Imagine Google has already indexed four hundred pages that say a particular founder started a particular company in a particular year. The four hundred and first page saying the same thing adds nothing. Google has known this for a long time, and it has said so in two places: a patent granted to the company in 2022 that describes an information gain score, which rates a document by how much new information it would give a reader who has already seen the others, and its public guidance on helpful content, which asks whether a page provides substantial value compared with other pages on the same subject.
The practical consequence is simple and most people get it backwards. The most valuable thing you can publish about yourself is not the polished version of what everyone already knows. It is the part that exists nowhere else.
What the web does not know about you
For most people with a public profile, the web knows a title, a company and a city, repeated across a dozen profiles that copied each other. It usually does not know:
- How your name is pronounced.
- Where you were born and where you studied, stated plainly.
- What you actually did at each company, as opposed to the title you held.
- The story of how you got into your field.
- The two or three subjects you have real opinions about, and what those opinions are.
- Which of the five people with your name you are.
- What you have appeared on, in one place, with dates.
Every one of those is a fact Google cannot find elsewhere. Put them on a page it can read and that page scores well on the only dimension where a personal site can beat a hundred bigger ones: it says something new.
Why a bio does not capture this
Bios are written to be safe and short. They compress a career into three sentences that could describe a thousand other people, and because they are so compressible, they get copied, which means the web ends up with twenty near identical versions and no new information in any of them. The things with high information gain are the specifics the bio format strips out.
How to get the specifics out
- Be interviewed, and record it. A conversation with a good interviewer surfaces the origin story, the pronunciation, the actual work behind the title and the opinions, because those are the things a curious person asks about and the things nobody writes down unprompted. If it is published as a podcast episode you also get a dated third party page in the bargain, which is a separate benefit.
- Distill the transcript into facts. Pull out every checkable fact and every specific that was not already online. Birthplace, schools, what each role actually involved, how the name is said, which subjects you have opinions on and what they are.
- Write them plainly on your home page. An at a glance block for the biographical facts, a career section that says what you did rather than what you were called, a short statement of the subjects you own. Keep the prose ordinary. The value is in the facts, not the writing.
- Mirror them in the machine layer. Each fact stated in prose can be stated once more in the structured data underneath the page, where Wikidata and the assistants can pick it up without parsing sentences.
How much is enough
Not every fact needs to be new for the page to score. A record naturally repeats your title and company, and should. What matters is the proportion. A page that is ninety percent restated profile and ten percent new reads as another copy; a page that states the common facts once and then spends most of its length on specifics nobody else has reads as a source. Aim for an at a glance block with eight to ten facts, at least half of which appear nowhere else online, and a career section that says what you did in each role in a sentence that could not be copied from a job title. That is enough to be the page the others cite, and it is more than most people with a Wikipedia entry can say.
What this does to the graph
Facts stated plainly on a canonical page are facts that can be cited. Wikidata can reference your birthplace to your home page only if the home page says it. A knowledge panel can show your education only if some source Google trusts states it. An AI assistant can answer "where did she study" only if the sentence exists somewhere it can read. New information on your own domain is both the highest scoring kind of content and the raw material everything else is built from.
The questions that produce new information
If you are running your own interview, or briefing a host who will, these are the questions that reliably surface facts the web does not have. Each one is answerable in a sentence and checkable afterwards.
- How do you say your name, and has anyone ever got it wrong in a memorable way?
- Where were you born, and where did you grow up if that was somewhere different?
- What did you study, where, and did you finish?
- Take me through the jobs on your LinkedIn, but tell me what you actually did in each one rather than the title.
- What was the first thing you built or sold that worked?
- Which company did you leave, and why?
- What are the two subjects people ask you about more than any other?
- What is the opinion you hold in your field that most of your peers disagree with?
- Who gave you your start, and where are they now?
- What have you written, recorded or been on that is not on your site?
An hour of this yields twenty or thirty specific facts. Half of them belong on your record as plain sentences. The other half belong in the transcript, which is published, searchable and attributed to you.
From transcript to record, with an example
The transcript sentence is the high information gain artefact. The record line is the citable version. You want both: the first because it is new, the second because it is checkable.
Three mistakes
Generating a bio with an AI tool and publishing it. The tool can only restate what is already indexed, so the result has, by definition, no information gain, and often introduces errors that then get copied.
Keeping the interesting parts for a book or a keynote. One sentence and a pointer on your record is enough to register the fact; the full story can live wherever you sell it.
Publishing specifics on a profile you do not control. A long bio on a platform's site helps the platform's page, not yours. Put the facts on your own domain first and let the profiles summarize.
A note on the patent, for the skeptical
The idea is sometimes dismissed as SEO folklore, so it is worth being precise about the basis. Google's public guidance on helpful content asks, in plain words, whether a page provides substantial value compared with other pages in search results. Separately, Google holds a granted patent describing an information gain score: a measure of how much additional information a document would provide to a user who has already seen other documents on the topic, used to adjust ranking. Nobody outside Google knows how heavily either is weighted, and this post does not claim a number. What it claims is narrower and safer: when two pages are about the same person and one of them says things the other does not, the machine has a stated reason to prefer the one that adds something. For a personal record, that is reason enough to write the specifics down.