This translation has not been editorially reviewed yet. The German version is authoritative. Deutsche Fassung →
GEO / AI visibility
Why do AI programs especially like reading lists and tables?
AI programs prefer lists and tables, because every piece of information there is already ordered and can be extracted individually. An answer engine like ChatGPT has to laboriously pick facts out of running text; from a list or table, it can take them directly. Our check therefore counts the lists and tables in the main content: from two such blocks onward, you get full marks.
Ordered facts are easier to extract
Compare two presentations of the same information. As running text: "We check your website for accessibility, load time, and privacy, among other things, as well as findability." As a list: accessibility, load time, privacy, findability — four lines, each understandable on its own. For a person, both versions are readable. For a machine, the list is unambiguous: the start and end of every item are marked, nothing has to be interpreted. The same goes for tables, a comparison of service packages, say. Answer engines happily pick up such blocks directly into their answer, often citing the source too. Running text with no structure at all, on the other hand, gets cited less often.
How our check assesses lists and tables
Our check system counts every bullet list, numbered list, and table when it visits the page. From two such blocks onward, you get full marks; with exactly one block, a large share of the points, with a note that more structure increases usability. With no list or table at all, the check reports that plain running text is harder for answer engines to use. Important: real list elements in the code count, not paragraphs with dashes or hand-typed bullet characters. A paragraph that only looks like a list visually stays an ordinary block of text for machines.
- Look on your most important pages for enumerations in running text — services, benefits, or steps, say — and turn them into real lists.
- Present comparisons, prices, or opening hours as a table instead of as text.
- Use your content management system's list function, not hand-typed dashes.
- Label every list with a subheading that says what's being listed.
- Aim for at least two lists or tables on every content page.
FAQ
How many lists should a web page have?
Our check gives full marks from two lists or tables per page onward. That's deliberately set low: a services overview and a step-by-step process alone already meet the requirement. More structure doesn't hurt, as long as the blocks carry real content. Empty token lists with nothing to say help neither readers nor machines.
Do dashes in the text count as a list?
No. Machines only recognise real list elements in the page code, generated by the content management system's list function. A paragraph with hand-typed dashes or asterisks looks like a list to people, but stays plain running text for machines. Always use your system's bullet-list button instead of typing characters by hand.
Aren't tables impractical on phones?
Wide tables can become unwieldy on narrow screens, that's true. But the solution isn't to avoid them — it's a slim table with few columns, or side-scrolling within the table. Your agency can set that up. The benefit stays: answer engines prefer to extract and cite ordered information.
What does "extractable" mean for web content?
Extractable means: a machine can cleanly pull out a single piece of information without having to interpret the surrounding text. List items and table cells have clearly marked boundaries and are therefore easy to extract. AI programs build their answers from building blocks like these. The more easily extractable blocks your page offers, the more likely it is to show up as a source.