Information Gain

Whether a page adds something beyond what is already ranking. Used by Lily Ray and relayed by Caleb Ulku; recorded as emerging because the two sources are not independent (Caleb is relaying Ray) and no official source in this wiki uses the term.

Definition

The term enters this wiki from It Works Until It Doesn't: AI Content Strategies That Backfire: AI-assisted content should “add original or unique information above and beyond what is offered by competing pages (information gain).” Her sharpest form is a pre-publish question: is there “any first-party data, expertise, or original perspective on this page that isn’t available on the first ten results already ranking for the query?”

In They Lied About AI Content (This Proves It) Caleb says the sites in Lily Ray’s study were not hurt for using AI; “what killed them is that the pages carried nothing that wasn’t already on the internet. We call this information gain.” He pairs it with a second question: does the page exist because a real customer needs it, or because a search engine or an LLM might cite it?

The operational form is Ray’s test, which he quotes: could a competitor publish a near-identical version of this page tomorrow using the same prompt and the same methodology? If yes, the page “has no reason to exist,” however well it reads.

His local-SEO argument is that service businesses are unusually rich in unpublished information, unlike B2B software companies whose every product fact is already on review and comparison sites. Neither he nor Ray cites Google documentation or a patent for the term, and the wiki holds no official source using it. The nearest official material is the people-first self-assessment in Creating helpful, reliable, people-first content.

Key Properties

• Independent of authorship: an AI-written page can pass and a human-written page can fail.
• Reproducibility test as the gate before publishing. (Ray)
• Ray’s other gates: would you be comfortable if Google, a journalist, or customers saw the full URL list in the subfolder; is the page transparent about its bias.
• Local sources of gain (Caleb’s examples): what a repair actually cost in a given neighborhood recently; failures that recur in houses of a given decade; which permit the city is slow on and which inspection item gets flagged; why installs take longer where original galvanized pipe remains; photos of the real truck; the named technician.
• Division of labor (Caleb): AI handles structure, category and service mapping, briefs, title tags, H1s, Internal Linking (Nav vs Editorial) plans, Schema Markup (Structured Data), cleanup of notes, and research into what already ranks. The business supplies what only it knows.
• Ceiling: AI “can’t know what happened on Tuesday” at a specific job.

Examples from Sources

In the Sources

Tension

Ray pitches no product or service in her piece. The remedy Caleb lands on is his own paid agent, so his framing doubles as a product pitch. The underlying idea does not depend on the tool: the examples he gives are things a business owner can write down without any software.

Related

Concepts: Scaled Content Abuse, Programmatic SEO, Core 30, 8-Pass AI Writing Pipeline, E-E-A-T (Experience, Expertise, Authoritativeness, Trust), Local SEO Entities: Caleb Ulku, Lily Ray, Core30 AI Tool, Google