Languages

Commodity Content / Information Gain Checker

Compare what your page says with 3–5 competitors. The deterministic first layer shows what appears restated, what remains lexically unique, and whether visible first-hand artifacts support an Experience proxy.

Built by Patrick Stox

Four different questions, four different axes

Each axis answers a different content question. Open the tool that matches the one you are asking.

Content GapMissing topics Content CompletenessMissing element formats Answer-ReadyCitability of what exists Commodity / Information GainNovelty of what exists

Pages to compare

Paste text or HTML for the most private path. When a paste is empty, the tool fetches that row’s public URL through the site’s bounded fetcher.

Your target page

Competitor 1

Competitor 2

Competitor 3

Competitor 4 (optional)

Competitor 5 (optional)

Checks run from our server; we fetch the URL you enter and don't keep the results. Pasted content stays in this browser. A URL is fetched only when its paste field is empty; the optional model pass sends only bounded extracted target claims after you request it. Anonymous run-level outcome counters may be used for aggregate research; URLs, domains, IPs, and identifiers are never included, and no statistic is released below 100 runs.

Method last reviewed Jul 18, 2026.

Feedback
Report a bug

Found something broken in Commodity Content / Information Gain Checker? Let us know what happened — this goes straight to a private triage queue, not a public list.

What will be sent 
 No tool inputs, uploads, pasted source, complete results, query parameters, or URL fragments are attached automatically. You can edit or remove the selected passage above. Browser and anti-abuse metadata is processed for spam prevention. 
  
 
  
 
 
  
 
 
   
  
 
 
Sample report
 
Example data — the same passage loaded by “Try an example”
 
The built-in example pastes one short passage as the target and the same passage as three competitors. Because every supplied page restates the same claims, the deterministic first layer reports the worst-case commodity signal:
 
     
  • Restated overlap — 100% of bounded target claims also appear across the supplied competitors.
    •  
  • Novelty — 0% of target claims are unique within the supplied corpus.
    •  
  • Experience share — 0% first-hand-with-artifacts (the sample carries no original numbers, described methods, or original media).
    •  
  • Commodity flag — Flagged, because overlap is high and artifact evidence is low.
    •  
 
With real, distinct pages the numbers spread out. Read the example as a demonstration of how each card behaves at the commodity extreme, not as a typical result.
 
 
 
How to use it
 
     
  1. Paste your target page (text or HTML) and three to five competitor pages. Leave a paste field empty and supply that row's public URL to use the bounded fetcher instead.
    2.  
  2. Choose Compare pages. The four score cards summarize restated overlap, novelty, the Experience proxy, and the commodity flag.
    3.  
  3. Read the Target claim evidence table and the competitor statements that may be more specific list to see where you restate others and where a competitor states something more precisely.
    4.  
  4. Optionally run the model-assisted Experience split. It classifies only the extracted target claims and is rejected unless every first-hand label carries a supported artifact type and an exact evidence quote.
    5.  
  5. Treat every candidate addition as a review cue. Verify each for accuracy, usefulness, and support before publishing.
    6.  
 
 
How it works
 
     
  1. Paste the target and three to five competitor pages, or supply public URLs for the bounded fetcher.
    2.  
  2. The browser extracts a bounded claim set and compares token overlap while preserving numeric evidence.
    3.  
  3. Visible original numbers, described methods, and original-media language support a separate Experience proxy.
    4.  
  4. Only high overlap combined with low artifact evidence creates the structured “add information gainInformation gain is how much new information a page adds beyond what a searcher has already seen in prior results on the same topic — novelty relative to the existing corpus, not general content quality. The term comes from a granted Google patent that Google has never confirmed using in live ranking.” task.
    5.  
  5. Review every candidate addition for accuracy, usefulness, and support before publishing.
    6.  
 
What the results mean
 
     
  • Restated overlap — the share of bounded target claims that also appear, lexically, across the supplied competitors. A high share means the page mostly says what every competitor already says.
    •  
  • Novelty — the share of target claims that are unique within the bounded supplied corpus. Unique is a review cue, not proof of originality, accuracy, or usefulness.
    •  
  • Experience share — an artifact-bound proxy for first-hand content: visible original numbers, a described method, or original-media language. It is explicitly not an E-E-A-T score; Google publishes no such score.
    •  
  • Commodity flag — raised when restated overlap reaches 65% while first-hand artifact evidence stays low. It flags “nothing only you can say,” not a ranking prediction.
    •  
  • “Add information gain” task — appears only when high overlap combines with low artifact evidence, prompting you to add something only you can say: your own data, your test, your before-and-after.
    •  
 
Features
 
     
  • Deterministic, repeatable lexical overlap that preserves numeric evidence — no model required for the first layer.
    •  
  • Unique-statement extraction across a bounded three-to-five competitor corpus.
    •  
  • Artifact-bound Experience proxy, corroborated (not overridden) by whole-page element detectors.
    •  
  • Optional, strictly validated model-assisted Experience split that rejects the whole response unless every label is evidence-backed.
    •  
  • A per-claim evidence table plus a “competitor states it more specifically” list.
    •  
  • Paste-first privacy: content is analyzed in your browser, and a URL is fetched only when its paste field is empty.
    •  
 
Limitations
 
     
  • Lexical, not semantic — paraphrased or reordered claims can lower the score even when the meaning overlaps.
    •  
  • Bounded corpus (three to five competitors with disclosed character and claim caps); results describe only the supplied pages.
    •  
  • Novelty is not originality or accuracy; a unique claim can still be wrong or unhelpful.
    •  
  • The Experience proxy is not an E-E-A-T score, and the commodity flag is not a ranking prediction.
    •  
  • The optional model pass sends only extracted target claims after you request it and adds no ranking claim.
    •  
 
Frequently asked questions
 
Is this a plagiarism checker?
No. It compares bounded claims using disclosed lexical overlap. It cannot establish authorship, publication order, copying, accuracy, or originality.
Does novelty prove information gain?
No. A statement can be lexically unique and still be inaccurate or unhelpful. Treat novelty as a review cue and verify every proposed addition.
What does the Experience share measure?
It is an artifact-bound proxy for visible original numbers, described methods, or original media. Google has no public E-E-A-T score, and this tool does not claim to reproduce one.
Where is my content processed?
Pasted content is analyzed in your browser. A public URL is sent to the bounded fetcher only when its paste field is empty. The optional model pass sends only extracted target claims after you request it.
 
 
Next stepAI Content Brief Generator — generate the corrected version.

People also explore
ToolCitation Gap CheckerToolAI Content AuditToolChunk TesterArticleInformation GainArticleContent FreshnessArticleSEO Audit Checklist
 
 
 
Feature requests for Commodity Content Checker
 

Upvote what you want most. New ideas can be submitted from the floating Feedback menu;
      requests appear here once approved, and the most-wanted rise to the top.

   
Loading…
  
  ➕ Request a feature   
New requests are reviewed before they appear here.
 
     
  
  
 
 
 
 
   
         
 
Where this tool helps
Common use cases
Benchmark a page against competitors
Compare one page with three to five competitors to see where its statements substantially overlap.
Find unique information
Surface claims and details that appear distinctive enough to deserve stronger placement or support.
Review the experience proxy
Inspect artifact-bound signs of first-hand experience without treating them as proof of authorship or quality.
Guide a differentiated rewrite
Use the overlap and unique-statement evidence to plan additions that contribute useful information rather than cosmetic wording changes.
Watch the full workflow
Commodity Content / Information Gain Checker walkthrough
Read the transcript

      
      

        
Commodity Content & Information Gain Checker

        

        
This tool compares one target page with three to five competitors using a deterministic first layer. It reports lexical restatement, uniqueness within the supplied corpus, an artifact-bound Experience proxy, and per-claim evidence. I’ll run the built-in example, explain each result, show the optional model boundary, cover privacy and limitations, and turn the findings into responsible editorial next steps.

      

      

        
Step 1

        

        
Use this checker when competing pages seem interchangeable, during content refreshes, or before publishing. It asks a different question from topic gaps, format completeness, or answer citability: what is novel in the supplied comparison?

      

      

        
Step 2

        

        
The axis map separates missing topics, missing element formats, citability, and novelty. This tool focuses on novelty and artifact evidence, so it should not replace the other specialized audits.

      

      

        
Step 3

        

        
Paste text or H-T-M-L for the most private path. If a paste is empty, its public U-R-L is fetched through the bounded endpoint. Choose a relevant, representative competitor set because every result is corpus-relative.

      

      

        
Step 4

        

        
Try an example loads the same short passage as the target and three competitors, then runs locally. This creates an intentionally extreme commodity result: every bounded target claim is restated and no first-hand artifact is present.

      

      

        
Step 5

        

        
The cards report restated overlap, novelty within this corpus, Experience share, and a threshold-based commodity flag. They are separate review cues, not Google scores or ranking predictions.

      

      

        
Step 6

        

        
Restated overlap is the share of bounded target claims meeting the disclosed lexical threshold. Novelty is the complement within the supplied corpus. Paraphrases can evade the measure, and unique wording can still be wrong or useless.

      

      

        
Step 7

        

        
The Experience proxy looks for visible original numbers, a described method, or original-media language. The commodity flag appears when overlap is high and artifact evidence is low. Neither reproduces E-E-A-T or proves content quality.

      

      

        
Step 8

        

        
The claim table shows classification, best supplied match, overlap, shared terms, and artifact evidence. Use it to locate the exact statements behind the summary instead of editing from percentages alone.

      

      

        
Step 9

        

        
The richer-statement list highlights related competitor claims that may be more specific, while whole-page element detectors provide corroboration. These are prompts to investigate, never text to copy or unsupported facts to add.

      

      

        
Step 10

        

        
The optional model task sends only extracted target claims after an explicit request. The entire response is rejected unless every first-hand label has a supported artifact type and an exact quote. Not run means no model result is implied.

      

      

        
Step 11

        

        
The comparison is lexical, bounded to three to five supplied competitors, and subject to character and claim caps. Novelty is not authorship, accuracy, or usefulness. The Experience proxy is not E-E-A-T, and the flag does not predict rankings.

      

      

        
Step 12

        

        
Review restated claims and candidate gaps, then add only accurate, useful, supportable information: original data, a documented method, first-hand examples, or owned media. Cite sources, preserve important consensus, and rerun with a representative corpus.

      

      

        
Add something useful that only you can support.

        

        
Use the evidence table to identify restated claims and richer competitor statements, then decide what readers actually need. Add your own verified data, test, process, images, examples, or before-and-after evidence where appropriate. Recheck the page against a representative corpus, but never treat lexical novelty, the Experience proxy, or the commodity flag as proof of originality, accuracy, or ranking impact.

      

    
 
         
 
Common issues & how to fix them
 
     
  •  Warning High restated overlap with little artifact evidence Fix: Add a supportable first-party artifact such as original data, a documented test, or a before-and-after result, then rerun the same bounded competitor set. 
    •