Why anonymous bylines fail YMYL
Every major LLM applies stricter extraction filters to financial, medical, legal and regulatory content. Part of that filter is an author check: who wrote this and is that person verifiable? Anonymous “Editorial Team” bylines fail the check almost universally. The page does not get extracted; a competitor with named experts does.
We have measured this on identical content. Two pages, same words, one with anonymous byline, one with schema.org Person markup pointing to a verifiable LinkedIn — the named-expert page outcites the anonymous page by 1.8–2.5× across ChatGPT, Perplexity and Google AI Overviews on YMYL prompts.
For non-YMYL content the lift is smaller — maybe 1.2–1.4× — but still systematic.
The implementation
For every named expert, deploy schema.org Person markup. Required fields:
name— full legal namejobTitle— current roleworksFor— Organization linked back to your site@iddescription— one-paragraph bio with concrete credentialsknowsAbout— array of topics they actually knowsameAs— array of verifiable external profiles
Optional but strong:
alumniOf— university with link to its homepageimage— real photo URLaward— published recognitions, with@type: Award
The verification standard for sameAs
sameAs is the field that decides whether the schema works. AI extractors check the URLs. Required:
- LinkedIn — always. The single most important external profile.
- One additional verifiable:
- Bar admission registry or equivalent regulatory body (legal / compliance / fintech)
- Professional society or board (medical / engineering)
- Wikidata Q-number (if the person has published / been written about)
- Conference talk page (with bio + abstract, not just a slide deck)
- Named industry directory listing (Crunchbase founder profile, AngelList, etc.)
What does NOT count:
- A profile you control on a low-authority site (the firm’s own about page does not count)
- A Twitter handle (most LLMs do not weight social as authority)
- An interview on a content-mill blog
- A directory listing without bio text
What we will not do
- Place named experts on pages they did not write or review
- Invent credentials to make a junior bio look senior
- Add sameAs URLs that 404 or that point to placeholder profiles
- Co-sign content with names from outside the firm without their consent
The E-E-A-T trap most agencies fall into: fabricating authority. Google’s spam team flags it; LLMs detect it; the brand pays the reputational cost. We refuse engagements where the client wants us to invent named-expert signals.
The HARO / Featured.com layer
Beyond on-site Person schema, the authority graph extends to where else the named expert is quoted publicly. Three services we use across engagements:
- Featured.com — formerly Terkel. Match-style platform connecting experts to journalist queries. We use it to place named partners as quoted sources in tier-1 publications.
- Source of Sources (formerly HARO) — same model, broader query universe. Lower hit rate, broader reach.
- Qwoted — better signal-to-noise for B2B SaaS and fintech.
Three to five published quotes per quarter per named expert is enough to anchor the authority graph. The quotes feed back into the sameAs / authority weighting, and the publication links become AI-trusted sources in their own right.
Time to implement
Per expert:
- 2–4 hours to gather verifiable sameAs URLs and write the bio
- 30 minutes to deploy the schema
- 15 minutes to validate
For a firm with five named experts, plan one full day to get the named-expert layer fully shipped. After that, quarterly maintenance — checking for link-rot on sameAs, updating credentials, refreshing photos.
The maintenance pitfall
sameAs URLs rot. LinkedIn URLs change when people update their handle. Conference talk pages disappear. Industry-directory listings get archived behind paywalls. Quarterly link-rot check is mandatory. A 404 on a sameAs link is the kind of thing that takes a page from “verified expert” back to “anonymous” in the LLM extractor’s view.
We bake the quarterly check into the Scale engagement calendar. On Growth or Starter the client owns it.
What you should do this month
- List your named experts (founders, principals, technical leads)
- For each, gather LinkedIn + one verifiable other (bar registry, talk page, Wikidata)
- Deploy schema.org Person markup on each
/team/[slug]page - Reference each expert’s
@idfrom any article they authored - Validate against Schema.org and Google Rich Results Test
That single workstream lifts citation rate on YMYL pages within sixty days. It is the single biggest lever on regulated content and the single cheapest move to ship if you have the people in-house.