Information Gain Auditor

Before you publish, see which paragraphs just repeat what already ranks — and which ones an AI answer would actually quote. Fix the echoes, keep what only you can say.

Coverage Map What the top covers Topical Map What your site covers

Information Gain What your draft adds

Commercial Audit What the buyer checks

Who are you up against?

The keyword — we'll pull the ranking articles. Or paste the URLs you want to beat.

We skip aggregators, socials and your own domain.

Your own questions (optional)

Scored alongside the questions we derive, marked as yours in the report.

Your article

A draft you haven't published yet is the best case — you can still fix it.

Keep the blank lines between paragraphs — each one is scored on its own. Min ~500 characters.

Cost: 35 credits per analysis

Your report

What to fix?Ordered by impact: the paragraphs that echo the competitors and carry no first-hand experience come first — they're the cheapest wins. Each one comes with a rewritten version you can paste straight in.

Format fit?Whether your page looks like what Google already ranks here — genre, length, media, FAQ, author. A guide rarely breaks into a SERP made of listicles, however good its content. Scored separately from gain on purpose: format is fixed first.

do you look like what ranks here?

Paragraph heatmap?Every paragraph scored against every passage of the ranking pages. Click any paragraph to see the closest competitor text, borrowed wording, and why it was flagged.

You vs the competitors?Every page scored the same way: against all the others. A score of 9 means little on its own — it means a lot next to competitors scoring 10 and 33. Gain and rank are different things: the top-ranking page often adds the least.

same scale

AI citability?Whether an AI answer engine could quote you. It lifts passages, not pages — so it wants self-contained paragraphs, hard numbers and named sources, not "experts say".

Topics you skipped?Themes most competing articles cover and yours doesn't. These are questions, not orders — if a topic is filler, ignore it. We never tell you to insert keywords.

covered by the competitors

Your strengths?Terms that appear in your text and nowhere in the competing articles. This is the source of your information gain — expand these rather than adding what everyone already has.

nobody else has these

Topic coverage?A third, independent axis. An LLM maps the branches a complete resource on this topic would address — independent of who currently ranks — then your passages are matched to each branch by embeddings. This is a map of gaps, not a word-count order: coverage ≠ length, and some branches are expert territory only you can judge.

How this was scored?The full method, including what we don't claim: Google has never confirmed information gain is a live ranking signal.

Every paragraph gets a verdict — and a rewrite

See which paragraphs repeat what already ranks and which carry something of your own. Every flagged paragraph comes with a rewrite you can paste — apply them and re-check for 10 credits instead of 35.

Would an AI answer quote you?

Know before publishing whether ChatGPT or Perplexity could lift a passage from your page — and whether the page even looks like what ranks here: genre, length, media, FAQ, author.

You and the competitors on one ruler

Your score means nothing alone and a lot next to theirs. You also get the terms only you use — the thing to expand — and the topics you skipped, as questions, never as orders.

See it on a real draft

Information Gain Auditor — originality score, per-paragraph verdicts, ready-to-paste rewrites and the competing articles you were measured against

What we don't do

×Don't count keyword density

We never tell you to repeat a phrase four more times. We compare meaning, not word counts — "budget housing in the suburbs" and "affordable property outside the city" are the same idea, and we treat them that way.

×Don't ask you to echo the SERP

A topic the whole top 10 covers isn't automatically worth covering. Missing themes arrive as a question — do you have something to say here? — not as an order.

×Don't score the article as one blob

Average a 3,000-word article into a single vector and your best paragraph dissolves into the filler around it. Every paragraph is scored on its own.

×Don't accuse your text of being AI

AI detectors are unreliable, and a false accusation is worse than no verdict. We only report what's checkable: first-person experience, named cases, real numbers.