How can buyers detect scraped content sites before they buy?
Part of MFA avoidance
The short answer
Scraped content sites republish articles taken from other publishers, often by automated means, either word for word or lightly reworded, and sell ads beside them. Buyers detect them by searching distinctive sentences in quotation marks, comparing publication dates and archived copies, and checking for a license or credit. Ads on these pages pay the copier, not the original publisher.
Key takeaways
- Google's spam policies define scraping as taking content from other sites, often through automated means, and hosting it to manipulate search rankings.
- An exact match search of a few sentences, plus the earliest dated copy, can show who published first.
- Licensed syndication is not scraping: Google's rater guidelines say licensed or syndicated content should not be considered copied.
- Scraped sites belong on the exclusion list, and the original publishers they copy can be strong candidates for the inclusion list.
What are scraped content sites?
Definition
- Scraped content site
- A website whose articles are taken from other publishers without permission, copied exactly or with small changes, and published to attract search or paid traffic that can be shown ads.
Google's spam policies (opens in a new tab), last updated August 28, 2026, define scraping as taking content from other sites, often through automated means, and hosting it with the purpose of manipulating search rankings. Their examples include republishing content from other sites without adding any original content or value, or even citing the original source, and copying content, modifying it only slightly, for example by substituting synonyms, and republishing it.
Google's Search Quality Rater Guidelines (opens in a new tab), September 11, 2025 edition, name the reworded version directly. When a few words or whole sentences are changed, or a find and replace swaps one word for another throughout, the guidelines call it copied with minimal alteration and note that such changes are deliberately done to make it difficult to find the original source. A scraped site may also copy from a changing source such as a news feed, which the guidelines still treat as copied content.
| Form | What the page looks like | What gives it away |
|---|---|---|
| Verbatim copy | The original article reposted whole, sometimes with the byline removed | An exact match search returns the original, published earlier |
| Light rewrite | Synonyms swapped, sentences reordered, headline changed | Odd word choices in fixed phrases, and the same quotes and figures in the same order |
| Feed copy | Headlines and openings pulled automatically from another site's feed | Posts appear minutes after the source's and stop at the same point |
| Stitched page | Paragraphs from several sources joined into one article | Changes of tone, tense and spelling between paragraphs, each matching a different source |
How can a buyer check whether content is copied?
The rater guidelines describe a method anyone can use without special software. It works best on articles with specific details, such as local reporting, reviews or interviews, because generic text can match other pages by chance.
A copied content check for one candidate site
- Copy a distinctive sentence. Choose one with names or details and no unusual characters, as the guidelines suggest, and try two or three sentences from the page.
- Search it in quotation marks, then without. The quoted search finds exact copies; the open search finds reworded ones that keep the same facts.
- Compare the matching pages. If most of the main content is the same, look for the source known for original work, such as a newspaper or magazine with named reporters.
- Find the earliest dated copy. Compare publication dates and use the Wayback Machine (opens in a new tab) calendar for each URL to see which version existed first.
- Look for a license or credit. A syndication line, a canonical link to the source or a named wire service points to a deal; no credit at all points to scraping.
- Repeat on several articles. One copied page can be a mistake. Most of the sampled pages matching other sites is a pattern.
Why do scraped sites earn ad revenue?
Scraping skips the reporting. A reporter's week of work can be copied in seconds, and a script can fill thousands of pages a day across many domains. Each page can draw search visits for the original's subject or receive bought traffic, and each carries ad slots. The trade groups' shared MFA definition (opens in a new tab) of September 26, 2023 lists generic, often syndicated, dated or non-unique content among the five characteristics of made for advertising sites.
Standard buying checks can miss it. The copied article matches the campaign's keywords because it was written for that subject. It passes brand safety filters if the original did. A scraped site can publish a valid ads.txt file, because ads.txt shows who is authorized to sell a domain's inventory, not where the domain's articles came from. Google's replicated content policy (opens in a new tab) does not allow Google-served ads on screens with embedded or copied content from others without additional commentary, curation, or otherwise adding value, but that rule governs Google's own ad products, not every exchange.
The harm reaches the original publisher twice. Ads that run beside the copy pay the copier instead of the newsroom that did the work. And in search, Google's canonicalization troubleshooting guide (opens in a new tab), last updated August 21, 2026, notes that in rare situations its algorithm may select a URL from an external site that is hosting your content without your permission, so the copy can be shown in place of the original.
How do rewritten articles differ from syndication?
Legitimate publishers also run articles they did not write. Wire services sell stories to member papers, and publishers license each other's work. The rater guidelines are explicit: licensed or syndicated content should not be considered copied, and they give news articles by AP or Reuters as examples of syndicated content in the U.S. The difference lies in permission and credit, not in whether the words match.
| Practice | Permission | Credit and links | What the publisher adds |
|---|---|---|---|
| Licensed syndication | A license or wire membership | Source named; sometimes a canonical link or noindex on the copy | Own reporting alongside; the syndicated piece is part of a wider site |
| Aggregation with commentary | Short excerpts, with or without a license | Headline and link back to the source | Selection, summary or analysis a reader can use |
| Light rewrite | None | Source credit missing or buried | Nothing beyond swapped words |
| Verbatim scrape | None | Byline and links can be stripped | Nothing; the page exists to carry ads |
Google's canonicalization guide adds a practical note for publishers that syndicate: the canonical link element is not recommended for avoiding duplication by syndication partners, and the most effective solution is for partners to block indexing of the content. A copy that carries a noindex tag and credits the source is therefore a sign of an agreement, while a copy that tries to rank in place of the original is a sign of scraping.
What should happen to a scraped site on a plan?
A site where most sampled articles are copied without license belongs on the exclusion list. There is little room for a review middle ground: the content has no editorial owner, so no section can be trusted to stay clean. Record the evidence so the decision holds up if the seller asks.
Evidence to record when excluding a scraped site
- The sampled URLs and the sentences searched.
- The original source for each match, with both publication dates.
- Archive captures that show which version existed first, where available.
- Whether any credit, license line or canonical link was present.
- Sister domains that carry the same copied articles, reviewed together.
- The date of the check and the name of the reviewer.
Then look at the other side of the match. The publishers being copied wrote the articles that made the scraped site relevant to the brief, so they can be strong inclusion candidates in their own right. Rights holders have their own remedies: in the United States, section 512 of the Copyright Act (opens in a new tab) sets up a notice and takedown system that allows rightsholders to send a notification to the online service provider. A buyer's role is narrower: keep the money off the copy and, where the original fits, put it on the source.
What does a scraped content review look like on a real brief?
Fictional example
A bookstore plans its holiday season display buy
A fictional independent bookstore with locations in Burlington and Plattsburgh asks its agency for an English display plan in Vermont and northern New York, beside book reviews, author interviews and local arts news, before the holiday gift season.
Among the candidates, four sites carry book reviews that rank well for the season's new titles. The planner searches three sentences from each. Two sites return exact matches on a regional literary magazine, published there days earlier; the copies drop the reviewer's name and keep the magazine's cover photos, cropped. A third site rewrites the same reviews with swapped words but the same quotes and plot points in the same order. All three go to the exclusion list with the evidence recorded.
The regional literary magazine is added as a candidate and passes the rest of the review. The fourth site turns out to run a licensed column from a book review syndicate, credited and linked, alongside its own reporting, and stays on the plan.
How does PlacementGraph help a team review copied content sites?
Copy checks are slow by hand across a long candidate list. PlacementGraph reads several articles per site and shows each result's best-matching pages.
Frequently asked questions
Is quoting another publisher's article scraping?
No. Quoting a passage with credit and a link, then adding reporting or analysis, is ordinary journalism. Google's spam policies target republishing without adding any original content or value. The test is whether the page would still be worth reading if the quoted passages were removed.
Can an AI rewrite of another article count as scraped content?
Yes. Google's spam policies list copying content and modifying it only slightly, for example by substituting synonyms or using automated techniques, as scraping. The rater guidelines add that content can be paraphrased by people or generative AI tools. A rewrite that keeps the original's facts, quotes and order without adding anything is still a copy.
Do scraped sites copy images and video too?
They can. The rater guidelines note that any type of content may be copied, that images may be slightly cropped or edited to avoid detection, and that videos can be altered. A reverse image search on a distinctive photo is a quick second check alongside the sentence search.
Should a buyer tell the original publisher about a copy?
It can help. An original publisher may not know where its work is copied, and the rights holder decides whether to act. Sharing the copy's URL and dates is a courtesy that costs little, and it can open a direct conversation with a publisher that fits the plan.
Sources
- Spam policies for Google web search (opens in a new tab)Google Search Central, last updated August 28, 2026
- Search Quality Rater Guidelines (General Guidelines, September 11, 2025) (opens in a new tab)Google
- Google-served ads on screens with replicated content (opens in a new tab)Google Publisher Policies Help
- Fix canonicalization issues (opens in a new tab)Google Search Central, last updated August 21, 2026
- Using the Wayback Machine (opens in a new tab)Internet Archive Help Center
- Section 512 of Title 17 (opens in a new tab)U.S. Copyright Office
- Leading Trade Groups Define Made For Advertising Websites (opens in a new tab)ANA, September 26, 2023
Researched and written by PlacementGraph Research. Read how we research, source and update these pages. Corrections: diana@placementgraph.com.