Google’s major local algorithm update, dubbed “Pigeon”, expanded to the United Kingdom, Canada, and Australia. The original update hit the United States in July 2014. The update was confirmed on the 22nd but may have rolled out as early as the 19th.
A Google representative said that Penguin had shifted to continuous updates, moving away from infrequent, major updates. While the exact timeline was unclear, this claim seemed to fit ongoing flux after Penguin 3.0 (including unconfirmed claims of a Penguin 3.1).
More than two years after the original DMCA/”Pirate” update, Google launched another update to combat software and digital media piracy. This update was highly targeted, causing dramatic drops in ranking to a relatively small group of sites.
More than a year after the previous Penguin update (2.1), Google launched a Penguin refresh. This update appeared to be smaller than expected (<1% of US/English queries affected) and was probably data-only (not a new Penguin algorithm). The timing of the update was unclear, especially internationally, and Google claimed it was spread out over “weeks”.
Google made what looked like a display change to News-box results, but later announced that they had expanded news links to a much larger set of potential sites. The presence of news results in SERPs also spiked, and major news sites reported substantial traffic changes.
Following up on the June 28th drop of authorship photos, Google announced that they would be completely removing authorship markup (and would no longer process it). By the next morning, authorship bylines had disappeared from all SERPs.
After months of speculation, Google announced that they would be giving preference to secure sites, and that adding encryption would provide a “lightweight” rankings boost. They stressed that this boost would start out small, but implied it might increase if the changed proved to be positive.
Google shook the local SEO world with an update that dramatically altered some local results and modified how they handle and interpret location cues. Google claimed that Pigeon created closer ties between the local algorithm and core algorithm(s).
John Mueller made a surprise announcement (on June 25th) that Google would be dropping all authorship photos from SERPs (after heavily promoting authorship as a connection to Google+). The drop was complete around June 28th.
Less than a month after the Payday Loan 2.0 anti-spam update, Google launched another major iteration. Official statements suggested that 2.0 targeted specific sites, while 3.0 targeted spammy queries.
Google confirmed a major Panda update that likely included both an algorithm update and a data refresh. Officially, about 7.5% of English-language queries were affected. While Matt Cutts said it began rolling out on 5/20, our data strongly suggests it started earlier.
Just prior to Panda 4.0, Google updated it’s “payday loan” algorithm, which targets especially spammy queries. The exact date of the roll-out was unclear (Google said “this past weekend” on 5/20), and the back-to-back updates made the details difficult to sort out.
Major algorithm flux trackers and webmaster chatter spiked around 3/24-3/25, and some speculated that the new, “softer” Panda update had arrived. Many sites reported ranking changes, but this update was never confirmed by Google.
As predicted by Matt Cutts at Pubcon Las Vegas, authorship mark-up disappeared from roughly 15% of queries over a period of about a month. The fall bottomed out around December 19th, but the numbers remain volatile and have not recovered to earlier highs.
Almost all global flux trackers registered historically high activity. Google would not confirm an update, suggesting that they avoid updates near the holidays. MozCast also registered a rise in some Partial-Match Domains (PMDs), but the patterns were unclear.
Multiple Google trackers picked up unusual activity, which co-occurred with a report of widespread DNS errors in Google Webmaster Tools. Google did not confirm an update, and the cause and nature of this flux was unclear.
After a 4-1/2 month gap, Google launched another Penguin update. Given the 2.1 designation, this was probably a data update (primarily) and not a major change to the Penguin algorithm. The overall impact seemed to be moderate, although some webmasters reported being hit hard.
Announced on September 26th, Google suggested that the “Hummingbird” update rolled out about a month earlier. Our best guess ties it to a MozCast spike on August 20th and many reports of flux from August 20-22. Hummingbird has been compared to Caffeine, and seems to be a core algorithm update that may power changes to semantic search and the Knowledge Graph for months to come.
Google added a new type of news result called “in-depth articles”, dedicated to more evergreen, long-form content. At launch, it included links to three articles, and appeared across about 3% of the searches that MozCast tracks.
Google confirmed a Panda update, but it was unclear whether this was one of the 10-day rolling updates or something new. The implication was that this was algorithmic and may have “softened” some previous Panda penalties.
Google’s Matt Cutts tweeted a reply suggesting a “multi-week” algorithm update between roughly June 12th and “the week after July 4th”. The nature of the update was unclear, but there was massive rankings volatility during that time period, peaking on June 27th (according to MozCast data). It appears that Google may have been testing some changes that were later rolled back.
Google announced a targeted algorithm update to take on niches with notoriously spammy results, specifically mentioning payday loans and porn. The update was announced on June 11th, but Matt Cutts suggested it would roll out over a 1-2 month period.
While not an actual Panda update, Matt Cutts made an important clarification at SMX Advanced, suggesting that Panda was still updating monthly, but each update rolled out over about 10 days. This was not the “everflux” many people had expected after Panda #25.
After months of speculation bordering on hype, the 4th Penguin update (dubbed “2.0” by Google) arrived with only moderate impact. The exact nature of the changes were unclear, but some evidence suggested that Penguin 2.0 was more finely targeted to the page level.
Google released an update to control domain crowding/diversity deep in the SERPs (pages 2+). The timing was unclear, but it seemed to roll out just prior to Penguin 2.0 in the US and possibly the same day internationally.
In the period around May 9th, there were many reports of an algorithm update (also verified by high MozCast activity). The exact nature of this update was unknown, but many sites reported significant traffic loss.
Matt Cutts pre-announced a Panda update at SMX West, and suggested it would be the last update before Panda was integrated into the core algorithm. The exact date was unconfirmed, but MozCast data suggests 3/13-3/14.
Right before the Christmas holiday, Google rolled out another Panda update. They officially called it a “refresh”, impacting 1.3% of English queries. This was a slightly higher impact than Pandas #21 and #22.
Google added Knowledge Graph functionality to non-English queries, including Spanish, French, German, Portuguese, Japanese, Russian, and Italian. This update was “more than just translation” and added enhanced KG capabilities.
Google announced an update to its original page layout algorithm change back in January, which targeted pages with too many ads above the fold. It’s unclear whether this was an algorithm change or a Panda-style data refresh.
After suggesting the next Penguin update would be major, Google released a minor Penguin data update, impacting “0.3% of queries”. Penguin update numbering was rebooted, similar to Panda – this was the 3rd Penguin release.
Google published their monthly (bi-monthly?) list of search highlights. The 65 updates for August and September included 7-result SERPs, Knowledge Graph expansion, updates to how “page quality” is calculated, and changes to how local results are determined.
Overlapping the EMD update, a fairly major Panda update (algo + data) rolled out, officially affecting 2.4% of queries. As the 3.X series was getting odd, industry sources opted to start naming Panda updates in order (this was the 20th).
Exact-Match Domain (EMD) Update — September 27, 2012
Google announced a change in the way it was handling exact-match domains (EMDs). This led to large-scale devaluation, reducing the presence of EMDs in the MozCast data set by over 10%. Official word is that this change impacted 0.6% of queries (by volume).
Google made a significant change to the Top 10, limiting it to 7 results for many queries. Our research showed that this change rolled out over a couple of days, finally impacting about 18% of the keywords we tracked.
After a summer hiatus, the June and July Search Quality Highlights were rolled out in one mega-post. Major updates included Panda data and algorithm refreshes, an improved rank-ordering function (?), a ranking boost for “trusted sources”, and changes to site clustering.
In a repeat of March/April, Google sent out a large number of unnatural link warnings via Google Webmaster Tools. In a complete turn-around, they then announced that these new warnings may not actually represent a serious problem.
Google rolled out yet another Panda data update, claiming that less than 1% of queries were affect. Ranking fluctuation data suggested that the impact was substantially higher than previous Panda updates (3.5, 3.6).
Google released their monthly Search Highlights, with 39 updates in May. Major changes included Penguin improvements, better link-scheme detection, changes to title/snippet rewriting, and updates to Google News.
In a major step toward semantic search, Google started rolling out “Knowledge Graph”, a SERP-integrated display providing supplemental object about certain people, places, and things. Expect to see “knowledge panels” appear on more and more SERPs over time. Also, Danny Sullivan’s favorite Trek is ST:Voyager?!
Google published details of 52 updates in April, including changes that were tied to the “Penguin” update. Other highlights included a 15% larger “base” index, improved pagination handling, and a number of updates to sitelinks.
After weeks of speculation about an “Over-optimization penalty”, Google finally rolled out the “Webspam Update”, which was soon after dubbed “Penguin.” Penguin adjusted a number of spam factors, including keyword stuffing, and impacted an estimated 3.1% of English queries.
In the middle of a busy week for the algorthim, Google quietly rolled out a Panda data update. A mix of changes made the impact difficult to measure, but this appears to have been a fairly routine update with minimal impact.
After a number of webmasters reported ranking shuffles, Google confirmed that a data error had caused some domains to be mistakenly treated as parked domains (and thereby devalued). This was not an intentional algorithm change.
Google posted another batch of update highlights, covering 50 changes in March. These included confirmation of Panda 3.4, changes to anchor-text “scoring”, updates to image search, and changes to how queries with local intent are interpreted.
This wasn’t an algorithm update, but Google published a rare peek into a search quality meeting. For anyone interested in the algorithm, the video provides a lot of context to both Google’s process and their priorities. It’s also a chance to see Amit Singhal in action.
As part of their monthly update, Google mentioned code-name “Venice”. This local update appeared to more aggressively localize organic results and more tightly integrate local search data. The exact roll-out date was unclear.
Google published a second set of “search quality highlights” at the end of the month, claiming more than 40 changes in February. Notable changes included multiple image-search updates, multiple freshness updates (including phasing out 2 old bits of the algorithm), and a Panda update.
Google released another round of “search quality highlights” (17 in all). Many related to speed, freshness, and spell-checking, but one major announcement was tighter integration of Panda into the main search index.
Google updated their page layout algorithms to devalue sites with too much ad-space above the “fold”. It was previously suspected that a similar factor was in play in Panda. The update had no official name, although it was referenced as “Top Heavy” by some SEOs.
Google announced a radical shift in personalization – aggressively pushing Google+ social data and user profiles into SERPs. Google also added a new, prominent toggle button to shut off personalization.
Google announced 30 changes over the previous month, including image search landing-page quality detection, more relevant site-links, more rich snippets, and related-query improvements. The line between an “algo update” and a “feature” got a bit more blurred.
Google outlined a second set of 10 updates, announcing that these posts would come every month. Updates included related query refinements, parked domain detection, blog search freshness, and image search freshness. The exact dates of each update were not provided.
After Panda 2.5, Google entered a period of “Panda Flux” where updates started to happen more frequently and were relatively minor. Some industry analysts called the 11/18 update 3.1, even though there was no official 3.0. For the purposes of this history, we will discontinue numbering Panda updates except for very high-impact changes.
This one was a bit unusual. In a bid to be more transparent, Matt Cutts released a post with 10 recent algorithm updates. It’s not clear what the timeline was, and most were small updates, but it did signal a shift in how Google communicates algorithm changes.
Google announced that an algorithm change rewarding freshness would impact up to 35% of queries (almost 3X the publicly stated impact of Panda 1.0). This update primarly affected time-sensitive results, but signalled a much stronger focus on recent content.
Google announced they would be encrypting search queries, for privacy reasons. Unfortunately, this disrupted organic keyword referral data, returning “(not provided)” for some organic traffic. This number increased in the weeks following the launch.
To help fix crawl and duplication problems created by pagination, Google introduced the rel=”next” and rel=”prev” link attributes. Google also announced that they had improved automatic consolidation and canonicalization for “View All” pages.
After experimenting for a while, Google officially rolled out expanded site-links, most often for brand queries. At first, these were 12-packs, but Google appeared to limit the expanded site-links to 6 shortly after the roll-out.
Google rolled Panda out internationally, both for English-language queries globally and non-English queries except for Chinese, Japanese, and Korean. Google reported that this impacted 6-9% of queries in affected countries.
After a number of social media failures, Google launched a serious attack on Facebook with Google+. Google+ revolved around circles for sharing content, and was tightly integrated into products like Gmail. Early adopters were quick to jump on board, and within 2 weeks Google+ reached 10M users.
Google continued to update Panda-impacted sites and data, and version 2.2 was officially acknowledged. Panda updates occurred separately from the main index and not in real-time, reminiscent of early Google Dance updates.
Google, Yahoo and Microsoft jointly announced support for a consolidated approach to structured data. They also created a number of new “schemas”, in an apparent bid to move toward even richer search results.
Google rolled out the Panda update to all English queries worldwide (not limited to English-speaking countries). New signals were also integrated, including data about sites users blocked via the SERPs directly or the Chrome browser.
Responding to competition by major social sites, including Facebook and Twitter, Google launched the +1 button (directly next to results links). Clicking [+1] allowed users to influence search results within their social circle, across both organic and paid results.
A major algorithm update hit sites hard, affecting up to 12% of search results (a number that came directly from Google). Panda seemed to crack down on thin content, content farms, sites with high ad-to-content ratios, and a number of other quality issues. Panda rolled out over at least a couple of months, hitting Europe in April 2011.
In response to high-profile spam cases, Google rolled out an update to help better sort out content attribution and stop scrapers. According to Matt Cutts, this affected about 2% of queries. It was a clear precursor to the Panda updates.
In a rare turn of events, a public outing of shady SEO practices by Overstock.com resulted in a very public Google penalty. JCPenney was hit with a penalty in February for similar bad behavior. Both situations represented a shift in Google’s attitude and foreshadowed the Panda update.
After an expose in the New York Times about how e-commerce site DecorMyEyes was ranking based on negative reviews, Google made a rare move and reactively adjusted the algorithm to target sites using similar tactics.
Google and Bing confirmed that they use social signals in determining ranking, including data from Twitter and Facebook. Matt Cutts confirmed that this was a relatively new development for Google, although many SEOs had long suspected it would happen.
A magnifying glass icon appeared on Google search results, allowing search visitors to quickly view a preview of landing pages directly from SERPs. This signaled a renewed focus for Google on landing page quality, design, and usability.
Expanding on Google Suggest, Google Instant launched, displaying search results as a query was being typed. SEOs everywhere nearly spontaneously combusted, only to realize that the impact was ultimately fairly small.
Although not a traditional algorithm update, Google started allowing the same domain to appear multiple times on a SERP. Previously, domains were limited to 1-2 listings, or 1 listing with indented results.
After months of testing, Google finished rolling out the Caffeine infrastructure. Caffeine not only boosted Google’s raw speed, but integrated crawling and indexation much more tightly, resulting in (according to Google) a 50% fresher index.
In late April and early May, webmasters noticed significant drops in their long-tail traffic. Matt Cutts later confirmed that May Day was an algorithm change impacting the long-tail. Sites with large-scale thin content seemed to be hit especially hard, foreshadowing the Panda update.
Although “Places” pages were rolled out in September of 2009, they were originally only a part of Google Maps. The official launch of Google Places re-branded the Local Business Center, integrated Places pages more closely with local search results, and added a number of features, including new local advertising options.
This time, real-time search was for real- Twitter feeds, Google News, newly indexed content, and a number of other sources were integrated into a real-time feed on some SERPs. Sources continued to expand over time, including social media.
Google released a preview of a massive infrastructure change, designed to speed crawling, expand the index, and integrate indexation and ranking in nearly real-time. The timeline spanned months, with the final rollout starting in the US in early 2010 and lasting until the summer.
In a major change to their logo-and-a-box home-page Google introduced Suggest, displaying suggested searches in a dropdown below the search box as visitors typed their queries. Suggest would later go on to power Google Instant.
A large-scale shuffle seemed to occur at the end of March and into early April, but the specifics were unclear. Some suspected Google was pushing its own internal properties, including Google Books, but the evidence of that was limited.
While not your typical algorithm update, Google integrated traditional search results with News, Video, Images, Local, and other verticals, dramatically changing their format. The old 10-listing SERP was officially dead. Long live the old 10-listing SERP.
Throughout 2006, Google seemed to make changes to the supplemental index and how filtered pages were treated. They claimed in late 2006 that supplemental was not a penalty (even if it sometimes felt that way).
Technically, Big Daddy was an infrastructure update (like the more recent “Caffeine”), and it rolled out over a few months, wrapping up in March of 2006. Big Daddy changed the way Google handled URL canonicalization, redirects (301/302) and other technical issues.
After launching the Local Business Center in March 2005 and encouraging businesses to update their information, Google merged its Maps data into the LBC, in a move that would eventually drive a number of changes in local SEO.
Google released a series of updates, mostly targeted at low-quality links, including reciprocal links, link farms, and paid links. Jagger rolled out in at least 3 stages, from roughly September to November of 2005, with the greatest impact occurring in October.
Also called the “False” update ? webmasters saw changes (probably ongoing), but Google claimed no major algorithm update occurred. Matt Cutts wrote a blog post explaining that Google updated (at the time) index data daily but Toolbar PR and some other metrics only once every 3 months.
Unlike previous attempts at personalization, which required custom settings and profiles, the 2005 roll-out of personalized search tapped directly into users? search histories to automatically adjust results. Although the impact was small at first, Google would go on to use search history for many applications.
“GoogleGuy” (likely Matt Cutts) announced that Google was rolling out “something like 3.5 changes in search quality.” No one was sure what 0.5 of a change was, but Webmaster World members speculated that Bourbon changed how duplicate content and non-canonical (www vs. non-www) URLs were treated.
Webmasters witnessed ranking changes, but the specifics of the update were unclear. Some thought Allegra affected the “sandbox” while others believed that LSI had been tweaked. Additionally, some speculated that Google was beginning to penalize suspicious links.
To combat spam and control outbound link quality, Google, Yahoo, and Microsoft collectively introduce the “nofollow” attribute. Nofollow helps clean up unvouched for links, including spammy blog comments. While not a traditional algorithm update, this change gradually has a significant impact on the link graph.
Although obviously not an algorithm update, a major event in Google’s history – Google sold 19M shares, raised $1.67B in capital, and set their market value at over $20B. By January 2005, Google share prices more than doubled.
Google rolled out a variety of changes, including a massive index expansion, Latent Semantic Indexing (LSI), increased attention to anchor text relevance, and the concept of link “neighborhoods.” LSI expanded Google’s ability to understand synonyms and took keyword analysis to the next level.
What Florida missed, Austin came in to clean up. Google continued to crack-down on deceptive on-page tactics, including invisible text and META-tag stuffing. Some speculated that Google put the “Hilltop” algorithm into play and began to take page relevance seriously.
This was the update that put updates (and probably the SEO industry) on the map. Many sites lost ranking, and business owners were furious. Florida sounded the death knell for low-value late 90s SEO tactics, like keyword stuffing, and made the game a whole lot more interesting.
In order to index more documents without sacrificing performance, Google split off some results into the “supplemental” index. The perils of having results go supplemental became a hotly debated SEO topic, until the index was later reintegrated.
The monthly “Google Dance” finally came to an end with the “Fritz” update. Instead of completely overhauling the index on a roughly monthly basis, Google switched to an incremental approach. The index was now changing daily.
This marked the last of the regular monthly Google updates, as a more continuous update process began to emerge. The “Google Dance” was replaced with “Everflux”. Esmerelda probably heralded some major infrastructure changes at Google.
While many changes were observed in May, the exact nature of Dominic was unclear. Google bots “Freshbot” and “Deepcrawler” scoured the web, and many sites reported bounces. The way Google counted or reported backlinks seemed to change dramatically.
Announced at SES Boston, this was the first named Google update. Originally, Google aimed at a major monthly update, so the first few updates were a combination of algorithm changes and major index refreshes (the so-called “Google Dance”). As updates became more frequent, the monthly idea quickly died.
1st Documented Update — September 2002
Before “Boston” (the first named update), there was a major shuffle in the Fall of 2002. The details are unclear, but this appeared to be more than the monthly Google Dance and PageRank update. As one webmaster said of Google: “they move the toilet mid stream”.