a draft but not for published data. For informational infographics, template-based editors such as Canva, Adobe Express, Genially, Venngage and Visme give a more predictable result. When the content is a series of numbers, Datawrapper, Flourish and RawGraphs start from the spreadsheet and produce charts that are correct by construction.
As a link building tactic they have lost ground to data pages in text, because a number inside a PNG does not make it into snippets and is not picked up by generative systems. They remain useful in three cases: when the spatial layout is the information, when the channel only accepts images, as on social media, and when they accompany text that already says everything. Text locked inside an image is also an accessibility problem: WCAG 1.4.5 asks for real text instead of images of text, and UK public sector sites must meet WCAG 2.2 AA.
Alt text cannot hold everything a complex infographic says. The workable solution is to duplicate: a short alt that describes what it is, and immediately below, in HTML, the full data as a table or list. It serves screen reader users and it serves the crawler, and it is a single piece of work. According to WebAIM, in the February 2026 edition of its analysis of a million home pages, 16.2% of images still have no alt text.
One thing the industry rarely says: the infographic has stopped being a format and has become an export format. The work that counts (choosing the figure, checking it, putting it to scale, stating the source) is the same work a table, a paragraph or a slide needs. Whoever keeps that work in a Canva file has a single copy and no original. Whoever keeps it in a spreadsheet and a published HTML table can export the image for Instagram, the PDF for sales and the slide for the meeting, and update them all by changing one cell. The tools on Google's first page sell the export. The part that decides whether anyone will cite you is the part you have to write.
