Content analyzer
Helpful-content scores, a guideline checklist, YMYL, page type and search intent for every crawled page.
The content analyzer scores every crawled page against Google's guidance on helpful content. It tells you how strong a page is, which guidelines it meets or misses, what kind of page it is and who it is for. Scans run automatically after a page is crawled.
Where to find it
- On a page: open a page from URLs and choose the Helpful content tab. A short summary also appears in the page's sidebar, with a link to the details.
- In the URL list: turn on the Helpful content, Page type and Search intent columns, plus a column for each score. Sort by Helpful content to find your weakest pages, and filter by page type or search intent.
- In exports: the scores, page type, search intent and YMYL can be included when you export URLs to CSV.
- In chat: ask which pages in a section or page type score lowest. The assistant ranks them by overall score and reads their scans to explain why.
Helpful-content scores
Each page gets an overall score out of 5 and a score for each dimension:
| Dimension | What it looks at |
|---|---|
| Originality | Original information, research or perspective, not a rewrite of other sites |
| Content effort | How much care and work went into the page |
| Helpfulness | Whether a reader can reach their goal from this page |
| Talent or skill | Whether the writing, method or tool shows the skill the page needs |
| Accuracy | Factual accuracy, and consistency with expert consensus on YMYL topics |
| Trust | Sourcing, expertise and author background that give a reason to trust the page |
| People-first | Whether the page is made for people or mainly for search visits |
| Spam | Visible spam techniques. Higher means cleaner. |
Most dimensions use the levels Weak, Limited, Fair, Strong and Excellent. People-first runs from Search-engine-first to Human-first, and Spam from Spam to Clean. Each level comes with a sentence that explains what it means for this page.
People-first is shown on the scan but marked Not in overall for now, so it does not change the overall score.
If the page's text does not show enough to answer a dimension's questions, Morrison says Not enough to judge instead of guessing. Low-confidence results are labeled as such.
The checklist
Behind the scores is a checklist of Google's own questions, such as Covers the topic substantially, Title describes the content, Shows first-hand expertise and depth and No easily-verified factual errors. Each question shows whether the page meets it:
- Met
- Partly met
- Missing
- No sign of a problem
- Not relevant for this page
The checklist tells you exactly what to fix, not just how the page scored.
YMYL
Morrison flags whether a page is YMYL ("your money or your life"): content that could affect someone's health, finances, safety or well-being. On YMYL pages, accuracy is judged against expert consensus and Trust counts twice in the overall score.
Page type and search intent
Every page also gets a page type and a search intent, each with a confidence.
Page types: Pillar guide, Support article, Product or service, Comparison or review, Glossary or definition, Tool or calculator, News or update, Category hub, Home, About or company, Legal or utility, and Other.
Search intents: Informational, Commercial, Transactional and Navigational.
Use them to filter the URL list, build segments, or check that a page's content matches the intent it targets.
AI suggestions
Click Get AI suggestions on a page's helpful-content view. Morrison opens a new chat with a prepared message about the page, ready to send. The assistant reads the page and its scan and suggests concrete improvements. You can keep going from there, for example by asking it to draft a new section in the canvas.
Compare with competitors
Ask in chat to score a competitor's page, for example "Score the top three results for 'invoice software' and compare them with our page". Morrison runs the same helpful-content scan on any public page, so you can see where yours falls short. Each outside page uses one indexed page.
Rescanning
Scans run after each crawl. To score a page again, re-scrape it from the Helpful content tab. If a scan fails, its text is kept, so Try again does not need a new crawl. When a newer scan fails, Morrison keeps showing the previous scores.
Turning scans off
Content scans are on for new websites. An organization admin can turn them off per website in Website settings under Content scans.
Each scan reads the page's title, meta description and main text, up to about 40,000 characters, which covers even long guides. Turning scans off stops new scores. Your pages are still crawled, and earlier scores stay until a page is deleted.
Related use cases
- E-E-A-T Content Assessment: assess content quality and trust signals
- Thin Content Identification: find low-value pages
- Search Intent Alignment: check that pages match what searchers want