Everything We Know About The Penguin Update

In case you missed it, the past couple of months have been wild and crazy in the world of SEO. We’ve seen an avalanche of unnatural links notices; Pandas and Penguins have completely monopolized the news; and negative SEO has become one of the most hotly debated topics of the year.

This post represents my attempt to organize the chaos by describing various events that have transpired. Specifically, I created a massive timeline that includes important events (e.g., algorithm updates) as well as noteworthy posts and videos that cover those events. In some cases, I simply summarize a given post or video, and in other cases, I add my own commentary.

But before we take a stroll down memory lane, it’s important to note that I have included events that are not directly related to the Penguin Update. Although this update was the most newsworthy event of the past few months, a number of other events influenced its coverage, and as a result, they also appear on the timeline.

Now, let’s jump in the DeLorean and get this party started…

March 10, 2012

Matt Cutts pre-announces a Google “over optimization” algorithm change in the Dear Google & Bing: Help Me Rank Better! session at SXSW.

Here is the entire audio of that session:

March 13, 2012

Ryan Clark writes an informative post about private blog networks getting deindexed by Google throughout February and March: Private Blog Networks Getting Deindexed?

In this post, Ryan details how private blog networks operate, and he itemizes the various risks associated with using the networks for link building. He also makes this somewhat prophetic statement: “I think Google is just going after people who are too greedy in their anchor text distribution.” (That’s not the last time we’ll see a statement like this.)

March 15, 2012

Dan Thies has the following Twitter exchange with Matt Cutts:

(Keep this exchange in the back of your mind because it comes up again very shortly.)

March 19, 2012

Google deindexes one of the most popular paid blog networks (BuildMyRank). As Ryan noted previously, Google had already deindexed numerous networks, but the BuildMyRank deindexing received the most attention due to the network’s size and popularity.

March 22, 2012

Jennifer Ledbetter gives her take on the paid blog network deindexings: OMG! The Sky Is Falling!?!?!

In the post, Jennifer explains what private blog networks are and why so many people use them. But the real value of her post is in this excerpt: “For one, when sites that have links to you get deindexed by Google, those links don’t ‘count’ anymore. Based on that, as I said earlier, your rankings may very well drop or tank. This doesn’t necessarily mean YOUR site is getting hit with a penalty because you have links from these blog networks – it DOES mean you lost back link POWER so your rankings drop.”

These comments are common knowledge for many of us, but that doesn’t make them any less valuable. Fundamentally, Google treats links as votes or endorsements by one site in favor of another. If your site loses these votes (regardless of the reason), it’s less likely to win its “election” in the search engine rankings.

March 26, 2012

Carson Ward provides his experience as a former “link network spammer” in this insightful post: Unnatural Link Warnings and Blog Networks.

In his post, Carson presents identifying characteristics of blog network posts (e.g., “Lots and lots of exact-match anchor text seemingly pointing to sites at random”), discusses the importance of removing low quality links, and gives helpful advice for building a legitimate link portfolio. Here’s an example of a blog network post:

Carson also presents an example of Google’s unnatural links notice (many webmasters began receiving them around the same time that Google was deindexing private blog networks).

April 15, 2012

In the aftermath of Google deindexing private blog networks and sending thousands of unnatural links notices, Modesto Siotos presents an approach for assessing and monitoring the riskiness of your backlinks: How to Check Which Links Can Harm Your Site’s Rankings.

Modesto’s technique is very straightforward. First, you identify the root domains of the sites linking to your site. Then, you repeatedly check the ToolBar PageRank (TBPR) and social metrics for those domains as well as the percentage of domains that have been deindexed.

As you repeatedly check this information, you’re looking for potentially troubling signs (e.g., the TBPR distribution skewing to lower numbers, an increasing percentage of deindexed domains, etc.).

April 17, 2012

Barry Schwartz comments on a “huge uptick” in the number of people claiming that their sites were deindexed by Google: Google Update April 2012? Over SEO Penalty?

Initially, the prevailing wisdom was that these deindexings were caused by the “over optimization penalty” (this hypothesis was subsequently rejected based on comments by Matt Cutts – see below).

April 18, 2012

Jammy posts the details of his negative SEO campaigns against Dan Thies and negativeseo.me in this Traffic Planet thread: CASE STUDY: Negative SEO – Results.

Jammy chose Dan Thies as a target for a variety of reasons (one of which relates to his Twitter exchange with Matt Cutts on March 15), and he chose negativeseo.me because that site was selling negative SEO services.

This thread is incredibly interesting for at least two reasons. First, although the industry was already having conversations about negative SEO, this thread put those conversations in the spotlight and spawned a wave of new posts and opinions about the subject.

Second, the results are somewhat alarming. Although it doesn’t appear that Jammy’s scrapebox blast negatively impacted Dan’s site (for the dan thies query), it did potentially impact negativeseo.me’s rankings (I say potentially only because we can’t account for every variable that might be responsible for the observed change).

Jammy also reports that Dan’s site lost significant rankings for 3 queries (seo, seo service, and seo book) and received an unnatural links notice from Google; however, as various commenters pointed out, this result was probably caused by thousands of Authority Link Network links targeting those phrases that were created by an unknown source between March 18 and March 23.

Obviously, we can quibble about various details of this experiment, but the key takeaway is that negative SEO is a very real problem.

---

Matt Cutts posts on Google+ that sites were losing search rankings (see Barry Schwartz’s post above on April 17) due to a problem with Google’s “classifier for parked domains” (and not the “over optimization” update that people were fearing).

April 19, 2012

Google quietly releases a Panda data refresh (i.e., Panda Update 3.5).

---

Rand Fishkin posts a video that details preemptive actions that SEOs should take before the “over optimization” penalty goes live: 6 Changes Every SEO Should Make BEFORE the Over-Optimization Penalty Hits – Whiteboard Friday.

Rand identifies 6 items that should be addressed:

April 24, 2012

Google finally releases the highly anticipated “over optimization” update. Here’s the official announcement: Another step to reward high-quality sites.

According to the announcement, this algorithm change is primarily targeted at webspam, and it will affect about 3.1% of queries in English.

April 26, 2012

Google gives the new webspam algorithm change an official name: Penguin.

---

Rand Fishkin posts a video about negative SEO: Negative SEO: Myths, Realities, and Precautions – Whiteboard Friday.

In the video, Rand discusses various ways to negatively affect a site, including exploiting a security vulnerability in the site’s Web server or hosting platform (to propagate malware or spam) as well as submitting spam reports about a site to the search engines.

He spends the majority of the video discussing the negative SEO campaigns described on Traffic Planet (see the description above on April 18). As part of this discussion, he identifies factors that make a site more susceptible (e.g., existing spammy backlinks, a limited number of brand signals, manipulative on-page activities, etc.) and less susceptible (e.g., a clean backlink profile, a high-quality UX, strong brand signals, etc.) to negative SEO.

Finally, he offers a few suggestions for monitoring potential negative SEO campaigns against your site, and he advocates maintaining an open dialog with the search engines to help minimize any damage inflicted by those campaigns.

April 27, 2012

Google releases another Panda data refresh (i.e., Panda Update 3.6) only a week after the previous one.

---

Aaron Wall gives his take on the Penguin Update in this post: The Google Penguin Update: Over-Optimization, Webspam, & High Quality Empty Content Pages.

One of the first things Aaron points out is an odd result in the Google rankings for make money online. As this screenshot shows, the first organic result is a completely blank blogspot blog:

In the remainder of the post, Aaron shows other examples of unhelpful or useless Google results, and he warns that SEO requires “threading the needle” (i.e., finding the sweet spot between being aggressive enough to get results and being aggressive enough to get penalized).

May 2, 2012

The folks over at Microsite Masters present an analysis of the Penguin Update: Penguin Analysis: SEO Isn’t Dead, But You Need to Act Smarter (And 5 Easy Ways to Do So!).

In this post, they investigate thousands of websites (the exact number is undisclosed) and make the following observations:

As part of these observations, they also provide this graph:

Finally, they offer a few suggestions for moving forward post-Penguin. The best suggestions include the following:

---

Dr. Pete provides a very nice overview of the Penguin Update (and uses one of my favorite titles): Penguins, Pandas, and Panic at the Zoo.

The first part of the post describes many of the details covered previously (e.g., the update’s timeline, impact on queries, staying power, etc.). The second part focuses on generally applicable advice, including the following list of DOs:

He also includes a nice list of DON’Ts:
Finally, Dr. Pete concludes with the best advice in the entire post: “From a diagnostic standpoint, it definitely helps to know whether you were hit by Penguin or another update, but after that, you have to fix what’s in your power to fix… Isolate the damage, find the problems you can fix, and get to work fixing them.”

---

Richard Baxter gives helpful advice for recovering from Penguin: So Your Site’s Been Penalised? Now What?

In this post, Richard offers the following suggestions:

May 4, 2012

Modesto Siotos offers guidance about identifying suspicious link characteristics: How To Survive Google’s Unnatural Links Warnings & Avoid Over-optimisation.

Specifically, Modesto explains how to check for the following 4 things:

---

Eric Enge publishes an interesting video interview with Bruce Clay about link pruning: Link Pruning is the Key to Addressing Penguin.

In this interview, Bruce explains what link pruning is (“the identification and removal of unnatural, non-organic, or generally spammy links from a link profile”), the difficulties associated with getting links removed, and the reasons why a site might not return to pre-penalty rankings after a round of link pruning:

May 8, 2012

Rand Fishkin posts a video on Google+ about the Penguin Update: What’s Unique & Noteworthy About Google’s Penguin Update.

In the video, Rand makes various observations, and here are the biggest ones:

---

Jon Cooper writes a helpful guide about the types of links we should be acquiring post-Penguin: What type of links should we be getting?

In his post, Jon makes the point that post-Penguin link building techniques are still very similar to the approaches being advocated before the update. Yes, anchor text distributions need to be more diversified (and more brand-oriented), but many of the old link building principles are still sound.

You can read the specific techniques he describes in this post, but you’d be better served reading his extensive list of link building strategies instead: Link Building Strategies – The Complete List.

May 10, 2012

Danny Sullivan interviews Matt Cutts about the Penguin Update, recovery, and negative SEO: Two Weeks In, Google Talks Penguin Update, Ways To Recover & Negative SEO.

First, the interview discusses the successfulness of the update, and according to Matt, “It’s been a success from [Google’s] standpoint.” Next, the interview explains that reconsideration requests won’t help initiate a Penguin recovery because it’s an algorithmic update (as others had already explained).

Then, the interview turns to the topic of negative SEO, and the responses are somewhat unsatisfactory. Matt claims, “[Google has] done a huge amount of work to try to make sure one person can’t hurt another person.” However, he declined to comment on the Dan Thies situation (chronicled above in various posts and videos).

Finally, the interview offers the same advice we’ve seen previously: (1) clean up on-page spam, (2) clean up bad links, (3) wait for the next Penguin data refresh and see if you recover, and (4) only file a report if you genuinely believe your site is a false positive.

One interesting resource that is included at the end of the post is this list of WordPress plugins that insert hidden links into a blog: Google Penguin Targeted Many WordPress Blogs With Hidden Links In Plugins/Themes.

---

Rob Kerry posts a video about the Penguin Update: The Penguin Update – Whiteboard Friday.

In this video, Rob makes a few different observations about the update. Here are the highlights:

May 13, 2012

SEO Cliff creates a site for tracking Penguin Update news and resources: Google Penguin Update.

May 17, 2012

Barry Schwartz reports on a query hack that allows searchers to see pre-Penguin search results: Hit By Penguin? Google Query Hack To Confirm?

For example, the following query: blue widget -amazon returns the pre-Penguin results for blue widget.

---

Bryant Dunivan offers 10 link building tips for a post-Penguin world: 10 Post Penguin Link Building Tips.

Some of the tips overlap so here’s a merged summary:

May 18, 2012

Alex Pyatetsky writes an interesting post about the inconclusive nature of previous Penguin articles: 4500 Words on Everything We Don’t Know About Penguin and What Not to Do About It.

Alex begins his post by describing what the Penguin Update is supposed to be about: (1) over optimized anchor text profiles and (2) links from irrelevant sites. Then, he highlights a few examples of “awful” search engine results (the most notable example he gives is the make money online query mentioned above by Aaron Wall on April 27).

Next, Alex dives into the results presented by Microsite Masters (see May 2 above). He is somewhat dismissive of the individual graphs presented by Microsite Masters because they do not conclusively prove that over optimized anchor text and/or irrelevant links directly lead to a Penguin-related penalty.

However, I believe this analysis is a bit shortsighted since it assumes that the Penguin Update is a single feature classifier (or a simple combination of a small number of signals). It is far more likely that the update is a weighted combination of numerous features, which include the presence (and severity) of over optimized anchor text as well as the percentage (and authority) of irrelevant links. Thus, the Microsite Masters graphs are valuable because they offer at least some evidence that Google is putting more emphasis on the optimization level of anchor text and the absence of relevant links.

I also want to caution against drawing quantitative “best practice” conclusions from any of this data. In this particular case, it’s interesting that the Microsite Masters data shows a cut-off at 60% for over optimized anchor text, but this specific number will quickly become a moving target (similar to the “best practice” keyword densities from back in the day). The important takeaway is that over optimized anchor text is being scrutinized (and not the specific number).

Alex also expresses his disbelief that Google would want to devalue links from irrelevant sites, labeling it a “Wallstreet Journal penalty.” Basically, his argument (and this opinion is shared by many others) is that links from large sites such as The New York Times, The Huffington Post, and Tech Crunch would be considered irrelevant, resulting in a devaluation by Google. However, I have a different opinion, and I believe this argument misses the fundamental intent of an irrelevancy penalty (assuming one exists).

First of all, it’s important to understand that a link’s value is determined by a variety of different factors, including the authority of the linking site (and the linking page), the trustworthiness of the linking site (and the linking page), the position of the link on the linking page, etc. Now, if we add relevancy (of the linking site and the linking page) to that mix of factors (or simply give relevancy additional weight), it doesn’t eliminate everything else. Thus, “some blog roll from any indus­try friend” might be more relevant than “a New York Times link” (emphasis on the might – more on the meaning of relevancy in a moment), but its overall value still pales in comparison to the New York Times link based on the various other link-related factors. Additionally, a link from a New York Times article about blue widgets is still incredibly relevant for a site about blue widgets. No, The New York Times is not in the blue widgets niche, but that specific article definitely is (i.e., the link might not have site-level relevancy, but it has page-level relevancy).

At the end of the post, Alex offers numerous suggestions, including:

May 23, 2012

Jonah Stein gives his take on the Penguin Update: When Penguins Attack.

In his post, Jonah begins by echoing the same timeline concerns as many others: numerous updates were released in a very short window so your site’s concerns are only Penguin-related if you began losing traffic on April 24.

However, unlike others, Jonah also makes the argument that the Penguin Update wasn’t the only major change that Google made on April 24. This quote says it best: “Evidence suggests that at the same time Google announced a crackdown on over-optimization, they also turned up the sensitivity on the over-optimized anchor text filter and changed to broad match.”

Next, he gives a list of link building techniques that have been penalized (or might be penalized in the future), including purchasing links, participating in a blog network or other linking scheme, running a link exchange program, and registering your site in hundreds of directories.

Jonah also emphasizes the need to understand the difference between a natural link profile and one that has been built. He cites the following 3 characteristics as being predominantly associated with an unnatural link profile: (1) use of keywords in anchor text, (2) use of site-wide links, and (3) inclusion in directories, article sites, and blog networks.

Here is another striking quote from the article: “With Penguin, sites are now being penalized for the cumulative debris of years of evolving link-building strategies using tactics that were previously successful and commonly considered necessary to survive.”

Finally, Jonah offers a familiar recommendation: clean your link profile, and strive for high-quality editorial links.

May 24, 2012

The query hack for displaying pre-Penguin search results (originally reported on May 17) stops working.

---

Jonathan Leger presents his findings about the Penguin Update’s impact (or lack thereof) on exact match domains (EMDs): Anchor Text Optimization – How Much Is Too Much?

In this post, Jonathan conducts an experiment to quantify the amount of exact match anchor text in the search engine results post-Penguin. As part of the experiment, he gathered 1,500 keywords from 30 different markets, including business, technology, transportation, food, etc.

Then, Jonathan submitted the keywords to Google to obtain the top-level domains ranking on the first page for each keyword. Finally, he checked each domain’s link profile to identify the percentage of links containing exact match anchor text.

Based on this experiment, Jonathan makes the following observations:

Jonathan’s hypothesis for why these EMDs are ranking (despite their high percentages of exact match anchor text) is that the exact match anchor text is being treated like brand-related anchor text. Using this logic, he essentially argues that if Google penalized bluewidgets.com for “blue widgets” anchor text, it would also need to penalize amazon.com for “amazon” anchor text.

May 25, 2012

Nick Eubanks details his experience with an over optimization penalty that affected his personal blog: Recovering from an Over Optimization Penalty – A True Story.

In this case study, Nick identifies a site-wide over optimized sidebar link as the culprit for his site’s penalty. After he changed the offending anchor text to a somewhat naked URL, his site recovered. Nick’s primary takeaway in this post is that you should make your link profile as natural as possible (i.e., use naked URLs, “click here”, “here”, brand names, etc.).

May 26, 2012

Google releases its first Penguin data refresh (i.e., Penguin Update 1.1).

May 28, 2012

Ross Hudgens writes a very helpful case study about WPMU.org’s recovery from the Penguin Update: How WPMU.org Recovered From The Penguin Update.

Ross begins the post by giving WPMU.org’s story. The site was hit by the Penguin Update, which resulted in a 81% drop in traffic (week over week).

After an initial investigation, Ross identifies “WordPress MU” as the most popular anchor text for the site. This anchor text was generated by WPMU’s themes, which include a “Powered by” link in the footer that points back to the corresponding theme’s page on WPMU. As Ross points out, these themes “generated a high volume of sitewide links on low quality sites.”

Fortunately, many of the site-wide links came from a single site (EDUblogs.org). And since this site is owned by the same owner as WPMU.org, they were able to instantly remove almost 15,000 site-wide, footer linking root domains from their link profile.

Ross also highlights a few changes made by WPMU that were not necessarily link-related, which include implementing canonical URLs to clean up crawl errors with the site, updating and submitting XML Sitemaps, changing duplicate titles tags, and more.

Finally, Ross reports the actual recovery that came with Penguin Update 1.1. This screenshot is quite compelling:

June 1, 2012

The post you’re currently reading goes live 😉

What Do You Think?

I would love to hear your thoughts about all of this craziness. What events really stuck out in your mind from the past few months? What do you expect in the months to come?