Content

What Is Information Gain in SEO?

Information gain is what a page adds beyond what the top results already say, and the gain earns trust only when the new claim carries its evidence. A Google patent describes scoring that difference per document, relative to the pages a user has already viewed.

Information gain, illustrated
Finding information gain as a flow: top results, agreement area, the gap, the evidence, and the gain block placed after consensus facts

What does Google's information gain patent describe?

Google's patent on link information gain describes scoring the additional information a document holds beyond the documents a user already viewed.

The patent is US 11,354,342, with a priority date in 2018. The score is relative: one document scores high for a user who has read nothing on the topic and low for a user who has read three pages saying the same thing. The patent describes using the score to choose which documents to present next.

Google's filing on its generative search describes a related behavior: content a profile or session shows the user already read is omitted, or the summary goes deeper.

Digital Vikingz treats the patent and the pending filing as design logic, not a guarantee. A page that restates the top results offers no reason to be read after them.

What counts as information gain on a page?

Information gain counts when a page states an attribute, value, method or result that the top results do not state.

Four kinds qualify. First-party data is a number the business measured itself. Method is how the business does the work, with its own decision rules. Coverage is an attribute of the entity that no ranking page states. Examples are worked cases built from the business's own niche.

Each kind is specific to one business, which makes the gain hard for a competitor to copy the week it is published.

Suppose every ranking page defines a topical map and none shows a filled node row. A page that shows the row, with its page decision, reason, page type and main component, adds an attribute the others lack. The definition is consensus, and the node row is the gain.

How does a page support an information gain claim with evidence?

An information gain claim carries its basis in the same block: the measurement, the method, the sample and the date.

Evidence turns a claim only one site makes from an assertion into a checkable fact. A sentence such as 'service pages with the main component first earned more calls' needs its page count, period, comparison and how calls were counted.

Google's guidance on helpful, people-first content asks if content provides original information, reporting, research or analysis, and asks content that draws on other sources to add substantial value instead of rewriting them.

How generative systems rate confidence in a claim belongs to AI citation readiness, and the page-side rule here is narrower: every unique claim sits next to its basis, in the same block.

Where does information gain sit relative to consensus wording?

Consensus facts come first, in the wording the field uses, and the information gain block follows them.

A page that opens with its unique claim asks for trust before showing it knows the basics. A page that opens with consensus establishes the entity and its agreed attributes, so the gain reads as an addition, not a contradiction.

The balance is set per section: each H2 opens with the agreed answer, and the paragraphs after it carry the method, data or example the top results lack.

The ratio differs by page type. A definition page is mostly consensus with one gain block, such as a worked example. A methodology page is mostly gain inside a short consensus frame. A service page carries its gain in the process and the deliverables.

How is information gain found before a page is written?

Information gain is found by listing what the top results already say, then listing what the business knows that they do not.

Read the top results and record every attribute and value they share: the agreement area. Record the attributes some cover and others miss, and the questions none of them answer.

List what the business holds: measurements from its own work, its process and decision rules, and the questions buyers ask on calls. The overlap between the gaps and that knowledge is the gain set, and gaps without evidence stay out.

The content brief records both lists: consensus facts as triples in the field's wording, and gain facts in a marked block with their evidence. A brief with no gain block is a brief for a page the search engine already has.

Which pages fail to add information gain?

Pages fail to add information gain when every fact on them already appears in the top results.

Rewording does not count, because a synonym for a consensus fact is the same fact. Added length does not count either, because the attribute set stays the same. An AI-generated summary of the top results holds nothing beyond them.

Google's rater guidelines name pages auto-generated in bulk without curation as showing no effort, and count the effort that goes into page functionality. A working calculator or filter is gain that text alone does not match.

Templated page sets fail the same way: a location page that swaps only the place name repeats its sibling. The gain for each page in a set is the facts true of that value alone.

Sources

Patents and filings describe systems Google has invented. None is a confirmed ranking factor.

Illustration for the knowledge article What is AI citation readiness?AI citation readinessA page whose sentences can each be verified against one passage.Illustration for the knowledge article What is a content brief?What is a content briefThe document that turns one row of the topical map into one page.Illustration for the knowledge article What is microsemantics?MicrosemanticsRelevance at sentence level, decided by word order, subject choice and dependency structure.Illustration for the knowledge article What are author and organization entities?Author and organization entitiesThe named person and the named company behind a page.Illustration for the Semantic Content Writing serviceSemantic Content WritingBriefs and pages written from your topical map.Illustration for the AI Search Visibility serviceAI Search VisibilityGet named in ChatGPT, Perplexity, Gemini and AI Overviews.Talk to us speech bubble: message the Digital Vikingz team on WhatsAppTalk to usMessage the team on WhatsApp

Questions about information gain

Does a small business produce information gain without original research?
Yes. A dental clinic's record of the questions patients ask at consultation, with the clinic's answers, is information no directory holds.
How often does information gain need refreshing?
Each time competitors copy it. A gain that spreads across the top results becomes consensus, and the 90-day review chooses the page's next addition.
Can information gain come from images or diagrams?
Yes, when the caption states the new fact in text. A diagram of a process no ranking page shows adds information. Google's generative search filing describes reading image captions and text inside images as page content.
Does information gain conflict with topical coverage?
No. Coverage decides which attributes a page holds, and gain decides what the page adds. A page with gain and missing consensus attributes is incomplete, and a page with every attribute and no gain is redundant.
Is information gain the same as E-E-A-T?
No. E-E-A-T is the rater guidelines' framework of experience, expertise, authoritativeness and trust. Information gain is a patent's measure of new information relative to what a user saw, and first-hand experience is one common source of it.
Does citing another site's data count as information gain?
Not by itself, because the citation is consensus with attribution. The gain is the page's own interpretation, comparison or measurement set against the cited figure.
Does information gain make pages longer?
Rarely by much, because one measured fact with its basis adds more than a paragraph of restated definitions.
Can two pages on one site share the same information gain?
No. Each gain block lives on one URL, and other pages link to it instead of repeating it, because Google's deduplication shows only the most relevant of very similar pages.
Do FAQs carry information gain?
Yes, when the questions come from buyer calls that no ranking page answers. Questions copied from People Also Ask boxes are consensus in question form.
Who supplies the information gain on a client page?
The client's subject-matter contact supplies the facts, and strategy finds the gaps they fill. The audit asks for a subject-matter contact for this reason.
Does translated content carry information gain?
A translation of the site's own page carries the same gain in the new language, and a translation of another site's page carries none.
Book a callWhatsApp