An SEO link network, in Polish “zaplecze”, is a set of sites that someone builds or controls so that they link to their main site. It can be a blog network on its own domains (PBN, private blog network), domains bought back after abandoned projects, presell pages written for a single link, or a slot in a link exchange system. Google classifies links created primarily to manipulate rankings as link spam, and since December 2022 it detects on both sides of such a network both the buyer and the pages that pass the links.
What is an SEO link network
The Polish word “zaplecze” means back room or backstage. In SEO it names auxiliary sites whose main job is to pass links to the target page, that is, the page that is supposed to sell or collect leads. Whoever runs the network decides the content, the anchor text and the position of the link, and so has a level of control that a link from an independent publisher does not give. That control is exactly what Google’s policies describe as the problem:
“Link spam is the practice of creating links to or from a site primarily for the purpose of manipulating search rankings.”
Google Search Central, Spam policies for Google web search
The definition rests on the purpose of the link. A link in an article written for readers and a link in a text written only to carry the link look the same in HTML. What separates them is the reason they were created, and that reason is what Google evaluates.
How a hyperlink is constructed in HTML
Hyperlinks, links and anchors are, in web terminology, a reference (address) to another element: an HTML page, an image, a sound, a film, and so on. An anchor is the term used for the destination of a hyperlink within a document. The <a> element in HTML is used to define both the hyperlink and the anchor.[1] Activating a link opens the target document.
The syntax of a link[2] looks like this:
<a href="value">link text</a>
The values that can go into the “value” part include:
- a direct URL, pointing to another page
- an indirect URL, pointing to a file within the current site
- a URL identifier[3], pointing to an element on the current page (a document fragment, for example
href="#section")
Today the <a> tag has its own attributes: href, target, rel, download, hreflang, ping, referrerpolicy and type. On top of these come the global attributes shared by all elements (among others accesskey, class, dir, id, lang, style, tabindex, title) and the event attributes for scripts (onclick, onfocus, onblur, onmousedown, onmouseover, onkeyup and others).[4] The attributes charset, coords, name, rev and shape, which older HTML courses listed, are obsolete. MDN recommends the global id attribute instead of name, and the <area> element in image maps instead of coords and shape.[4]
To work correctly, the syntax must end with the closing element </a>.
In a link network the most important attribute is rel, because that is where the linking page tells the search engine what kind of link it is. Google writes in its documentation on links that its crawler can reliably follow an <a> element with an href attribute, and that it does not reliably read links triggered only by JavaScript events (Google Search Central, crawlable links).
The rel attributes: sponsored, ugc, and nofollow
For years rel="nofollow" was the only way to mark a link that should not pass value. On September 10, 2019, Google announced two new values: sponsored for links created as part of advertising and ugc for links in user-generated content. In the same post it changed the status of all three:
“[The link attributes] are treated as hints about which links to consider or exclude within Search.”
Google Search Central, Evolving “nofollow”: new ways to identify the nature of links
Since March 1, 2020, nofollow is also a hint for crawling and indexing. The current documentation on qualifying outbound links assigns these uses to the values:
| Value | When Google recommends it |
|---|---|
rel="sponsored" | advertising links and paid placements, that is, paid links |
rel="ugc" | links in user-generated content, for example comments and forum posts |
rel="nofollow" | when the other values do not apply and the page does not want to be associated with the target or send the crawler there |
Values can be combined, separated by a space or a comma, for example rel="ugc nofollow". Nofollow remains an accepted way to mark paid links, but Google prefers sponsored. The same documentation points out that the target page can still be found by other means, such as a sitemap or links from other sites.
For the topic of link networks, the key sentence is in the spam policies. Buying and selling links for advertising purposes is described as a normal part of the web economy, and such links do not violate the policies when they carry rel="nofollow" or rel="sponsored" (spam policies, link spam). A paid link is therefore not a problem in itself. The problem is a paid link posing as an editorial one.
On July 26, 2021, Google extended this to affiliate programmes. In the post A reminder on qualifying links and our link spam update, it asked sites with affiliate links to qualify them as sponsored, whether they are created manually or automatically, and warned that missing qualification can lead to manual actions. The same post covers sponsored articles and guest posts: sites that publish many of them without the appropriate attributes may face algorithmic or manual actions.
Historical link classification in search engine optimization
Tomasz Frontczak, in the book Marketing internetowy w wyszukiwarkach (2006), expanding[5] the list presented by Todd Malicoat, divides links by the factor that decides their presence on a page and by the type of page the link sits on. Below is the list as we recorded it in 2009. The links are:
- Natural (links that are not bought and whose appearance is driven by a strong or unique offer of the site they point to), one-way links from a highly ranked site, placed in the content. The link contains a keyword in the anchor. The site’s content contains keywords we care about.
- Natural, one-way links from a highly ranked site, placed in the content. The site’s content contains keywords we care about.
- As in point 1, but the site’s content does not contain the keywords.
- As in point 2, but the site’s content does not contain the keywords.
- Advertising (bought) links on a highly ranked site, surrounded by text with keywords.
- Deep links (links to subpages of the site) from so-called presell pages.
- Advertising links with a keyword in the link, on a highly ranked site.
- One-way links with a keyword in the link, from a thematically related site, not necessarily highly ranked.
- One-way links with only the domain name and no keywords, from a site with good rankings.
- One-way links with a keyword, from an ordinary site with no thematic relation.
- One-way links with the domain name, from an ordinary site with no thematic relation.
- Links with a keyword placed in a paragraph of text on a links page.
- Links placed on a links page in the form: Title/Description.
- Links in discussion forum signatures.
- Return links (backlinks) on a links page.
- Rotating links from link exchange systems.
- Links in blog comment signatures.
The list shows how the value of a link was thought about between 2006 and 2009: what counted was the ranking of the source page, the keyword in the anchor text and thematic closeness. Most of these points can be achieved by controlling the source page yourself, and building link networks grew out of that reasoning. Link types split into smaller variants are described in the article types of SEO backlinks.
Which link types violate Google spam policies today
The spam policies give examples of link spam. Comparing them with the list above shows that a large share of the classic link types is now named directly:
| Link type from the list | What Google’s policies say today |
|---|---|
| Natural links in the content (points 1 to 4) | Allowed, as long as nobody paid for them with money, goods or services. The starter guide describes links from other sites as something that appears naturally over time. |
| Advertising, bought links (points 5 and 7) | Buying or selling links for ranking is link spam. The same link with rel="sponsored" or rel="nofollow" is allowed. |
| Deep links from presell pages (point 6) | The policies mention creating low-value content primarily to manipulate linking and ranking signals. A presell page written only for the link fits that description. |
| One-way links from related and unrelated sites (points 8 to 11) | The assessment depends on how the link was obtained. Money, goods or services in exchange for links, and a product in exchange for a text with a link, are on the list of examples. |
| Links pages, descriptions and return links (points 12, 13 and 15) | The list includes links from low-quality directories and bookmark sites, excessive link exchanges (“link to me and I’ll link to you”) and partner pages created only for cross-linking. |
| Forum signatures (point 14) | Named directly: forum comments with optimised links in the post or the signature. |
| Rotating links from exchange systems (point 16) | Fall under excessive link exchanges and the use of automated programs or services to create links. |
| Blog comment signatures (point 17) | Comments with optimised links are on the list, and Google recommends that the CMS add ugc or nofollow to user links automatically. |
The list of examples also includes two practices that were not in the 2009 summary and that often come with link networks: widely distributed links in the footers or templates of various sites, and hidden or keyword-stuffed links in widgets distributed across various sites. Link spam also covers requiring a link in the terms of service or a contract when the content owner cannot choose the link attribute.
How Google detects private blog networks and purchased links
The spam policies say Google detects violations both with automated systems and, when needed, with human review, and that sites violating the policies may rank lower or not appear in results at all. In the guide to Google Search ranking systems, Google lists separately the link analysis systems, including PageRank, which according to that document remains part of its core ranking systems, and the spam detection systems, including SpamBrain.
On December 14, 2022, Google announced an update in which it used SpamBrain against links:
“Today, we’re leveraging the power of SpamBrain to neutralize the impact of unnatural links on search results.”
Google Search Central, December 2022 link spam update releasing for Google Search
In the same post, Google wrote that SpamBrain can detect both sites buying links and sites used for the purpose of passing outgoing links. For a link network this means both sides are assessed: the target page and the pages linking to it. The post also warned that rankings may change when spammy links are neutralised and the value they passed is lost. The update covered all languages, and the rollout was expected to take about two weeks.
The previous update, from July 2021, was meant according to Google to identify and nullify link spam more effectively across multiple languages. For someone running a network, neutralisation means the cost of domains, hosting and texts has already been paid while the links stop passing value. An overview of penalties is in the article common link building mistakes and penalties.
Do expired domain networks violate Google policies
A common way to set up your own network was to buy expired domains that still had links from previous years. On March 5, 2024, Google described this practice as a policy of its own:
“Expired domain abuse is where an expired domain name is purchased and repurposed primarily to manipulate Search rankings by hosting content that provides little to no value to users.”
Google Search Central, What web creators should know about our March 2024 core update and new spam policies
The policies give three examples: affiliate content on a site previously used by a government agency, commercial medical products sold on the domain of a former non-profit medical charity, and casino content on the domain of a former elementary school. In all three cases the new owner uses a reputation that someone else built, for a subject that has nothing to do with it. The definition, as with link spam, rests on the main purpose of the purchase.
Understanding site reputation abuse
Some link networks moved from their own domains to other people’s domains: instead of building a blog network, posts are bought on strong portals. The site reputation policy applies to third-party content published on a host site mainly because of the ranking signals the host earned with its own content. Third-party content is, according to Google, content created by a separate entity from the host, for example users, freelancers, white-label services, or people not directly employed by the publisher. Google has enforced the policy since May 5, 2024. Since August 30, 2026, a manual action under this policy affects only search results shown to users outside the European Economic Area; within the EEA the affected section may be treated separately from the rest of the domain and rank on its own (Google Search Central, August 2026).
As examples of violations, the policies mention among others an educational site hosting a page of sponsored reviews of payday loans and a medical site hosting a low-quality third-party advertising page. Considered compatible are wire services and press release services, news publications syndicating content from other news publications, columns and editorial texts, and sites designed for user-generated content, such as forums. The manual actions documentation adds that moving such content to a subdomain, a subdirectory or another established site may not solve the problem (Manual Actions report).
What to do after receiving a manual action for unnatural links
A manual action is a decision by a person at Google that pages of a site do not comply with the spam policies. It appears in the Manual Actions report in Search Console, and the notification also reaches the message centre. For links, Google distinguishes two cases (Manual Actions report):
- Unnatural links to your site. Google detected a pattern of artificial or manipulative links pointing to the site. Recommendation: export the links from Search Console, ask the owners of the pages to remove the links or add nofollow, and only list in the disavow tool the links that could not be removed.
- Unnatural links from your site. Google detected paid or manipulative outgoing links. Recommendation: remove them or prevent them from passing PageRank, for example with nofollow. This is the case of a network page that sells or passes links.
A reconsideration request only makes sense once all pages are fixed, because fixing only some of them does not give a partial return to the results. The pages must be accessible to Google, without a login, paywall, robots.txt block or noindex. The request describes the problem, the steps taken and the outcome. The review takes from several days to a few weeks, and you should not send another request while it is in progress.
When to submit a disavow file
The disavow links tool is, according to Google, an advanced feature. The documentation says most sites do not need it and that it should only be used when a considerable number of spammy, artificial or low-quality links point to the site and have caused, or are likely to cause, a manual action (Google, disavow links). The first step is always to try to remove the links at the source.
The file is plain text in UTF-8 or 7-bit ASCII, with one URL or domain per line. A whole domain or subdomain is given with the domain: prefix, for example domain:example.com. A new list replaces the previous one for that property, and Google may take a few weeks to incorporate the list into its index.
How to earn organic backlinks without building satellite networks
Google’s SEO starter guide points out that Google finds new pages mainly through links from pages it has already crawled and that links from other sites appear naturally over time. As one of the most effective and lasting forms of promotion, it names word of mouth, that is, people who know the site telling their friends about it, and community engagement as the basis for that. The same guide warns that it is possible to overdo promotion and harm the site, because search engines may see some practices as manipulation.
In practice, three routes remain that do not need a link network:
- A sponsored post is allowed when the link carries
rel="sponsored". It brings visits and visibility with the publisher’s audience, but not ranking value. - External links from your own page, for example to sources, strengthen credibility. Google recommends nofollow only for sources the page does not trust, not for all outgoing links.
- Links from publishers are earned with content worth citing: data, a tool, a case study, an expert comment.
How this work looked in one furniture shop is described in the link building case study for the furniture industry.
See our SEO and GEO services.
[1] https://www.w3schools.com/HTML/html_links.asp; 03.11.2009 [2] https://www.w3schools.com/tags/att_a_href.asp; 01.11.2009 [3] MDN Web Docs, the a element, https://developer.mozilla.org/en-US/docs/Web/HTML/Reference/Elements/a; 12.10.2026 [4] MDN Web Docs, the a element, attributes and deprecated attributes, https://developer.mozilla.org/en-US/docs/Web/HTML/Reference/Elements/a; 12.10.2026 [5] Tomasz Frontczak, Marketing internetowy w wyszukiwarkach, 03/2006; 08.11.2009
