Skip to content
Learn

Content quality and the content lifecycle

In this lesson

Section titled “In this lesson”
  • Explain what E-E-A-T is, why trust is its most important part, and why it is not one ranking factor.
  • Add information gain to a page with primary data: research, surveys, your own datasets, case studies and custom visuals.
  • Find content decay, and choose between a refresh, a merge, a redirect, noindex or removal.
  • Tell normal duplicate content from scraping, which is spam.
  • Use AI writing tools and content optimizers without making scaled, low-value pages.

Day 04b made one page clear. This lesson asks a harder question: is the page worth reading at all? Then it follows the page over its life.

How Google judges content quality

Section titled “How Google judges content quality”

Google’s automated ranking systems try to rank “helpful, reliable information that’s created to benefit people”. Google calls this people-first content. The opposite is search engine-first content: pages made mainly to get search traffic.

Google gives a list of self-assessment questions. Some examples:

  • Does the content give original information, reporting, research or analysis?
  • Does it give “substantial value when compared to other pages in search results”?
  • If it uses other sources, does it add value, or does it only rewrite them?
  • Will a reader leave with enough information to reach their goal?

Google also lists warning signs. Two of them are important for this lesson. You change “the date of pages to make them seem fresh” with no real change. You add or remove a lot of content to make the site “seem fresh”. Google’s answer: “No, it won’t.”

The main content is the part of a page that helps the page do its job. In October 2026, Google’s guide names four things that its quality raters check in it:

Attribute What it means
Effort How much human work went into the content. A custom tool is high effort. AI text with no review is “little to no effort”.
Originality Information or views that are not on other sites already.
Talent or skill Clear writing, good video, tools that work. Not all topics need an expert.
Accuracy Facts are correct. For important topics, they agree with expert consensus.

So what matters most? Original value and trust come first. The tools and the lifecycle work in this lesson only help if the content itself is good.

E-E-A-T

Section titled “E-E-A-T”

E-E-A-T means experience, expertise, authoritativeness and trustworthiness. Google added the first E, experience, in December 2022. Experience means first-hand knowledge: you used the product, or you visited the place.

Part Question
Experience Did the creator use, do or visit the thing?
Expertise Does the creator know the topic well?
Authoritativeness Do others see the creator or site as a go-to source?
Trustworthiness Is the page accurate, honest and safe?

Three facts from Google keep E-E-A-T in proportion:

  1. “Of these aspects, trust is most important.” The others support trust. A page does not need all four.
  2. “E-E-A-T itself isn’t a specific ranking factor.” Google’s systems use a mix of other signals that agree with good E-E-A-T.
  3. Google gives E-E-A-T more weight for YMYL (“Your Money or Your Life”) topics. These topics can affect health, money, safety or society.

Search quality raters are people who check if Google’s results are good. They use the Search Quality Rater Guidelines. Google says that raters “have no control over how pages rank”. Their data is feedback, like comment cards in a restaurant. Read the guidelines to self-assess, not to find ranking tricks.

How do you show E-E-A-T? Google suggests the “Who, How and Why” questions:

  • Who made it? Add a byline that links to an author page. Do not invent authors. Google calls AI headshots and false credentials “a form of deception”.
  • How was it made? Show your method. For a product review, say how many products you tested and show photos.
  • Why was it made? Google calls this “perhaps the most important question”. The answer must be “to help people”, not “to get search traffic”.

Information gain

Section titled “Information gain”

Information gain is what a page adds beyond the pages that a searcher already saw. A page that repeats the top results gives the searcher nothing new.

The term comes from a Google patent, “Contextual estimation of link information gain”, granted in June 2022. It describes a score for “additional information” in a document “beyond information contained in other documents that were already presented to the user”. But a patent is not a product. Google has never confirmed that it uses an information gain score. Ahrefs, which wrote about the patent, says the same.

You do not need the patent to act on the idea. Google’s own questions ask for “original information” and “substantial value when compared to other pages”. The roadmap calls this “information gain”. Google calls it “originality”. It is the same advice.

Your draft about descaling an espresso machine includes

Example data, not real results. Turn facts on and off in your draft. Notice that the facts every result already has add nothing new, and that the new facts all come from your own work.

Primary data and visuals

Section titled “Primary data and visuals”

Primary data is data that you collect yourself. It is the most reliable way to add information gain, because nobody else has it. Secondary data is data from someone else’s report. Its value is in your analysis, not in the numbers.

Type What it is Effort Example
Original research and surveys You run a study or ask people questions. High A survey of 1,200 espresso machine owners.
Proprietary datasets Data only you have: usage, sales, support tickets. Medium, if you have the data “Our app logged 2 million brews. Here is when people brew.”
Case studies One project, with the problem, the work and the result. Medium How one café cut waste by 20%.
Benchmarks Your results compared with averages or with competitors. Medium Load times of 50 coffee shop sites.
Custom charts and images Visuals that you make from your data, or photos you take. Low to medium A chart of your survey results.

The example numbers in the table are invented, to show the type.

Original research has a second effect: other sites cite it. Backlinko explains that journalists and bloggers need data to support their points, so a good study earns links. Day 06 covers links.

Do these things to make primary data useful:

  • Explain your method: sample size, dates and questions. This is the “How” of E-E-A-T.
  • Put the main numbers in text or in an HTML table, not only in an image. Google uses the text near an image to understand it, as day 04b showed.
  • Give every chart alt text that states the main finding.
  • Use your own photos instead of stock images. Google’s guide gives photos as evidence of real testing.

Small items need not be expensive. A tested recipe with your own photos has more information gain than a long summary of other recipes.

The content lifecycle

Section titled “The content lifecycle”

The content lifecycle is the life of a page: plan, write, publish, measure, maintain, and at the end merge or remove. The roadmap item for this has the wrong text in October 2026, so this section uses Ahrefs’ model. Ahrefs names five traffic stages after publication: early traction, growth, traffic peak, slow plateau and decline.

Ahrefs says that most teams “invest heavily in the first three phases and almost nothing in the last two”. The last two are where you lose traffic. So a content plan needs a maintenance step, for example a review of the top pages every three months.

Freshness and decay

Section titled “Freshness and decay”

Content freshness is how current a page is. It matters only for some queries. Google has “query deserves freshness” systems. They show fresh content “for queries where it would be expected”. A search for a new film needs new reviews. A search for “how to boil an egg” does not.

Content decay is the slow loss of rankings and traffic over months or years. It is not the same as a sudden drop after an update, which day 02 covered. Ahrefs names these causes:

  • Better competitors: someone publishes a better page. Ahrefs calls this the most common cause.
  • Intent shift: the meaning of the query changes. Ahrefs’ example: “LLM” once meant “Master of Laws”.
  • Age: for queries that expect fresh content, newer pages win. Prices, versions and examples go out of date.
  • Cannibalization: two of your own pages compete for the same query and split the signals.

To find decay, compare a page’s clicks and impressions in Google Search Console with the same period a year ago. Day 08 teaches Search Console.

How do you refresh correctly? Update the facts, examples and missing subtopics. Keep the URL, so that the page keeps its links. Then show a “Last updated” date. Google’s byline date guide asks for a visible, labelled date, and dateModified in structured data. Do not change the date if the content did not change.

At month 24, the team

Example data for one article, not a forecast. Move the slider through the stages. Then pick an action at month 24 and move past it. Notice that a new date alone changes nothing, but a real refresh or a merge does.

Duplicate content

Section titled “Duplicate content”

Duplicate content is the same or almost the same content at more than one URL. Most duplicates are on one site and are not deliberate. Examples are HTTP and HTTPS versions, URLs with tracking parameters, and print pages.

There is no “duplicate content penalty” for this. Google said so in 2008. Its canonicalization guide still says that some duplicate content “is normal and it’s not a violation of Google’s spam policies”. Google groups the duplicates into a cluster and picks one canonical URL to show. Day 03 taught how to signal your choice. The real costs are smaller: split signals, wasted crawling, and confusing data.

Duplicates across sites are different. Scraping is taking content from other sites and republishing it “without adding any original content or value”. It is a spam policy violation. So is scaled content made with “synonymizing, translating” or other tricks.

Consolidation, pruning and deindexing

Section titled “Consolidation, pruning and deindexing”

Content consolidation merges two or more pages about the same intent into one stronger page. Then you add a 301 redirect from each old URL to the new one. Google calls a redirect a strong canonical signal, so the links to the old pages count for the new page. Consolidation is the usual fix for cannibalization.

Content pruning removes or merges low-value pages. Deindexing removes a page from the search index. Google gives two messages, and both are true:

  • In 2022, Google said that “removing unhelpful content could help the rankings of your other content.” Its helpful content system looked at the whole site.
  • Google’s core updates guide says that deleting content “is a last resort”. If you want to delete whole sections, those sections were probably made for search engines first.

So prune with care. Ahrefs reports cases where large prunes helped. Two were very large sites with crawl budget problems. In a third, the site made other changes at the same time. Ahrefs advises pruning in batches and measuring each batch.

Pick the right tool for each page:

Situation Action
Still useful, but out of date Refresh it. Keep the URL.
Two pages for one intent Merge them. 301 the weaker URL.
Not useful, but has links or traffic 301 to the closest relevant page.
Not useful, no links, no traffic Delete it. Return 404 or 410.
Useful on the site, not in search Add noindex.
Must leave search today Use the Search Console Removals tool. Its effect lasts about six months.

Do not use robots.txt to remove a page. Google cannot see a noindex on a page that it cannot crawl.

AI tools for content

Section titled “AI tools for content”

AI tools for content analyze top-ranking pages and suggest what a draft lacks. Many of them also write drafts. Google’s position has two parts.

First, Google rewards quality “however it is produced”. It said this in February 2023. Not all automation is spam: sports scores and weather forecasts are automated and useful. Generative AI “can be particularly useful when researching a topic” and to add structure.

Second, the risks are real. Using AI “to generate many pages without adding value” can break the scaled content abuse spam policy that day 02 covered. Language models “predict a likely sequence of words”, so they can make up facts. Google says that it is “critical to manually factcheck” all AI content, also titles, descriptions, structured data and alt text.

The roadmap names four tools, which are often called content optimizers:

Tool Main job, from its own site in October 2026
Surfer SEO Content Editor with a real-time “Content Score” for terms, structure and word count. Also AI drafts.
MarketMuse Content planning: finds topic gaps and clusters across your whole site and competitors’ sites.
Clearscope Term suggestions and content grades. Also tracks brand mentions in AI answers.
Frase Briefs, AI drafts and audits. Watches pages for drops and drafts a fix.

Use these tools for research and checklists. Do not treat their scores as Google’s rules. A content score compares your terms and length with the top pages. That pushes you toward the consensus, which is the opposite of information gain. And Google says that it has no preferred word count.

Key terms

Section titled “Key terms”
  • People-first content: content made mainly to help people, not to get search traffic.
  • Main content: the part of a page that directly helps the page do its job.
  • E-E-A-T: experience, expertise, authoritativeness and trustworthiness; trust is the most important part.
  • YMYL: topics that can affect a person’s health, money, safety or society.
  • Search quality raters: people who check Google’s results with the rater guidelines; they do not change rankings.
  • Information gain: what a page adds beyond the pages that a searcher already saw.
  • Primary data: data that you collect yourself, for example with a survey or your own logs.
  • Content lifecycle: the life of a page from plan and publication to refresh, merge or removal.
  • Content freshness: how current a page is; it matters for queries that expect new information.
  • Content decay: the slow loss of a page’s rankings and traffic over time.
  • Cannibalization: two pages of one site that compete for the same query.
  • Duplicate content: the same or almost the same content at more than one URL.
  • Scraping: republishing other sites’ content without added value; a spam policy violation.
  • Content consolidation: merging pages about one intent into one page, with redirects.
  • Content pruning: removing or merging low-value pages.
  • Deindexing: removing a page from a search engine’s index.
  • Content optimizer: a tool that compares a draft with top-ranking pages and suggests changes.

Exercise

Section titled “Exercise”

Exercise · about 15 min

Audit one article for quality and decay

Pick an article on a site that you know, or on your own site. A “how to” or “best of” article works well.

  1. Open the article. Find the author. Is there a byline? Does it link to an author page?
  2. Find the dates. Is there a published or “last updated” date? Do the facts look current for that date?
  3. Search Google for the main query of the article. Open the top three results. Write down the facts that all three give.
  4. Now list what the article adds that the top three do not: data, tests, photos, a method or a view.
  5. Copy one unusual sentence from the article. Search for it in quotes, for example "exact sentence here". Do other sites show the same text?
  6. Search site: with the domain and the main keyword. Does the site have two pages for the same intent?
  7. Pick one action for the article: keep, refresh, merge, redirect, noindex or delete. Write one sentence about why.
Show a possible solution

Answers depend on the article. Here is a possible result for a typical “best espresso machines” article.

  • Who: a byline with a link to an author page that shows years of coffee work. Good for E-E-A-T.
  • Dates: “Last updated” two years ago. Two of the listed machines are no longer sold. The facts are out of date.
  • Top three results: all name the same five machines and give the same price ranges.
  • Information gain: the article has the author’s own photos and a test of brew temperature. This is primary data, and it is worth keeping.
  • Quoted sentence: two small sites show the same paragraph. They probably scraped it. This does not cause a penalty for the original site.
  • site: search: the site has a second, shorter “top espresso machines” page. The two pages compete.
  • Action: refresh the article with current machines and new tests. Keep its URL. Merge the short page into it with a 301 redirect. Change the “last updated” date, because the content really changed.

Check yourself

Section titled “Check yourself”
A site adds author bios to every page and expects a ranking boost from its new 'E-E-A-T score'. What is wrong with this idea?
Two articles answer the same question. Article A summarizes the top five results well. Article B adds the results of the author's own test of six machines. Which statement is best?
An article lost traffic slowly over a year. Its facts are old, but the topic still matters to the audience and the page has good links. What should you do?
Your shop has the same product page at three URLs because of tracking parameters. What happens in Google?
A team uses an AI tool to write 500 city pages. Each page has the same text with a different city name, and nobody reviews them. Which risk is the biggest?

Sources

Section titled “Sources”

Next

Section titled “Next”

Day 05b moves from pages to things: entities, the Knowledge Graph and structured data with schema.org.