Enjoyed the Read?

Don’t miss our next article!

Case Study: How Penguin 4.0 Has Affected Anchor Text Optimization

Matt Diggity
Matt Diggity is a full time SEO specializing in affiliate website ranking for both micro-niche sites and medium-level authority sites. Learn more about Matt at DiggityMarketing.com.
    This is a guest contribution to Ahrefs Blog. Author’s views are entirely their own and may not always reflect the views of Ahrefs team.
    Penguin’s latest 4.0 release shook-up the SERPs of many niches by adjusting their average anchor text ratios.  In this article, I will showcase 5 case studies that were affected by this update.  You’ll see the role that anchor text optimization played in their SERP displacements and the exact steps that were taken to recover their rankings.

    Penguin Then and Now

    On September 23, 2016 the face of SEO was changed as Google released the always-running version of the Penguin algorithm:  Version 4.0.

    While many webmasters rejoiced as we now had a 24/7 police officer available to punish our naughty competitors, others were hit with ranking changes… and not all good.

    Penguin was originally designed to help eliminate spam by attempting to identify unnatural offsite SEO signals.


    • The distribution of anchor text sent to a given page should look natural (more on this later).
    • The quality of link sources linking to a page should be high.

    The interesting thing about Penguin is that historically it was run periodically:

    • Penguin 1.0 — April 24, 2012
    • Penguin 1.1 — May 26, 2012
    • Penguin 1.2 — October 5, 2012
    • Penguin 2.0 — May 22, 2013
    • Penguin 2.1 — October 4, 2013
    • Penguin 3.0 — October 17, 2014
    • (Current version) Penguin 4.0 — September 23, 2016

    Essentially, there would be huge gaps (sometimes longer than a year), where Penguin wouldn’t run and people would be able to get away with (for lack of a better word) murder, using overly-aggressive anchor text and poor quality links.

    That is, until Penguin swung back around and it was time to pay the price.

    The biggest gap we have seen has been between our current Penguin release 4.0 and the previous 3.0 release with roughly a two year timespan between iterations.

    So what has actually happened in this two year gap?

    Niche Specific Target Anchor Text

    Offsite optimization essentially boils down to how natural the incoming anchor text looks for a given web page while still being optimized for the intended keywords.

    When trying to optimize for Google (or any other search engine for that matter), you could say we’re playing poker with our competition, but we can see their hands.

    Let me explain…

    Many people try to decide how to optimize their anchor text by using “rules of thumb” based on what they might think looks like an average breakdown of anchor types.

    Typically it might look something like this:

    • Target: 10%
    • LSI: 10%
    • URL: 25%
    • Brand: 25%
    • Topic: 20%
    • Misc: 10%

    These numbers are completely hypothetical and more often than not, do not nearly represent your competition’s anchor distribution — which has already gotten them to #1.

    So instead of guessing, simply reverse engineer the top ranking websites, and determine an average anchor text distribution for the niche.

    When the cat’s gone away, the mice come out to play

    As mentioned earlier, Penguin is the policeman that looks at anchor text and wreaks havoc on the perpetrators of its rules.  For roughly two years, Penguin was on sabbatical.

    That’s a long time in the evolution of the SERP.

    For two years, the average anchor text distributions for niches were slowly changed over time.

    More often than not, they were becoming more aggressive as people were sneaking in the SERPs with high proportions of target anchor text.

    It was very 2010 SEO, but hey, it worked.

    So what has happened now that Penguin is back and running 24/7?

    It’s been a few months since Penguin 4.0 has released and the dust has settled, so let’s take a look at how this has changed offsite anchor text optimization.

    In this case study, I will examine 5 niches and will be taking a look at how the ideal anchor text distribution for a given niche has changed in a pre-and-post-Penguin 4.0 environment.

    These niches each contain live ranking websites that we manage at my company LeadSpring.  The test suite includes both affiliate sites of varying sizes (including a 1000+ page sample) and a local search website.

    Test Methodology: Determining the Niche-specific Target Anchor Text

    I’ll be doing a comparison of each niche’s average anchor text distribution before and after Penguin 4.0.

    The methodology can be summarized as follows…

    Step 1. Download anchor text data for the website in position #1 from Ahrefs Site Explorer

    Example: Ketogenic Diet

    Ahrefs Site Exploer Anchors Report
    Step 2. Categorize each anchor by # of referring domains and type
    anchors categorized
    Step 3. Create a pie chart of the anchor text distribution breakdown
    ketogenic pie

    Step 4. Repeat for sites #2–5
    Note: for sites that are hiding backlinks, I simply skipped over them for the purpose of this study.

    Step 5. Compute the average of sites #1–5 to determine the average anchor distribution for the niche
    The niche-specific target anchor text was then compared before and after Penguin 4.0’s release date.

    keto average

    Case Study 1: High Competition Affiliate Niche in the Home Improvement Space

    case 1

    SERP Reaction after Penguin Rollout

    This niche was shaken up pretty hard when Penguin eventually rolled out to this niche on October 29th, 2016.

    Our site was ranking steady at #1 for the previous 2 months and immediately took a plummet to #6 where it stayed for a few weeks while we diagnosed the issue.

    The new site that replaced ours was actually previously sitting at #5.  Sites that were previously at slots #3 and #4 vanished from page 1 entirely.

    By comparing the niche-specific anchor text distribution averages before and after the rollout, some definite patterns started to emerge.

    • A huge decrease in the average number of target anchors: ‑15%
    • Slight increase in the use of LSI synonyms: +4%
    • Brand + URL keywords increased significantly: +8%
    • Slight increase in topic anchors: +3%


    We can see that this SERP result was in desperate need of a Penguin visit.

    Websites, including ours, were getting away with aggressive 52% target anchor text for over a year.  I honestly wasn’t surprised when we tanked.

    Sites previously occupying #3 and #4, which I mentioned earlier were knocked off of page 1 entirely, had target anchor text ratios of 62% and 68%, respectively.  These two bad boys were largely responsible for throwing off the previous distribution’s average.

    Once Penguin stopped by, this aggression was diluted in favor of LSI, brand and URL anchors.

    We followed suit by removing our aggressive anchors from the links we had access to and replacing them with new LSI, brand and URL links. 

    Case Study 2: Foreign Affiliate (Brazil) in the Health and Wellness Space

    case 2

    SERP Reaction after Penguin Rollout

    Penguin caught up to this niche late in the roll out (November 13, 2016).

    We were, quite frankly, extremely surprised to see any fluctuations in this niche as the niche-specific anchor text distribution was well-rounded before the roll out.

    Personally, I like to push the envelope when it comes to optimization, but for this niche I was on my best behavior.

    Once Penguin hit, our site moved from #2 to #4 and the sites that moved up had one huge commonality: a high amount of local citations.


    We’ve diagnosed this issue to being a cause-and-effect relationship between two algorithms.

    The Google.com.br engine had increased the need for locally relevant domains.  Namely, domains with the .com.br TLD extension as well as being in the Portuguese language.

    Two newcomers to the top 5 results both had extremely heavy usage of business citations in their homepage profiles.  These citations were providing both local relevance and language relevance simultaneously.

    One company had 13 business citations and the other had 19 – both very high numbers for Brazil.

    As a result, the average breakdown for the niche saw a 15% increase in URL anchors (the anchor type that is most commonly used in a business directory link).

    Our play: Copy what worked.

    We had our team go out and build 20 business citations and indexed those bad boys immediately.

    Case Study 3: Medium Competition Affiliate Niche in the Beauty Space

    case 3

    SERP Reaction after Penguin Rollout

    Once Penguin rolled out to this niche, we actually experienced an increase in rankings.  Win.

    The homepage of our partial match domain (PMD) in the women’s beauty niche was ranked #6 for the main keyword.  Post-update, we shot up to #2.

    This was particularly surprising as we felt we were highly over-optimized for this niche with 6% target anchor text, while the niche was asking for 3%


    As mentioned earlier, Penguin is a two-part algorithm.

    The first aspect checks a backlink profile’s anchor distribution for anomalies.

    The second checks for low quality links.

    In this niche, previously we noticed that the top 2 – 4 slots had exhibited a large amount of low quality blog comments as exhibited by the Ahrefs reports.  Many of these referring URLs had in excess of 50 out bound links, sharing comments with “ugg boots”, “louis vutton bags” and other usual suspects in the spam space.

    Once Penguin moved out slots 2–4 and kicked them back to page 3 and beyond, we were left with a new top 5 that had clean links.

    This new top 5 had a significant decrease in the NA anchor category (-22%) which can be attributed to the loss of “author name” links which are commonly filled out in blog comments.

    In addition, the new top 5 happened to have more aggressive anchors than before, which suited our homepage quite well.

    Case Study 4: Medium Competition Affiliate Niche in the Beauty Space (Inner-page of Case Study 3)

    case 4

    SERP Reaction after Penguin Rollout

    The interesting thing about Penguin 4.0 is that it’s applied granularly, on a page-per-page basis.

    This search associated with this inner page of Case Study #3 has a completely different anchor text distribution from the homepage, before and after the roll out.

    In fact, Penguin caught up to this page much later than the homepage (October 27, 2016 vs. October 2, 2016, respectively).

    This particular search is your typical Amazon affiliate “<product> review” keyword.  We were sitting at #3 and dropped down to #7 after the algorithm paid its visit.

    We found 3 new players that jumped into this niche’s top 5:

    If you’re doing affiliate SEO, then you’ll know that this is pretty bad news.  Authority sites of this magnitude are difficult to topple.


    As can be seen from the pie charts, the anchor categories that were affected the most were:

    • Target anchors: ‑19%
    • Brand: +23%

    This is very typical when you have these types of authority sites in your niche. People link to Amazon with brand anchors like “Amazon”, “at Amazon” or “on Amazon”.

    Because we are using a partial match domain (PMD), we don’t necessarily get a brand anchor.  The best we can do is to try to imitate this 23% increase by substituting in naked URL anchors.

    Case Study 5: Local Search in the Health and Wellness Space

    case 5

    SERP Reaction after Penguin Rollout

    This case study, followed a very similar pattern to case study #1, despite this niche being locally-based and the latter being national search affiliate.

    Before Penguin 4.0, we had been #2 and we dropped down to #10 on November 5, 2016.

    Key changes:

    • A large decrease in the average number of target anchors: ‑36%
    • Slight increase in the use of LSI synonyms: +8%
    • Brand + URL keywords increased significantly: +14%


    While our site completely tanked after this update, it actually helped answer a long outstanding question we’ve had about this niche’s optimization parameters.

    The site we’re trying to rank is a yoga studio.  We’ll call the brand “Prana Yoga” for purpose of illustration.  We were trying to rank for the keyword “<city> yoga”.

    Most sites in our niche, ourselves included, were using a major of what we believed to be brand anchors:

    • Prana Yoga”
    • Prana Yoga Studio”
    • visit Prana Yoga”

    The sites that weren’t using “yoga” in their branded anchors, received huge rewards in this update and the folks like us were kicked off the top 5.

    We now see that anchors like ones we were using should have been considered target anchors since they contained the word “yoga” in them and are essentially longtails.

    Our game plan: removal of anchors and dilution with simply “Prana” anchors.


    After adjusting the anchor text ratios for each of our sites, we found that we had an 80% (4 of 5) success ratio at recovering each niche.

    We spaced out the anchor removals and replacement over the course of 2–6 weeks (depending on how many anchors needed to be adjusted) and found that the average response time was 16.4 days from the time the last anchor was corrected until SERP recovery.

    We’re still working on that final testcase (#4) that hasn’t recovered.  Our hypothesis is that we might have been “checkmate’d” since we cannot follow suit with the niche’s demand for more brand anchors since we’re a PMD.

    You win some, you lose some.

    We are considering redirecting a branded domain in this same niche to this URL, but it’s such a long shot that it’s probably not worth the effort.  Either way, I will report back with the results.

    What to do if you were affected by Penguin 4.0

    First things first, figure out which aspect of Penguin might have caused your downfall.

    Low quality links

    Do a complete audit of the incoming links to your site.

    Are these links are coming from trusted resources?

    Are they coming from relevant sites?

    If not, either get them removed or break out the disavow tool.

    Unnatural Anchor Text

    If you’re getting your links from trusted sources, it’s likely that anchor text is the culprit.

    At this point you need to assess what the niche-specific target anchor text distribution is currently expecting for your keywords.

    Download your competitors anchor text with a backlink crawler tool like Ahrefs.  Sort these anchors into categories and come up with the niche-specific average.

    Once you have a clear eye on the target, it’s time to start modifying your distribution to follow suit.

    This might involve diluting your distribution by earning more URL anchors.  If this is the case, a little business citation work can save the day.

    If you need to reduce the density of your target anchor texts, I recommend getting them removed by contacting the owner of the referring domains.  Don’t bother with getting the anchor texts changed within existing articles.

    As you can see, this Penguin update is highly anchor-focused.  Based on the average of these 5 case-studies, target anchor text usage dropped by 15% while “pillow” (Brand/URL/topic/misc) anchor usage increased by the same amount.

    The great thing about the live-version of the algorithm is that recovery can be swift if a proper anchor text diagnosis is made and applied to your own site.

    Editor’s Note
    Lasse Nielsen from Skjoldby & Co shared a spreadsheet with us, which can easily create very similar pie-charts for anchor text distribution. It will come in really handy for visualization and for reporting.

    You just need an Ahrefs account and Microsoft Excel installed.

    You can download the file here.

    Nick Churick

    Shows how many different websites are linking to this piece of content. As a general rule, the more websites link to you, the higher you rank in Google.

    Shows estimated monthly search traffic to this article according to Ahrefs data. The actual search traffic (as reported in Google Analytics) is usually 3-5 times bigger.