Author Attribution

​

Methodology

• A JSON-LD block with an author field (either a string like "Jane Doe" or a Person object like {"@type": "Person", "name": "Jane Doe"})
• A <meta name="author" content="…"> tag in the page head
• An element with itemprop="author", rel="author", or any class containing “author” (case-insensitive) — but only the first two count as schema-backed; the class-based fallback is visible-only

​

Verification

[...document.querySelectorAll('script[type="application/ld+json"]')].flatMap(s => { try { const d = JSON.parse(s.textContent); return Array.isArray(d) ? d : [d]; } catch { return []; } }).map(d => d.author).filter(Boolean)

​

Technical detail

• JSON-LD author field — every <script type="application/ld+json"> block is parsed, and any author property is extracted. If author is an object, its name field is used; if a string, the string itself is the author name. This sets authorInfo.schemaFound = true.
• <meta name="author"> — if no JSON-LD author found, the page’s meta[name="author"] tag is read. A non-empty content attribute sets authorInfo.schemaFound = true as well; the scanner treats meta-author as schema-equivalent because both are structured declarations.
• DOM element fallback — if neither of the above match, the scanner queries (in this order): [itemprop="author"], [class*="author" i], [rel="author"]. The first element’s trimmed text content is the author name. This sets schemaFound = false (visible-only).

• Author as an array — Schema.org allows "author": [{"@type": "Person", "name": "A"}, {"@type": "Person", "name": "B"}] for co-authored content. The crawler reads the first author’s name and sets schemaFound = true; co-authorship isn’t enumerated but the page still passes the signal.
• Organization as author — "author": {"@type": "Organization", "name": "Editorial Team"} is schema-valid and counts as schema-backed attribution. AI models accept organizational authorship for editorial brands, even though individual experts typically rank higher for expertise queries.
• Empty <meta name="author"> — a meta tag with content="" doesn’t set schemaFound. The scanner requires non-empty trimmed content.
• Generic class names — a <div class="author">Posted by</div> element matches the class selector but the text content is “Posted by” (not a name). The scanner extracts this as the author name with no validation; the page passes visible-only with a non-name string. AI models filter this out at the consumption side, but Cited reports it as detected.
• Server-rendered vs client-rendered bylines — bylines added by JS after page load are detected because the scanner reads the post-hydration DOM (3-second Puppeteer delay). Bylines that lazy-load on scroll may be missed.

• Author bio depth. A byline linking to a 500-word bio page with credentials scores the same as a byline with a plain text name. The scanner counts attribution presence, not the strength of the author’s credentials.
• Author authority across sites. AI models cross-reference authors across sites to assess authority (an author who writes for Wired and Cited’s blog ranks higher than one who only appears on Cited’s blog). The scanner only sees this page.
• Generic attributions. “Admin”, “Staff”, “Editor”, “Team” all count as visible-only attribution if they appear in a class-matched element. AI models treat these as effectively anonymous, but the scanner reports them as detected.
• Author images or social links. Visible photos and links to LinkedIn / X profiles add credibility but aren’t scored. Pages with strong bios beyond what the scanner detects underperform their attribution score in actual AI citation.