When a CMS is too hard to use, teams find a workaround: they paste text into a graphic and upload it as an image. The page looks right. But to a search engine - and now to every AI answer engine - that content simply does not exist.
Text in images SEO is a silent failure mode: image-based content - text baked into a JPEG or PNG instead of written as real HTML - is one of the most damaging shortcuts a marketing team can take. It is invisible to Google's crawler, invisible to screen readers, and, increasingly important, invisible to the AI answer engines that more and more people use instead of search. If your key messaging lives inside pictures, you are absent from both the search results and the AI-generated answers your customers now read.
This matters more than ever because discovery is splitting into two channels. Traditional SEO gets you ranked in search results. GEO - generative engine optimization - gets your content surfaced and cited by tools like ChatGPT, Perplexity, Google's AI Overviews, Gemini, and Copilot. Both channels read text, structure, and links. Neither reads pixels. We see this pattern often when a Drupal site feels uneditable and marketing teams upload mockups as images instead of fighting the CMS. This post breaks down exactly what image-based content costs you across SEO, GEO, UX, and content management, and how to fix it with structured, editable content in Drupal.
In this article:
- Why do teams end up with text trapped in images?
- What does image-based content cost you in SEO?
- How does image-based content hurt your AI visibility?
- What are the UX and accessibility costs?
- What does image-based content cost in day-to-day management?
- How do you fix it with structured content?
- How do you measure the impact?
- Frequently asked questions
Why do teams end up with text trapped in images?
Teams resort to image-based content when editing real content is harder than faking it. It is almost never laziness - it is a rational response to a broken tool.
The pattern usually forms like this:
- the CMS editing experience is confusing or only half-implemented, so editors avoid it;
- designers deliver page sections as finished mockups, and the fastest path is to upload the mockup as-is;
- pasting an image is genuinely quicker than fighting field by field with an admin form;
- the result looks pixel-perfect, identical to the design, which feels like a win.
It is far more common than people assume, especially on corporate sites running a CMS mandated by headquarters that nobody configured well. When Drupal Paragraphs are installed but not wired up for editors, teams default to screenshots and graphics with text baked in. The content appears on the page, the deadline is met, and the damage stays invisible - until someone checks the analytics or the search rankings.
What does image-based content cost you in SEO?
Text inside an image cannot be crawled, indexed, or ranked, which means image-based pages are effectively blank to search engines. Every ranking signal you would normally earn from your content disappears.
The specific SEO losses are:
- No indexable text. Search engines read HTML, not the words rendered inside a picture. Your messaging, product descriptions, and keywords contribute nothing to ranking.
- No heading hierarchy. An image has no H1, H2, or H3. Search engines lose the structural map they use to understand what a page is about.
- No keyword signals. With no machine-readable text, there is nothing for ranking algorithms to match against a query.
- No internal links. You cannot place a hyperlink inside a flat image, so image-based sections break your internal linking and the crawl paths it creates.
- Weaker Core Web Vitals. Large image files inflate Largest Contentful Paint (LCP) and slow the page, and performance is itself a ranking factor.
The net effect: a visually polished page that contributes almost nothing to your organic visibility. A modern CMS should expose SEO features like editable metadata, clean URLs, and structured content - image-based sections bypass all of them.
How does image-based content hurt your AI visibility?
AI answer engines build their responses from machine-readable text, so content locked in images is excluded from the answers - and the citations - they generate. This is the GEO cost, and it is growing as users shift from searching to asking.
When someone asks ChatGPT, Perplexity, or Google's AI Overview a question, the engine retrieves text passages, summarizes them, and often cites the sources. To be retrieved and cited, your content has to be:
- Readable as text, not rendered as a graphic;
- Clearly structured, with descriptive headings and short, self-contained paragraphs an engine can lift cleanly;
- Answer-first, stating the key point directly rather than burying it;
- Backed by lists, tables, and definitions, the formats these systems extract most reliably;
- Marked up with schema (such as Article and FAQ structured data) so machines understand the content's role.
Image-based content fails every one of these tests. Even when a model can run OCR on a picture, the retrieval and citation pipelines that decide who gets quoted run on clean HTML, headings, and links. If your competitors publish structured text and you publish pictures, they get cited in the AI answer and you do not.
Alt text helps accessibility and gives crawlers a hint, but it is a caption, not a substitute for real content - and you can generate and improve alt text with AI modules in Drupal once your actual content lives in HTML where it belongs.
What are the UX and accessibility costs?
Beyond discovery, image-based content degrades the experience for the humans who do reach the page, and it can put you in breach of accessibility law.
- No responsiveness. Images do not reflow. On a phone, text inside them shrinks to the unreadable or overflows the screen.
- Accessibility failures. Screen readers cannot parse text inside an image, which is a direct WCAG violation. Building an inclusive, accessible site is impossible when your content is pictures.
- Slow loading. Heavy image files punish users on mobile networks, exactly where much of your traffic is.
- No interaction. Visitors cannot select, copy, or search text that is really a picture.
- No translation. Multilingual sites cannot auto-translate words inside an image, so managing a multilingual site breaks down completely.
What does image-based content cost in day-to-day management?
Image-based content is expensive to maintain because every change is a design task, not an edit. The operational drag compounds over time.
- changing a single word means recreating and re-exporting the entire graphic;
- there is no meaningful version history for content stored as pixels;
- different people produce slightly different images, so styling drifts and consistency erodes;
- the same message cannot be reused across pages without duplicating images, multiplying the maintenance burden.
What should be a thirty-second edit becomes a ticket, a designer, and a redeploy. That is the opposite of the component mindset shift: editors should combine reusable blocks, not request new graphics for every tweak.
How do you fix it with structured content?
The fix is to replace images of text with real, editable text built from structured components. In Drupal, that means Drupal Paragraphs: each section becomes a configurable component with proper fields instead of a flat picture.
Done well, this gives you everything the image approach destroyed:
- A real heading hierarchy (H1 → H2 → H3) that search engines and AI engines use to understand the page.
- Crawlable, indexable, citable text that works for SEO and GEO at the same time.
- Responsive, accessible, translatable content by default.
- Structured data opportunities, since real content can carry Article, FAQ, and other schema markup.
- Reusable building blocks so the same content can appear across pages without duplication.
This table sums up the difference:
| Capability | Text baked into an image | Structured HTML content |
|---|---|---|
| Indexed by search engines | No | Yes |
| Readable by AI answer engines | No | Yes |
| Heading hierarchy (H1–H3) | None | Full |
| Responsive on mobile | No | Yes |
| Screen-reader accessible | No | Yes |
| Auto-translatable | No | Yes |
| Editable without a designer | No | Yes |
| Can contain links | No | Yes |
The practical migration approach is page by page, starting with your highest-traffic pages, where the SEO and GEO upside is largest. You do not need to convert the whole site at once. A phased CMS modernization - replacing the hero and main content sections of your top pages with structured components - delivers measurable, quick wins while the rest follows.
Read also: a quick way for editing and customizing a Drupal paragraph and what is technical SEO and how to do it on your website.
How do you measure the impact?
Track the change with before-and-after metrics across crawling, performance, and rankings, and give SEO results time to materialize.
Measure these:
- Indexation. Indexed pages and the volume of indexable text, via Google Search Console.
- Page speed. Lighthouse scores and Core Web Vitals (LCP, CLS, INP) before and after.
- Rankings. Positions for your target keywords on the converted pages.
- Mobile usability. Mobile usability and accessibility scores.
- AI presence. Whether your pages start appearing in AI Overviews and being cited by tools like Perplexity for relevant questions.
Set expectations on timing: SEO changes typically take four to twelve weeks to show, as engines recrawl and reassess. The UX, accessibility, and performance gains, by contrast, are immediate. In one production rebuild we worked on, replacing image-based content with structured, responsive components on the first redone landing page coincided with an approximately 24% increase in conversion rate - the combined effect of real content, faster loading, and proper mobile layout.
Frequently asked questions
Is text in images bad for SEO?
Yes. Search engines cannot read or index text rendered inside an image, so that content provides no keywords, no heading structure, and no ranking signals. The page is effectively blank to Google.
Can AI tools like ChatGPT or Perplexity read text in images?
For the purpose of being cited in answers, generally no. Even when a model can OCR an image, the retrieval and citation systems that decide which sources to quote rely on clean, structured HTML text - so image-based content is left out of AI answers.
What is GEO (generative engine optimization)?
GEO is the practice of structuring content so it gets surfaced and cited by AI answer engines such as ChatGPT, Perplexity, Google AI Overviews, Gemini, and Copilot. It rewards machine-readable text, clear headings, direct answers, lists, tables, and schema markup.
Does adding alt text fix the problem?
No. Alt text improves accessibility and gives crawlers a short description, but it is a caption, not a replacement for real content. Your actual messaging needs to live in HTML text.
How long until SEO improves after switching to real text?
Usually four to twelve weeks, as search engines recrawl and reindex. Performance, accessibility, and mobile improvements appear immediately on launch.
Ready to turn your images back into content that ranks and gets cited?
Text trapped in images is a silent tax on your visibility - it costs you search rankings, AI citations, accessibility, speed, and the ability to update your own site. Converting it to structured, editable content fixes all of those at once, and it is usually far less work than teams expect.
Want to audit where image-based content is hurting your site, or rebuild your key pages as editable, SEO- and GEO-ready components? Our team specializes in structured, search- and AI-friendly Drupal implementations. Visit our Drupal development services to see how we can help.
