High-Quality Content: Improve Thin Content, Fact Density and Source Quality

Many websites do not have too little text. They have too little substance.

An article can be 2,000 words long and still add hardly anything of its own. A short page can still give exactly the right answer. Thin contentThin content is content with too little independent value. A page is thin when it fails to satisfy a search intent adequately or adds nothing substantial to existing pages.Open glossary entry → is therefore not purely a word-count question.

RankScan has three measures on this topic. Each one measures a clear, simple signal:

• “Little text in the main content – check substance”: the main content of the page has fewer than 300 words. Navigation and footer do not count. This is a prompt to check, not a quality verdict.
• “Too few concrete details in the text”: the main text names fewer than 0.5 concrete details per 100 words, such as numbers with a unit or years.
• “No links to external sources”: an expert article with at least 300 words on the page does not link to a single external website in its main content. The rule does not check the quality of the source.

All three signals are prompts to check. Deliberately short pages can serve their purpose. Where content is missing, all three measures ask for the same thing: expand the pages. Add information, concrete details and evidence. This article explains the measurement, the procedure and the expertiseExpertise is the subject-matter competence behind a piece of content. It is the second E in E-E-A-T and should be visible through author details and supporting evidence.Open glossary entry → behind it.

The right framing matters:

High-quality contentHigh-quality content comes from better information value, not from more words. It fully answers the search intent, provides concrete facts, shows expertise and supports key statements with sources.Open glossary entry → is not created by more words, but by better information value.

What Is Thin Content? #

Thin content refers to content with little unique value. A page is thin if it does not sufficiently answer a search intentSearch intent is the goal behind a search query: does the user want to learn something, buy something or reach a specific page? Content only ranks well if it matches this intent.Open glossary entry →, provides hardly any concrete information or adds nothing significant to existing content.

Google does not use the term “thin content” as its own category in its spam policies. Related practices are listed there, though, such as “scaled content abuse”: many pages created mainly to manipulate rankings. There are also “doorway abuse” and “thin affiliation”, meaning affiliate pages without added value of their own.
Source: Google Search Central: Spam Policies for Google Web Search

Thin content can take many forms:

• very short pages without a clear answer,
• long texts with a lot of filler,
• copied manufacturer texts,
• automatically generated mass content,
• almost identical location pages,
• affiliate pages without their own review,
• blog articles without a new perspective,
• guides without sources,
• glossary entries without context,
• AI texts without editorial review.

Thin Content Is Not a Word-Count Problem #

A common mistake is the assumption:

few words = thin content
many words = high-quality content

That is not true.

A short answer can be high quality if it is precise. A long article can be thin if it only contains repetition, empty phrases and commonly known statements.

Google says so explicitly. Its guide to helpful contentHelpful content refers to content created primarily for people that delivers a clear answer and shows substance, originality and trust, rather than being optimized mainly for search engines.Open glossary entry → asks whether you write to a particular word count because you heard Google has a preferred word count. The answer there: “No, we don't.”
Source: Google Search Central: Creating helpful, reliable, people-first content

So why does RankScan count words? The word count is a simple trigger, not a quality verdict. A page with fewer than 300 words rarely explains a topic completely. Whether it is still enough, you decide when you read it.

Example:

Weak:
Content quality is very important because good content is rated better by Google and convinces users.

Stronger:
Content quality emerges when a page fully answers the search intent, provides concrete facts, shows recognizable expertise and supports key statements with reliable sources.

The second sentence is not much longer, but it is much more concrete.

What Does “Little text in the main content – check substance” Mean? #

The RankScan measure “Little text in the main content – check substance” means: RankScan counts fewer than 300 words in the main content of the affected pages. The title names the measurement. RankScan only checks the word count and makes no judgement on content depth.

Few words are a prompt to check. Often topic, benefit or context then remain incomplete. But deliberately short pages are legitimate, such as a concise answer page or a form with a short introduction. Whether the page is enough, you decide when you read it.

Rating in RankScan: medium severity · high effort · work type Editorial.

What RankScan measures #

• RankScan counts the words in the main content. The main content is the <main> area or an element with role="main". If both are missing, RankScan counts the page without navigation, header, footer and sidebar.
• Below 300 words, the page counts as affected.
• Only page types that need running text count: home page, posts, product and service pages, FAQ pages and pages without a detected type.
• Legal, contact, category, confirmation and functional pages such as a gallery do not count. Internal search and error pages do not count either.
• If RankScan does not detect such a page, you remove it from the list yourself with “Skip”.
• RankScan does not check pages with noindex for this measure.

Procedure #

1. First check whether the page should appear in search results. Otherwise, skip it, for example forms, galleries, overviews or reviews.
2. For each such page, choose “Skip” under “Affected pages”. From the next scan on, it no longer counts.
3. Open each remaining page. Check whether it is deliberately short and serves its purpose. Otherwise, note which questions about the topic remain open.
4. Add information, definitions, examples or benefits that answer these questions.
5. Add common questions as a short FAQ section where it fits.
6. Publish the page and mark the measure as done.

How RankScan checks it: The next full crawl counts the words of the affected pages again. RankScan then reports “Fixed” or “Not fixed yet” with the number of pages still missing.

Variant: references #

If at least half of the affected pages sit under a path such as /referenzen/, /references/ or /projects/, the measure names the goal “Expand … pages with little text, mostly references”. Reference pages often consist only of images and a project name. The steps then name what makes a reference meaningful:

1. First check whether the page should appear in search results. Otherwise, skip it, for example forms, galleries, overviews or reviews.
2. For each such page, choose “Skip” under “Affected pages”. From the next scan on, it no longer counts.
3. Open each remaining reference and note what is still missing about the project.
4. Describe the starting point: who is the client, and what was the goal?
5. Name the particular challenge and the chosen solution.
6. Add concrete details such as materials, construction method, floor area or construction time.
7. Add a short client quote if the client agrees.
8. Publish the page and mark the measure as done.

The overview page of the references does not count as a reference itself. Concrete details such as floor area and construction time also help fact density.

What Does “Too few concrete details in the text” Mean? #

Fact density is not an official Google metric. RankScan uses it as an editorial criterion.

The RankScan measure “Too few concrete details in the text” means: the main text of the affected pages names hardly any figures, dimensions, durations or dates. Menu and footer do not count.

Rating in RankScan: low severity · medium effort · work type Editorial.

What RankScan measures #

• RankScan reads only the main content, i.e. the <main> area or an element with role="main". If both are missing, RankScan reads the page without navigation, header, footer and sidebar.
• The main content must have at least 300 words. This measure does not rate shorter pages.
• A page with fewer than 0.5 concrete details per 100 words is affected. At 600 words, that means fewer than three details.
• Contact, legal, confirmation and functional pages do not count. The home page counts.

RankScan counts as a concrete detail:

• numbers with a unit or currency, such as “120 m²”, “15 minutes”, “3 weeks”, “12 km”, “20 %” or “CHF 450”,
• other numbers with two or more digits, such as a year in “since 1923”,
• definition sentences at the start of a sentence following the pattern “Cross-laminated timber is a …”.

These do not count as a detail:

• phone and article numbers, and numbers with seven or more digits in a row,
• single digits without a unit,
• bold or coloured words without a number,
• sentences such as “Our solution is the best.” or “This is a test.”. They start with a pronoun and do not count as a definition.

Batch of the most important pages #

If many pages are affected, RankScan shows a batch: “Add concrete details to the 10 most important of … pages”. The most important pages are at the top. The page list shows how many details fall on how many words for each page.

Procedure #

1. Work through the list from top to bottom. The most important pages are at the top.
2. Mark general statements such as “many years of experience”, “many projects” or “central location”.
3. Replace them with details that carry a number and fit the page.
4. Service page: year founded, number of projects, duration in days or weeks, area in m².
5. Technical article: measured values, percentages and the year of a standard or study.
6. Surroundings and location: walking or driving time in minutes, distance in kilometres.
7. Job page: workload in percent, start date, duration and number of holiday days.
8. Publish the pages and mark the measure as done.

How RankScan checks it: The next full crawl counts details and words of the affected pages again. After the check, the page list shows the result.

High fact density does not mean overloading every sentence with numbers. Every number must be correct and fit the page. Do not invent values just to reach the threshold.

What Does “No links to external sources” Mean? #

The RankScan measure “No links to external sources” means: RankScan found no link to an external website in the main content of the affected expert articles. The rule is met as soon as any external link appears in the main content. It does not check the quality of the source. So the finding does not say that a claim is unsupported. It only says that no external link was found.

Rating in RankScan: medium severity · medium effort · work type Editorial.

What RankScan measures #

• The measure only applies to pages that RankScan detects as a post or expert article. Service pages do not count. There, references are the better proof of competence.
• The page has at least 300 words on it.
• RankScan looks for a link to an external website in the main content. Main content means <main>, otherwise <article>, otherwise the whole <body>. Links in navigation, header, footer and sidebar do not count.
• Links to your own subdomains, anchors, email and phone links do not count as a source.
• If no such link exists, the article is affected.

RankScan does not check whether a source is good. A single external link is enough for the measurement. Whether the source supports the claim is up to you to check. The sections on source qualitySource quality shows whether important statements are backed by reliable evidence. It does not mean adding links indiscriminately, but providing robust support for verifiable claims.Open glossary entry → below help with that.

Procedure #

1. Skip pages without statements that need evidence, for example your own news or press releases.
2. Mark the statements on the remaining pages that need evidence: figures, standards, studies.
3. Find studies, official sources or reliable expert references for them.
4. Link the source right next to the statement.
5. Publish the pages and mark the measure as done.

How RankScan checks it: The next full crawl checks the links of the affected articles. RankScan then reports “Fixed” or “Not fixed yet” with the number of pages still missing.

For trustworthy content, Google mentions “clear sourcing” and evidence of the expertise involved, among other things.
Source: Google Search Central: Creating helpful, reliable, people-first content

Why High-Quality Content Matters for SEO and AI Visibility #

Search engines and AI systems evaluate content differently, but both benefit from clear, reliable and structured information. This applies to classic search engine optimization (SEO)Search Engine Optimization (SEO) covers all measures that help a website become more visible in unpaid (organic) search results.Open glossary entry → and to AI visibility alike.

For classic SEO, high-quality content helps because it:

• better fulfills search intent,
• can earn more internal and external links,
• enables better snippetsA snippet is the compact display of a result in the search results, usually made up of a title, URL and description. It is what users see before clicking.Open glossary entry →,
• builds more topical relevance,
• strengthens trust.

For AI visibilityAI Visibility describes whether and how visible a brand, website or person is in AI-generated answers. It measures mention, position, description and source linking across multiple platforms and prompts.Open glossary entry →, it helps because AI systems often retrieve, summarize and cite content section by section. Passages are especially suitable when they:

• answer a clear question,
• contain concrete definitions,
• present facts in a structured way,
• name sources,
• contain few empty phrases,
• provide independent information.

According to Google, the same SEO best practices apply to AI OverviewsAI Overviews are AI-generated answer boxes from Google that appear above the classic search results. They summarize information from the web and can change click distribution.Open glossary entry → and AI Mode as to the rest of Search. There are no additional requirements or special optimizations.
Source: Google Search Central: AI features and your website

Important: high-quality content does not guarantee a mention in AI answers. But it improves the conditions for being considered an understandable and trustworthy source.

Increase Fact Density: From Empty Phrases to Citeable Statements #

Fact density emerges when statements are concrete, verifiable and useful. The examples on the right contain details that RankScan counts.

The numbers in the table are examples. On your page, only use values that are true for your offer.

A simple rule:

Delete every statement that sounds interchangeable, or make it more concrete.

Source Quality: Which Sources Are Useful? #

Not every source has the same value. The following table is a general professional assessment. RankScan does not rate the quality of a source.

For important statements:

If the statement must be verifiable, it needs a reliable source.

Which Statements Need Sources? #

Sources are especially important for:

• numbers,
• percentages,
• trends,
• studies,
• benchmarks,
• legal statements,
• medical statements,
• financial statements,
• technical standards,
• statements about Google updates,
• statements about AI systems,
• market comparisons,
• historical developments,
• product or tool comparisons.

An external source is usually not needed for:

• your own recommendations,
• clearly labeled experience values,
• internal processes,
• examples,
• checklists,
• simple definitions,
• editorial assessments.

That is why the measure's procedure starts with skipping: your own news or press releases usually need no external evidence.

High-Quality Content by Page Type #

Not every page type needs the same content. The following table is general expertise. It helps you decide what to expand an affected page with.

A good product page needs different substance than an expert guide.

Assessment: Which Content Problems Weigh Heavily? #

The following table is a general professional assessment. RankScan shows a single severity per measure: medium for “Little text in the main content – check substance” and “No links to external sources”, low for “Too few concrete details in the text”.

The most important rule:

Start with pages that have visibility potential, business relevance or trust risk.

Improve Content: Added Value Instead of More Words #

All three measures ask you to expand the page. The cause decides what you expand it with.

Expand Instead of Hiding #

Hiding a thin page with noindex or a canonical does not solve the quality problem. That is why the measure asks you to expand.

A special case is many almost identical pages on the same topic:

/seo-consulting-lucerne
/seo-consulting-zug
/seo-consulting-zurich
/seo-consulting-basel

Here, too, expanding is the first path:

• add real local information,
• show references for each location,
• name contact persons or location data,
• give travel time in minutes or distance in kilometres.

General expertise, not part of the measure: If a page has no purpose of its own and only repeats another page, merging it with a 301 redirectHTTP status 301 signals a permanent redirect: a URL has been permanently replaced by another. Search engines transfer the signals to the target URL.Open glossary entry → is a common SEO path. That is a separate decision about the structure of your website. Make it deliberately, not to close a measure.

What to Do After a RankScan Finding #

Each of the three measures has its own procedure above. If several measures hit the same page, this order helps.

Step 1: Read the measure and its variant #

Read the goal and steps of the measure in RankScan. If the goal says “mostly references”, the procedure for references applies. If it says “the 10 most important”, you work through a batch from top to bottom.

Step 2: Read the page list #

The page list names the affected pages. For fact density, it also shows how many details fall on how many words. This shows you how far a page is from the threshold.

If a page should not appear in search results, for example a form, choose “Skip” for it. From the next scan on, it no longer counts.

Step 3: Note open questions #

Open the page. Note which questions about the topic remain open. General expertise helps here:

• What does a visitor expect on this page?
• Which page types rank for the topic?
• Which information is missing on our page?

Step 4: Add substance #

First add the content that answers questions. Then replace general statements with details that carry a number. In expert articles, link sources right next to the statement. One text often covers several measures at once this way.

Step 5: Publish and mark as done #

Publish the page. Then mark the measure as done in RankScan. The next full crawl checks the affected pages. RankScan then reports “Fixed” or “Not fixed yet” with the number of pages still missing.

What RankScan Measures and What It Does Not #

RankScan measures three simple signals on this topic:

Much of what makes good content is not measured by RankScan in these measures. The following list is a general checklist for your own review:

• search intent and page type in the search results,
• quality and freshnessFreshness describes how current a piece of content is relative to the search intent. It is not equally important for every query.Open glossary entry → of sources,
• overlap with your other pages,
• own experience and own data,
• impressionsImpressions indicate how often a website's search result was shown in Google Search, regardless of whether it was clicked. They are a core Google Search Console metric and reflect a page's raw visibility for given queries.Open glossary entry → and clicks in Google Search ConsoleGoogle Search Console (GSC) is a free Google tool that shows how a website performs in organic search, with data on impressions, clicks, click-through rate and average position, plus indexing and technical issues.Open glossary entry →,
• whether a number in the text is correct.

RankScan has separate measures for neighbouring topics, such as freshness, summaries, authors and trust signalsTrust signals are visible elements of a website that build credibility, such as an imprint, contact details, privacy information, sources, reviews, update dates and transparent company information.Open glossary entry →. Learn more in content updates, summaries and key takeaways and E-E-A-T.

Example: Expanding References With Little Text #

Initial Situation #

A timber construction company shows its projects under /referenzen/. Each reference consists of an image gallery, the project name and the location.

RankScan reports:

“Little text in the main content – check substance”: … pages with little text to expand, mostly references
“Too few concrete details in the text”

Analysis #

• At least half of the affected pages sit under /referenzen/. That is why RankScan shows the references variant.
• The references do not say what makes the project special.
• Service pages speak of “many years of experience” and “many projects”, without a number.

Solution #

1. Describe the starting point, challenge and solution for each reference.
2. Name materials, construction method, floor area and construction time.
3. Add a short client quote if the client agrees.
4. On the service pages, name the year founded, number of projects and typical duration.
5. Publish the pages and mark both measures as done.

Result #

The references now prove the company's competence. Service pages name verifiable details instead of empty phrases. This does not guarantee better rankings. But it creates much better conditions for classic SEO and AI visibility.

Common Mistakes With High-Quality Content #

Mistake 1: Treating Word Count as the Goal #

More words do not automatically mean better content. The threshold of 300 words is a trigger, not a goal.

Mistake 2: Merely Retelling the Top 10 #

If you only summarize existing rankings, no information gain is created.

Mistake 3: Publishing AI Texts Without Review #

Automated content without editing, an own perspective and added value can be problematic.

Mistake 4: Leaving Out Sources #

Especially for numbers, trends, legal, financial or health topics, this weakens credibility.

Mistake 5: Only Changing the Date on Outdated Content #

A new date without a real update is not a quality improvement.

Mistake 6: Hiding Thin Content With Canonical or Noindex #

A page with noindex drops out of the “Little text in the main content – check substance” measurement. But the page does not get better. Expand it, as the measure asks.

Mistake 7: Inventing or Merely Highlighting Numbers #

Bold words do not count as a detail. Invented numbers damage trust. Only name values that are true.

Checklist: Review High-Quality Content #

This checklist is general. RankScan measures only the marked items in the three measures.

• Which search intent should the page fulfill?
• Does the page answer the open questions on the topic?
• Does the page have at least 300 words on it? (measured by RankScan)
• Are there concrete definitions and examples?
• Does the main text name details with a number, at least 0.5 per 100 words? (measured by RankScan)
• Does an expert article link at least one external source? (measured by RankScan)
• Do the sources really support the statements?
• Is there own experience or own data?
• Is the content up to date?
• Are author and trust signals visible?
• Is there a clear H2 and H3 structure?
• Are there internal links to suitable pages?
• Has the measure been marked as done after publishing?

In addition, content updates, summaries and key takeaways, Entity SEO and Schema Markup help narrow down the cause and choose the next SEO measures.

Conclusion: High-Quality Content Is Documented Substance #

Thin content is not a problem of texts that are too short. It is a problem of missing substance.

The RankScan measures “Little text in the main content – check substance”, “Too few concrete details in the text” and “No links to external sources” use simple thresholds to show where you can start. What is good, you decide while expanding.

The best approach is:

1. read the measure, its variant and the page list,
2. note open questions about the page,
3. add information, definitions and examples,
4. replace empty phrases with details that carry a number,
5. support strong statements in expert articles with sources,
6. publish the pages and mark the measure as done.

The next full crawl checks the pages. This creates content that is not just longer, but better: more concrete, more trustworthy, more useful and more citeable.

Sources and Further Reading #

• Google Search Central: Creating helpful, reliable, people-first content
• Google Search Central: Spam Policies for Google Web Search
• Google Search Central: AI features and your website
• Google Search Quality Rater Guidelines