Checks
Readability yield
Measures what share of a page's visible text is actual main content versus navigation and boilerplate, specifically from the lens of what an AI content extractor can isolate.
What this check looks for
This check compares the length of readability-extracted main content against the page's total visible text, requiring at least 200 characters of visible text to run at all (thinner pages report "na"). If readability extraction found no main content whatsoever, it fails outright. Otherwise, a yield of 50% or higher passes; between 25% and 50% warns; below 25% fails, in each case reporting the exact ratio.
Why it matters
This check measures essentially the same underlying signal as the content.boilerplate-ratio check, but from a different angle: it specifically models what a content-extraction algorithm (the same family of tools AI systems use to isolate an article's body from surrounding chrome before summarizing or quoting it) is able to pull out of the page. A page where extraction yields little or nothing means an AI system reading it gets navigation links and footer text instead of your actual message — the content might be excellent, but if it can't be isolated from the surrounding page, it's effectively invisible to automated readers, even though a human visitor could easily spot the "real" content just by looking at the layout.
How to fix it
Structure the primary content so it's the dominant, clearly-delineated block on the page:
<!-- Content buried among heavy nav/sidebar chrome -->
<body>
<nav>...</nav>
<aside>...</aside>
<div><p>Actual article content here.</p></div>
<footer>...</footer>
</body><!-- Fixed: wrapped in <main>/<article> with substantive paragraphs -->
<body>
<nav>...</nav>
<main>
<article>
<h1>How Noise Cancellation Works</h1>
<p>Substantive, multi-paragraph explanation of the topic...</p>
</article>
</main>
<footer>...</footer>
</body>See how your site scores on this check