The next wave of Panda releases starts with Panda2.1
After the game-changing Panda2.0 update that caused a reduction in Google search engine referral traffic to a newly affected and larger set of websites, more Panda updates were already in the pipeline from Google. The next update to Panda, Panda2.1, came about a month later on May 10th, 2011. Many search engine industry watchers thought it was a major change and began calling it Panda3.0, but Google called it Panda2.1, implying that it was more a minor update or tweak on the same type of changes that were made in the Panda2.0 update. Calling it a minor tweak doesn’t diminish the fact that the Panda2.1 update still affected a large number of search queries and continued to create havoc by reducing formerly steady amounts of search referral traffic to well established websites.
Machine Learning Algorithm
Information began to come out from Google and other Google-connected search engine industry watchers indicating that the Panda updates were accomplished with a machine learning algorithm. Without getting too much into the details of machine learning algorithms
, Google was taking big sets of data and feeding that data into the Panda ranking algorithm and letting it crunch the data for a long period (week or weeks). Once that data and number crunching was done, Google engineers would then review, test over several days, and then release widely the new ranking data to affect search results. Each time that Google ran through this cycle it would take 4-5 weeks, and each release would become known as Panda2.0 iterations with the release of Panda2.1, Panda2.2, Panda2.3, and the latest release as of the writing of this article, Panda2.4.
Panda 2.2, 2.3, and 2.4 releases
The next update in line was Panda2.2, which was released on June 16th, 2011. Google’s Matt Cutts indicated
that Panda2.2 would target websites that scraped and re-published content that was outranking the originally published content. These scrapers were and still are a continual thorn in the side of Google and website publishers, and Google was able to placate angry website publishers with the tweaks in this release. About a month after Panda2.2, Panda2.3 was released on July 23rd, 2011. The Panda2.3 release was noted as a smaller release that tweaked Panda's signals to differentiate between low-quality and high-quality websites, and caused some websites to increase and some to decrease in traffic. The latest current release at the time of this article being published is Panda2.4, which was released on August 12th, 2011. Panda 2.4 was notable because it contained the Panda algorithm updates for all non-English languages except for Japanese, Chinese, and Korean languages. Those three languages are still in the testing phase in regards to Google’s Panda machine learning algorithm and how it affects Google search results in those languages. It was also noted that depending on the language, between 6-9% of searches would be affected by the Panda2.4 changes. This is less than the 12% of search queries affected when Panda was rolled out to English language searches, but still a significant algorithm update.
Have any websites recovered from the Panda Update? How?
A very small number of websites have reported recovering from Pandalization
(Google Panda algorithm penalty) following the Panda2.2 update. One large website, hubpages.com
, (profiled on a Wall Street Journal blog
) reported having recovered their search engine referrals from Google after Panda2.3 by eliminating some “low-quality” content, but also moving significant amounts of their content to multiple subdomains. Mentioned in the WSJ blog was Matt Cutts’ recommendation of this strategy as a way to separate out content into different categories to avoid having one set of content taint other sets of content. For example, topicA.hubpost.com
would be separated out for Google’s consumption from topicB.hubspot.com
, and Google would treat those two subdomains as different domains, and rank the content under each domain separately in all areas, including the Panda algorithm. This could allow the topicA area to grow and flourish if it had “high-quality” content even if Google Pandalized
the topicB area because of “low-quality” content. Having all of the content of topicA and topicB combined under hubspot.com would likely mean all of it would be Pandalized
, rather than just topicB when subdomains are used. Other than hubpages.com and one or two other reports, there have been a notable lack of Panda-related recoveries in search engine referrals since the original Panda 1.0 update continuing through the Panda2.4 update time period.
More on the overall ramifications and future of Google’s Panda update series in Part 5 of the Google Panda Algorithm Update.