Paste the HTML of a page and this tool reports how its facts and figures are marked up: real tables or styled boxes, proper lists or paragraphs, numbers with units or without. It scores seven checks and tells you which ones to fix.
Why markup matters for data
A person looking at a pricing grid sees a table whether it was built with a <table> element, a stack of <div> elements or a screenshot. A crawler does not. It reads the HTML, and only a real table tells it that “₹4,500” belongs to the row “Prototype” and the column “Price”. Built from divs, the same grid becomes a run of unrelated words. In an image, it is nothing at all.
This matters more for AI assistants than it did for classic search, because assistants answer specific questions. “What tolerance can they hold on a 400 mm part” has an answer on your page, and whether the assistant can find it depends on whether the figure is attached to its label.
The seven checks
| Check | Points | What passes |
|---|---|---|
| At least one real HTML table | 25 | A <table> with two or more rows |
| Figures are present in the text | 20 | Two or more numbers per 100 words |
| Lists are marked up as lists | 15 | At least one <ul> or <ol> |
| Tables have header cells | 10 | Every table uses <th> |
| Figures carry units or currency | 10 | Numbers appear with ₹, %, mm, kg, days and so on |
| Headings break up the page | 10 | A heading at least every 350 words |
| No data built from div elements or images | 10 | No blocks styled as tables without table markup |
The weights are Brilad’s own judgement of what makes data extractable. They are not taken from any search engine, and no search engine publishes a score like this.
How to fix the common failures
- Data in div elements. Page builders often produce pricing tables this way. Replace them with a real table, or ask your developer to change the block’s output. Styling can make a real table look identical.
- Data in an image. Keep the image if it looks good, and add the same figures as a table below it.
- No header cells. Change the first row from
<td>to<th>. Without headers, a column of numbers has no name. - Few numbers. Replace vague words with figures you can stand behind. “Fast turnaround” becomes “12 working days”.
- Numbers without units. “Tolerance of 0.01” is ambiguous. “Tolerance of 0.01 mm” is a fact.
What this tool cannot tell you
It measures structure, not substance. A page can score 100 with a table of meaningless figures, and a superb essay with no data will score low and be none the worse for it. Use this on pages whose job is to state facts: product and service pages, specifications, pricing, comparisons. Do not use it to judge an opinion piece.
The score is not a ranking factor and a higher score does not cause better rankings by itself.
Frequently asked questions
How do I get the HTML of my page?
In Chrome, right click the main content, choose Inspect, right click the highlighted outer element, then Copy and Copy outerHTML. Paste that into the box.
Why not just enter a URL?
Because pasting keeps everything in your browser. Your HTML is never sent to our server, which matters for pages that are not yet public.
My table is wide and breaks on mobile. Is a div layout not better?
Keep the table and wrap it in a container that scrolls sideways. You get the mobile behaviour without giving up the markup.
Related
- Tables and data-dense formatting for AI search
- Machine readability and clean HTML
- Content Chunking Analyser: the same check for your prose
- All free SEO tools