If your article says the same as the top 10, why would Google or an LLM cite it?
Information Gain compares your content sentence by sentence with the pages already ranking for your keyword and measures what percentage adds new information. It shows which passages are repeated, what your competitors cover that you do not, and which data only you have.
Information Gain Score
46/100
Rewriting the top 10 in other words is not new content.
Search engines and generative engines reward sources that add something. If your page only summarizes what already exists, there is no reason to pick it.
How content is usually written
- •Open the top 5 results and "merge" them into a single article.
- •Content tools push you to cover the same terms as everyone else.
- •Nobody measures how much of the final text is actually information already in the SERP.
What LLMFY Information Gain does
- Measures originality with real embeddings, sentence by sentence, against your SERP competitors.
- Flags passages that are near-identical to a competitor, with the matching domain.
- Separates what you are missing (gaps) from what sets you apart (unique data) so you know what to expand and what to protect.
How it is calculated
A deterministic, explainable method. No black boxes.
Your URL + your keyword
Pick country and language. If you prefer, list competitor URLs manually instead of using the SERP.
Competitor cohort
We take the organic top 20, keep up to 8 distinct domains and discard badly extracted pages (cookie walls, bot protection).
Sentence by sentence
We split your content and the cohort into sentences, generate embeddings and find the most similar competitor sentence for each of yours.
Score + insights
Your score is the % of original sentences. On top: duplicated passages, gaps, unique data and the questions nobody answers.
Every sentence, in one of three classes
The maximum cosine similarity of each sentence against the whole cohort decides its class. So you know exactly where your score comes from.
Original
similarity < 0.70Adds information that is not in the SERP. This is what counts towards the score.
"In our test with 214 stores, a one-page checkout lifted conversion by 11%."
Shared
0.70 – 0.85Common topic, different wording. Needed for context, but it does not set you apart.
"Reducing checkout steps usually improves the conversion rate."
Duplicative
similarity ≥ 0.85Practically the same sentence as a competitor. A candidate to rewrite or remove.
"Checkout is the final purchase process in an online store."
Reading the score
70 – 100
Highly original
40 – 69
Moderately original
0 – 39
Mostly shared
Everything you get in each analysis
A score and, above all, what to do with it.
Information Gain Score 0–100
Percentage of sentences that add new information compared to the SERP.
Per-sentence distribution
How many sentences are original, shared or duplicative, with their percentage.
Duplicative passages
The most copied-looking sentences, with the competitor domain they match.
Content gaps
Ideas your competitors repeat that your page does not cover, grouped together.
Your unique data
Figures and data points that only appear on your page. Your competitive edge: protect it and surface it.
Uncovered questions
Topics no SERP result answers well: your chance to add something new.
Real SERP or manual cohort
Analyze against the organic top of the chosen country or against the URLs you pick.
Sentence-level detail
Every sentence with its similarity and class, to review the text precisely.
PDF export
Full downloadable report to share with writers or clients.
Frequently asked questions
Find out how much of your content is really yours
Get started with the Starter plan (50 credits for €49/mo). Each Information Gain analysis costs 1 credit.
Want to go further? Ampliaciones GEO checks your page against the sources ChatGPT, Gemini, Perplexity and Claude cite. Discover Ampliaciones GEO