Information gain is a Google ranking concept where the algorithm evaluates whether a page adds new, unique information not already covered by other pages ranking for the same query. Pages with original research, proprietary data, and distinctive expert perspectives outrank those that merely repackage existing published content. Google filed a patent on this concept in 2020.
The "skim-the-top-10-and-rewrite" content strategy is dying. Google's information gain concept is why. When the system compares your page to everything else ranking for a query and finds you're saying the same things in different words, it has no reason to rank you above the pages that said it first.
What is information gain?
Google filed a patent in 2020 describing an "information gain score" — a system that evaluates top-ranking results and measures how much new information each document adds compared to documents a user has already encountered in the search session. A higher information gain score produces a ranking boost.
The concept shifts what earns ranking power. Previously, intent-matching and keyword presence dominated. Information gain adds a third dimension: novelty. You must match intent and introduce something the other ranking pages don't say.
Google's Helpful Content guidance explicitly rewards content providing "a unique or at least original" perspective. Combined with the 2020 patent, this signals that information gain is embedded in how Google evaluates content quality — not a future signal, but a present one.
Why information gain matters for content strategy
Information gain changes the competitive equation in three specific ways:
- It penalises duplicate-approach content. If you studied the top 10 results for a keyword and wrote an article covering all the same points, you've created zero information gain. Google sees this as an article that exists, not one that contributes.
- It rewards first-party data. Original surveys, client benchmarks, A/B test results, and proprietary case studies are high-information-gain assets by definition — nobody else has them.
- It makes expert quotes a ranking signal. A quote from a practitioner expressing an opinion absent from any other ranking page adds measurable information gain. This is why "expert roundup" formats, done properly, earn rankings.
How information gain works — the mechanics
The patent mechanism
Google's system compares documents within result sets, measuring new information each document adds versus documents a user has already seen. If a user clicks results 1 through 4 and all four say the same things, result 5 that says something genuinely different receives a higher information gain score — and may rank above results that matched intent but added nothing new.
What counts as information gain
- Original statistics and self-collected data
- Unique client case studies with real numbers
- Expert quotes that appear nowhere else online
- Novel frameworks or proprietary methodologies
- Contrarian analysis backed by evidence
- Specific benchmarks from your own testing
- Unique angles on topics everyone covers the same way
What does not count
- Rephrasing existing common points in different words
- Adding extra subheadings without substantive content under them
- Padding word count with generic advice
- Content that replicates what the top 10 already say
Types of information gain by source
| Source | Example | Gain level | Feasibility |
|---|---|---|---|
| Proprietary client data | Benchmark from 500 client sites | Very high — fully unique | Requires client pool |
| Original survey | Survey of 200 practitioners | High — hard to replicate | Moderate cost |
| Expert quote | Exclusive comment from named expert | High if genuinely novel | Low cost, high effort |
| Personal test results | Your A/B test findings | Medium — others can replicate | Low cost |
| Curated public data | New synthesis of existing studies | Low-medium | Easy |
| Rewritten top-10 content | Same points, different words | Near zero | Easy but worthless |
Real information gain examples
1. SEO agency page speed benchmark
An SEO agency published page speed improvement content with a specific claim: "Original benchmark data from 500 client sites: reducing JavaScript bundle size by 40% improved LCP by 1.2 seconds on average." That specific data point existed nowhere else online. The result was high information gain — a fact Google could extract and cite that no competing page could match.
2. Local HVAC pricing page
A local HVAC company replaced generic "$100–$500" service ranges with actual quoted prices from their own service providers, broken down by job type and city. Competitors all showed ranges. This page showed real numbers — and ranked above national directory listings with far more backlinks.
3. SaaS B2B content play
A B2B consultant documented exact LinkedIn post reach increases after applying a specific posting cadence: from 200 to 2,000+ views per post. The before/after numbers with a named person's account made the claim verifiable and unique — standard information gain.
Information gain vs. topical authority — how they interact
These are different levers on the same ranking machine.
Information Gain
- Evaluated per page, per query
- Rewards novelty and originality
- Measured against competing pages
- Strongest for individual article performance
- Built through original data and expert insight
Topical Authority
- Evaluated across a site's content cluster
- Rewards depth and comprehensiveness
- Measured against site coverage of a topic
- Strongest for ranking across a keyword category
- Built through volume and internal linking
The ideal play combines both: publish enough content to build topical authority, and ensure each piece adds something novel to earn information gain scores on individual queries.
7 ways to build information gain into every piece of content
- Run a pre-write gap analysis. Before writing, open the top 5 results for your target keyword. List every claim they make. Your article must say something none of them say.
- Mine your own client data. Even one anonymised case study with real numbers creates information gain that 99% of competitors can't replicate.
- Conduct micro-surveys. A 10-question LinkedIn poll of your audience creates first-party data. Publish the results; they're automatically unique.
- Quote practitioners, not pundits. Generic expert quotes add nothing. A quote from a named practitioner sharing a specific opinion they haven't published elsewhere — that adds gain.
- Test and document. Run an experiment. Document the setup, the result, the anomalies. Your testing documentation is inherently unique.
- Take a position. Contrarian views backed by evidence add information gain. "Contrary to common advice, X doesn't work when Y" — if supported — is a high-gain signal.
- Update with new data, not just new dates. Refreshing content by changing the year in the title adds zero information gain. Refreshing with new statistics, new case studies, and updated positions adds genuine novelty.
Adding sections and subheadings to hit a word count target adds zero information gain if the content under each heading repeats what competing pages already say. Google measures novelty per statement, not per word. One original data point outweighs 2,000 words of repackaged knowledge.
Common information gain mistakes to avoid
- The top-10 rewrite — researching competing pages and summarising their points in your own words is explicitly what information gain penalises.
- Fake originality — adding a "unique take" section that just restates conventional wisdom in a slightly different order. Google's comparison is semantic, not syntactic.
- Generic expert quotes — "According to experts, consistency is key" adds no information gain. A named expert's specific, documented, unpublished opinion does.
- Ignoring the session context — Google evaluates gain relative to what a user has already seen in a session. If your page is result #5, you need to add something results #1–4 don't say.
- Treating information gain as one-time work — competing pages update. Your information gain advantage erodes as others catch up. Refresh with new data regularly.
Frequently asked questions
Google filed a patent for information gain scoring in 2020. Their Helpful Content Update explicitly rewards content providing "a unique or at least original" perspective. While the exact patent implementation is unconfirmed, the principle appears embedded in Google's ranking philosophy and aligns with observable ranking patterns.
Include original data, proprietary research, expert quotes, unique case studies, or novel analytical frameworks. Documenting your own test results counts. The essential element: state something no other ranking page currently expresses. A single genuine data point can outweigh thousands of words of repackaged content.
Not directly. A 500-word article with genuine unique insight outperforms a 3,000-word article repeating common knowledge. Length helps only when every added section contributes genuinely new value. Google measures novelty per statement, not words per page.
Rephrasing existing points, adding extra subheadings without new substance, padding word count, or replicating what the top 10 results already say. Information gain requires genuine novelty — a claim, number, or perspective that doesn't appear in competing pages.
They are complementary signals. E-E-A-T signals who created the content and whether they have real experience. Information gain signals whether the content itself adds something new. The strongest pages demonstrate both: an experienced author sharing genuinely original insights unavailable elsewhere.
Related glossary terms
Sources
