Rewind SEO’s Encyclopedia of Risky Links

Updated: July 3, 2023.  Originally published: August 31, 2014

Use the quick navigation bar on the right to skip to a specific link type  -> 

This page lists link types, uses, and patterns that make up the most the most common causes for Google Penalties in 2014. It can serve as a rough do-it-yourself penalty diagnoses guide of your penalized or at-risk site to determine broad penalty causes or impending risks. The information is equally valuable for screening SEO services or backlink packages to determine how risky they could be for your site. Though keep in mind that many SEO services will be using the more risky links on second and third link tiers that will be considerably safer than pointing directly at the site. 

The intent of this guide is that even if you are a laymen with little understanding of SEO strategies or concepts, you can still get a good idea about the causes of a penalty or impending risks. Obviously an analysis on this scale is NOT a substitute for a comprehensive penalty assessment which will be necessary for an accurate disavow. 

Triggers for Google Penalties

Penguin Analysis

The following list contains the common link types, uses, and patterns that trigger algorithm penalties and sometimes incur manual webspam penalties like ‘unnatural links’. Note that these are general rules where you should be thinking about the link profile as a whole, not individual links. A few risky links won’t matter in an otherwise strong organic looking link profile but when these link types and patterns are used at scales where it becomes a significant proportion of the links, these are the most common causes of penalties we see.

To skip to a specific section or link type, use the quick navigation box in the top right sidebar of the page.


Commercial Anchor Over-Optimization

Unsafe anchor text distribution is a common penalty trigger and the first thing you should check for, particularly for Penguin recovery. It involves over-optimized anchor text ratios for one or more exact match keywords. It is particularly common for commercial keywords (AKA ‘money’ keywords). We commonly find this problem with sites that have used cheap automated link building services that only allow a few different keywords, and also older link profiles that have experienced long-term penalties. Over optimized anchor text ratios are prime targets for Google Penguin and can easily be fixed or greatly improved with careful application of the Google Disavow tool. Additional improvements can be achieved with carefully diversified link building.
General Penguin-Safe Anchor Text Guidelines:

  Anchors with ranking keywords (AKA ‘money keywords’) should ideally be making up less than half the link profile. Higher percentages do not guarantee a penalty but do put the site at higher risk.

  No single anchor text keyword or phrase except brand name and naked URL(s) should be over 10% of the total link profile. Money keywords should be under 5%.

‣  Exact money keywords should be mixed with other related keywords as well as natural phrases that include those keywords.

‣  Keyword+natural phrase and keyword+brand name anchor variations should approach similar numbers or even exceed exact-match keyword anchors. It is still extremely rare for low-budget SEO services to even include ANY of these anchors.

‣  If the brand name is also an exact-match ranking keyword it should be treated much more carefully as if it is a ranking keyword and not a brand.

‣  15% – 30% naked URLs (url address with no anchor phrase) is safe and natural. This is very important for safe SEO and Penguin recovery projects.

‣  15% – 30% generic anchor text like ‘click here’, ‘home page’ and ‘learn more’ is safe and natural. Again, very important for safe SEO and Penguin update recovery.

‣  When in doubt, SEO analyze the 3 top ranking competitors in the site’s niche next to your own site’s link profile. Compare ratios of the different anchor text types and try to match their levels.

Because anchor text ratios (and link types) naturally vary across different sectors of the internet, the best way to determine normal and safe anchor text ratios is always a competitor analysis. However, generalized ‘all purpose’ guidelines similar to what we have just summarized are listed below from Geoff Kenyon’s excellent case study via Moz:

Anchor distribution for penguin update recovery

Generalized Safe Anchor Type Distributions – these ratios are rough guidelines and can vary considerably from the listed values.   For example, not every site will have or need a brand name.

We cannot emphasize enough how significant it is to use safe ratios for anchor text, particularly exact commercial keywords and keyword phrases! This is almost 100% a common denominator across every algorithm penalty to date and most manual action penalties as well. It is almost certainly the single most significant contributing factor to receiving a penalty. If you want to be completely safe, just don’t use money keywords at all for any follow links!


Deindexed Domains

Deindexed domains have been completely removed from Google’s search index, usually for penalty reasons, though they may intentionally be set that way with a noindex tag. Based on limited testing, links form deindexed domains are still counted by Google and can even still help a site’s rankings in Google. So in theory, not every link from a deindexed domain is bad. But in general, links from deindexed domains should be considered a significant risk. Links from deindexed domains will accumulate much more rapidly in sites being spammed with made-for-SEO links. In older high-risk link profiles, the numbers of deindexed domains can end up being very significant. Such sites may have well over 50% of their linking domains deindexed and penalized in Google. This in turn can trigger a penalty due to poor link quality.

In general, links from de-indexed sites are are prime candidates for a Google disavow list with very few exceptions.


Sitewide Links – Header, Sidebar and Footer Links

After Penguin 2.0, mass site wide links over-optimzed with money keywords appeared to be the single most common cause of new algorithm penalties for our penalized clients at Rewind. These sitewide links suddenly became THE common denominator in nearly 75% of our analyses at Rewind SEO. The rank losses ranged from minor SERP drops of 1-2 pages which could just be link devaluation, up to severe SERP losses almost off the radar. In short, large scale sitewide links with money keywords are an extreme risk!

I do recommend sitewide links be nofollow but do not discount the legitimate dofollow site wide backlinks. If your business uses sitewide backlinks, you should definitely take a look at our detailed case studies on the exact reasons these links get penalized and how you can avoid these penalties without losing your links.

For much much more info on these header, footer, and sidebar links, see our detailed case studies and review on safe and risky uses of sitewide links.


General Web Directories

General link directories, also called web directories, or sometimes self-proclaimed as “SEO friendly directories” are a long-outdated link building tactic. These directories usually have categories on a wide range of topics covering potentially any type of site. Google’s official stance is that most general directories exist purely for SEO purposes. Accordingly, these directory links are commonly cited in the example links section as a cause for manual penalties and many directory sites get deindexed or heavily penalized in Google SERPs. It is highly recommended to disavow all except the top general link directories you find in the penalty analysis including paid directories. (Open Directory Project), (yahoo’s business directory), and (Best of the Web directory) are some of the only exceptions to this rule.

In fact, directories are commonly used in negative SEO cases we see, though also appear often in penalized sites that have outdated and now considered manipulative link profiles.

It is important to distinguish general SEO directories from niche directories and many business directories/citations. Both of these directories are legitimate for SEO and completely different from ‘link directories’. Niche directories are restricted to a specific vertical while business directories are proper citations, only allowing businesses with a physical brick-and-mortar address and phone number.


Social Bookmark Links / ‘SEO Bookmarks’

In the distant past before Google Search became ubiquitous, social bookmarking sites were a way to organize online resources, both publicly and privately. Later, they became a major spammed link type with most SEO bookmarking sites being used exclusively for automated spam. Penguin 3.0 in 2012 was a major turning point for these links. Shortly before Penguin 3.0 we saw large numbers of social bookmark sites getting deindexed and since then, bookmarking sites are almost all bad with many sites getting penalized from them.

As of 2023, most bookmarking links have been out of style for a long long time and most of the former bookmarking sites have long-since expired. Therefore, this link type is mostly irrelevant but occasionally SEO bookmarking sites still pop up on older link profiles. They are very easy to identify just from the URL. In case you need to identify them, a lot of older SEO bookmarking sites used the same platform and have the one of following lines in their link page URL: 


(note: * is for convenient excel wildcard search and not part of the URL itself)


Wiki Links

Since around 2010, wiki links have been a cool “new” and innovative link type to be spammed to every imaginable wiki with the cheapest auto-spun content. While their effectiveness quickly dropped off, they are still being spammed to every imaginable wiki with unintelligible content and are still common spam links even in 2023. Though now it is mostly pure automation. Just in 2023, there has been a new wave of wikipedia scraper clone sites, all with 0 TrustFlow. If you see hundreds of wiki links to your site from domains with random letters and numbers, all with 0 TrustFlow, that is it. Just disavow them. Also disavow spammed made-for-SEO wiki links (page content will be gibberish and links use commercial anchor text). 

On the other hand, obviously not all wiki links are bad. A meaningful entry (and link) on a topically relevant wiki site is going to be good link of course. Likewise Wikipedia is a great (nofollow) link and other wiki sites can be good. But most are spam which will usually be evident from both the content and context. 


Blog Comments on Irrelevant Blogs OR with Commercial Keyword Anchor Text

This was CONFIRMED by Matt Cutts (google) at the end of 2013, though one hopes it was already common sense to any SEO because this is one of the most common spam link types in SEO history. Blog comments are not risky links in general but when the commenter’s ‘name’ is a money keyword, the link becomes a spam signal. The safest anchor text for a blog comment is obviously your name (or any name). In some cases, a company name may be appropriate though Matt Cutts has warned against this as well. High numbers of blog comments on topically irrelevant blogs and sites is another risk. 


Automatically Approved Blog Comments

Auto approve (AA) blog comments are unmoderated pages where comments are automatically approved and posted. These are extreme risk links due to comment spamming programs like scrapebox. Some of these pages can gain hundreds of new links every day and have tens of thousands of total links. They can be found in the link analysis by checking for high outbound links counts (OBL) and will usually have your link placed on a non-relevant page spammed next to hundreds, even thousands of other completely irrelevant spam links.

The idea that auto approve blog comments are safe on pages with low outbound links is also incorrect. This was an old Scrapebox tactic that never made since in the first place. These links may not be immediately risky, but over time they will be spammed more and more so at best it’s a ticking time bomb. Either way, high proportions of blog comments in general are a risk. Most recent AA blog comments I see now are used for negative SEO

If you’re going to do blog comment links, comment on a moderated page, on a topically relevant site and just use your name as the anchor text. The problem with blog comments is that no one wants to manage comment moderation anymore because of the hundreds of spam comments that will be mixed in with the real ones. So realistically, don’t expect to get any real links from blog comments.


Automatically Approved Article Directories

Non-editorial article directories are risky because they are usually filled with unreadable spun content and therefore just another type of spam site. If the particular article linking to your site is unique and readable that’s one good thing. However, these sites will often have manual or algorithm penalties, get no human traffic, and are just risky link neighborhoods in general due to massive overuse. It’s easy pick out many of these sites by the simple fact that they’ll have “article” in the name and in the URL. A few of these links  may be passable but a large number of AA article directories will help to trigger a penalty.

In general, article directories are bad links and considered manipulative by Google. Back in the day, EzineArticles(.com) was considered the “best” article directory but we have seen this site specifically listed as example in a reconsideration request rejection notice. To be safe, just stay clear of all article directories since even editorial directories can be referenced in a manual penalty. As of 2023, article directories are mostly a dead/outdated link type that is rarely seen anymore but sometimes they do pop up.


Bulk Forum Posts

Mass forum post backlinks are similar to AA blog comments though usually with smaller counts of outbound links. Originally these spam links were generated by programs like xrumer and to this day in 2023, they are still common spam links. Bulk forum post links are pure spam links, commonly used in negative SEO attacks.

Forum posts in general are low quality links that only have real value with moderated forums and topics relevant to your site. Generally do not use commercial keyword anchor text when linking from forums. On the other hand, organic link profiles still commonly see real/organic forum links from topic-relevant forums. Generally if you’re making a real forum post with a topic-relevant link, just use the plain URL or a neutral (non-keyword) anchor text. 


Bulk Forum Profiles (profile links)

Forum profile links are different then forum posts because the link is in the user’s profile rather than a public post. These ‘entry level spam links’ are one easiest types to create in bulk on the order of thousands or even hundreds of thousands. There are a lot of common footprints to find these links by URL alone. Any Penguin recovery project must get rid of these links as a high priority.

To clarify, forum profiles which have real forum posts and activity are not a problem. If you have a few forum profile links that isn’t a problem. But bulk forum profiles refers to profiles with no posts (or a single spam post) that come in tens, or hundreds of links. Get rid of all of those. 


Bulk General Profile Links, AKA ‘Web 2.0 Profiles’

These are basically inactive/empty profile links on a variety of different types of sites (not just forums). Link building services like to call them web 2.0 profiles or similar names. These are considerably lower risk than forum profiles and also have more SEO value (due to site type diversity). However, they are low quality links and in large numbers are still a major risk for Penguin. Due to common CMS platforms, web 2.0 profiles are readily visible in an analysis by various URL footprints.


‘SEO’ Press Releases (Dofollow Links)

Since mid-2013, Google has CONFIRMED that Press Release links using DoFollow links are manipulative. Considering the way press release links were spammed in the past, anyone could have seen this coming. PR links have been massively over-spammed to free non-editorial press release sites and now are considered risky links. Google specifically states in their webmaster guidelines that dofollow money keyword anchors in press releases are considered link schemes. In the decade since Google’s announcement, PR links have now fallen out of style, but they aren’t all bad…

What about Good Press Releases?

Real editorial press releases are great for press exposure and traffic but these should generally be avoided for SEO purposes. That said, from our observations through 2023, Google has rarely penalized sites that use limited numbers of DoFollow press releases and neutral anchor text. Write some newsworthy press releases and submit them on real PR sites. No worries if a few of them are DoFollow links. Also do NOT under any circumstances use commercial anchor text for press release links (this includes EMD companies/sites where their “Brand” name is actually a general keyword).


Exposed Link Networks and Blog Networks

Blog networks are powerful but risky Blackhat backlinks that can look like legitimate blog post links. Google classifies these links as ‘manipulative’ links because they are not organic blogs and are created purely for SEO. If PBNs are created “correctly” they will have almost no footprints by using different/hidden domain registrant details, hosting IP address, Class-C IP block, DNS, Google Analytics code, and more. They must also be large enough that they don’t share too many outbound links in common across the sites. That means any single link cannot be placed on every single site in the network for footprint reasons, and overlap on outbound links must be minimized.

If any of these precautions are not taken, PBN links become a liability to every site they are linking to. Cheap, abandoned, or already penalized blog networks will have many of these commonalities in plain view, leaving obvious footprints to anyone that is looking (Google). Deindexed link networks obviously should also be removed/disavowed as they have likely been penalized, leading to their deindexed status. Google’s web spam team is always looking for poorly hidden footprints to reveal a link network. When they find these networks, the links are devalued and sites using those links have a high chance of getting slapped with a manual penalty or may also be penalized in the algorithm. 

Fortunately these same footprints are also how to find these links in a penalty analysis in order to get these links removed or disavowed. Links from exposed link networks always need to go, and the best time to get rid of them is BEFORE a penalty… On the other hand, properly hidden link networks can still be effective (even in 2023 at the time of this update). Of course, all link networks are blackhat links so use at your own risk. Also, if your site receives a manual penalty, all link network links need be disavowed with attempted removal before a reconsideration request. Link networks are very common causes for manual penalties.

Should You Use Private Blog Network Links in 2023? 

As noted above, PBN links are blackhat links that Google classifies as manipulative links. This means they are a risk no matter what and ideally should be avoided. However, higher quality private blog networks with no footprints still work and are still effective for SEO (even in 2023). Unlike general SEO spam links, private blog networks can still help rank your site if you are willing to accept the risks. Because of this, we do not disavow PBNs we find in our own analyses unless they have specific risk issues or have very poor link metrics, or, of course, if the site has a manual penalty. So if you absolutely must you these links, keep in mind the precautions noted above.

Maybe the biggest risk about PBN links is that a lot of PBN sellers are selling links to exposed or deindexed sites but will not disclose that fact because they don’t care as long as they can continue to make sales. When it comes to PBNs, long term reputation is everything. It is also advisable to avoid networks that are filled only with posts using money/commercial anchor text, which is unfortunately still very common even in 2023. 


Lack of Quality Links

This is a very general but important consideration. A lack of high quality links with tons of low quality links can also lead to a penalty and is very common in penalized sites. I will not go into what a quality link is here, except that it must include good and original content, page and domain quality metrics, and/or topical relevance of the domain. Cemper Power and Trust or Moz Authority and MozTrust are both reliable metrics from Link Research Tools and Moz respectively. On the other hand, Google Page Rank is not a reliable quality metric! Page rank is almost a year out of date and should be considered a legacy metric with little significance.

In general, the presence of even small percentages of high quality links, editorial links, and links from relevant authorities can make a significant difference in how safe the link profile is as a whole and how much ‘padding’ you get for the risky links. When in doubt, we always recommend a side-by-side competitor analysis putting your site next to top ranking competitors.


Link Type Diversity

One of the most fundamental principles in SEO; link profiles with very little diversity in link types usually appear very artificial. Ratios will vary based on site type and topic so again, when in doubt, check what the top ranking competition is doing.


This concludes our list of the common causes for Google penalties. Remember again that this guide is intended for general diagnostic purposes. It is a ‘rule of thumb’ guide. If you are actually making a disavow list for a Penguin recovery project or a manual link removal list for a reconsideration request, you will need a more comprehensive assessment backed by link metrics in order to avoid disavowing potentially good links and also getting any less common or hard to find risks. In this case, consider our comprehensive SEO Analysis and Penalty Recovery Solution



All statements and suggestions are backed or derived from our experience at Rewind SEO working on hundreds of Penguin recovery projects and countless more manual penalty analyses. However, this information may not be universally applicable or true. Ultimately it is Google that determines ‘good’ and ‘safe’ uses of backlinks and they seem to change their ‘rules’ quite often so there is no guarantee that all safe or effective SEO practices today will be valid for future algorithm updates. Certainly in the past, we have seen huge changes from Google in terms of how different backlinks and backlink types are treated in their search and penalty algorithms. Always keep this in mind when building links. The safest type of link is one that looks as organic as possible within your full link profile and serves a purpose beyond SEO.

About Daniel Delos

Daniel is the founder of Rewind SEO and has worked on hundreds of Google penalty analyses and recovery projects, recovering both manual and algorithm penalties. He has almost two decades of total SEO experience and has worked almost exclusively on risk auditing and penalty prevention/recovery since 2014.

Similar Posts