The Google Panda update launched on 23 February 2011 and, in its first US release, noticeably affected 11.8% of queries. It was a site-level quality classifier aimed at content farms and thin, duplicate or ad-heavy sites, and it ran as roughly 28 confirmed versions and refreshes until 2015, when Google folded it into its core ranking systems. Panda no longer runs on its own, but the way it judged a whole site by the quality of its pages is still how core updates behave.
Panda is one of the most important entries in my full list of Google algorithm updates. Below you will find every version with its date, what Panda looked for, and a recovery process that still works for sites hit by core updates today.
| Detail | Information |
|---|---|
| Update name | Google Panda Update (first nicknamed “Farmer”) |
| Type | Content quality (site-level classifier) |
| Rollout started | 23 February 2011 (Panda 1.0, US English) |
| Rollout finished | No single end date: last confirmed refresh Panda 4.2 from 18 July 2015; became part of core ranking in 2015, confirmed publicly in January 2016 |
| Rollout length | Roughly 28 confirmed versions and refreshes over 2011 to 2015 (trackers number them differently) |
| Confirmed by Google | Yes: the launch and most versions were confirmed; Panda #25 was pre-announced but not formally confirmed |
| Official source | Search Engine Land coverage of Google’s launch announcement |
What the Panda update changed
Before Panda, a site could rank by publishing huge volumes of short articles targeted at popular searches. Content farms did exactly that, and their pages often outranked better sources. Google’s answer was a classifier that scored the quality of a site as a whole and used that score in ranking.
Three things made Panda different from what came before:
- It worked at site level. Google described it as “an algorithm that’s applied to sites overall”. A large share of weak pages could pull down the strong pages on the same domain.
- It rewarded originality and depth. The stated aim was to reduce low-quality sites and reward sites with original content and in-depth analysis.
- It ran in batches. Until it was folded into core ranking, a site’s Panda score was only recalculated when Google ran a refresh, so changes on your site showed up weeks or months later.
Search Engine Land called it “Farmer” at first. Google later named it after Navneet Panda, the engineer behind it.
In May 2011 Google’s Amit Singhal published 23 questions describing what a high-quality site looks like. They asked, among other things, whether you would trust the information, whether it was written by an expert, whether the site had duplicate or overlapping articles on similar topics, whether it had spelling or style problems, whether it offered original analysis, and whether it had excessive ads. Those questions are the best guide to what Panda was measuring.
Timeline
Dates below are the most commonly cited ones. Where trackers disagree, I have noted it.
| Version | Date | Notes |
|---|---|---|
| Panda 1.0 (Farmer) | 23 Feb 2011 | US English launch; 11.8% of queries noticeably affected. Google’s post was dated 24 Feb. |
| Panda 2.0 | 11 Apr 2011 | All English queries worldwide; added signals such as sites users blocked. About 2% of US queries. |
| Panda 2.1 | 9 May 2011 | Industry called it “3.0”; Google said 2.1. |
| Panda 2.2 | 21 Jun 2011 | Improved scraper detection. Search Engine Land lists about 16 June. |
| Panda 2.3 | 23 Jul 2011 | New signals to separate higher- and lower-quality sites. |
| Panda 2.4 | 12 Aug 2011 | Expanded to all languages except Chinese, Japanese and Korean. |
| Panda 2.5 | 28 Sep 2011 | Followed by “Panda flux” around 3 and 13 October. Wikipedia lists 30 Sept. |
| Panda 3.0 | 19 Oct 2011 | New signals; recalculated Panda’s effects. |
| Panda 3.1 | 18 Nov 2011 | Minor; under 1% of searches. |
| Panda 3.2 | 18 Jan 2012 | Data refresh. |
| Panda 3.3 | 27 Feb 2012 | Data refresh. SEJ lists 23 Feb. |
| Panda 3.4 | 23 Mar 2012 | About 1.6% of queries. |
| Panda 3.5 | 19 Apr 2012 | Data refresh; minimal impact. |
| Panda 3.6 | 27 Apr 2012 | Data refresh; minimal impact. |
| Panda 3.7 | 8 Jun 2012 | Google said under 1% of US queries; trackers saw more. |
| Panda 3.8 | 25 Jun 2012 | About 1% worldwide. Ignite lists 18 June. |
| Japanese and Korean | 18 Jul 2012 | About 5% of queries in those languages. |
| Panda 3.9 | 24 Jul 2012 | About 1% of results. |
| Panda 3.9.1 | 20 Aug 2012 | Under 1% of queries. |
| Panda 3.9.2 | 18 Sep 2012 | 0.7% of queries; last of the 3.x numbering. |
| Panda #20 | 27 Sep 2012 | About 2.4% of queries (Wikipedia); same week as the EMD update. |
| Panda #21 | 5 Nov 2012 | About 1.1% of English queries. |
| Panda #22 | 21 Nov 2012 | Data refresh. |
| Panda #23 | 21 Dec 2012 | About 1.3% of English queries. |
| Panda #24 | 22 Jan 2013 | About 1.2% of English queries. |
| Panda #25 | 14 Mar 2013 | Pre-announced by Matt Cutts; not formally confirmed. Panda to roll out more continuously. |
| Panda “Recovery” (#26) | 18 Jul 2013 | More finely targeted; softened some earlier effects. |
| Panda 4.0 | 20 May 2014 | About 7.5% of English queries; “gentler” for some sites. SEJ lists 19 May. |
| Panda 4.1 | 25 Sep 2014 | Announced 25 Sept, rollout from about 23 Sept; 3 to 5% of queries; meant to help good small and medium sites. |
| Panda 4.2 | 18 Jul 2015 | Last confirmed refresh; 2 to 3% of English queries; took months to roll out. Some list 17 July. |
| Part of core ranking | 11 Jan 2016 | Google confirmed Panda was a core ranking signal. Its ranking systems guide says this happened in 2015. |
Who was affected
As reported across the industry, the sites that lost most were:
- Content farms publishing thousands of short, formulaic articles on popular searches.
- Article directories where anyone could post articles, mostly to get a link.
- Scraper and aggregator sites that republished other people’s content with little of their own.
- Sites with many thin pages: near-empty tag and archive pages, doorway-style location pages, product pages with only a manufacturer description.
- Ad-heavy sites where content was hard to find among adverts.
Ordinary businesses were caught too, usually because one weak section dragged down the rest. The first release was US English only; Panda 2.0 took it to all English queries, 2.4 to most other languages, and July 2012 to Japanese and Korean.
How to tell if the Panda update hit your site
Panda as a separate refresh is gone, so nobody is newly hit by “Panda” today. But the same pattern appears after core updates, and the diagnosis is the same.
- Match the date. For a historical drop, line your analytics up against the versions above. A Panda hit usually showed as a sharp step down within days of a refresh date, not a slow decline. Today, compare Search Console clicks for the two weeks before a core update started with the two weeks after it finished.
- Check whether the drop is site-wide. This is the Panda signature. If almost every section fell by a similar proportion, including good pages, the problem is a site-level quality judgement rather than individual pages.
- Count your pages. Compare the number of indexed URLs (Search Console Page Indexing report, or a
site:search as a rough guide) with the number of pages you would genuinely want a visitor to land on. A site with 400 useful pages and 9,000 indexed URLs has a quality ratio problem. - Find pages that earn nothing. Export a year of page data from Search Console. Pages with no clicks and almost no impressions over 12 months are your first candidates for review.
- Rule out other causes. Check for manual actions, migration or tracking changes, and seasonality. Penguin (links) launched in April 2012 and the two were often confused; a link problem tends to hit specific keywords and pages rather than the whole site evenly.
How to recover from the Panda update
Recovery from Panda was, and from core updates still is, slow and earned. Before 2016 it usually showed only at the next refresh, so sites waited months. Now improvements are picked up through core ranking over time and often show around the next core update. The work itself has not changed much.
Step 1: Audit every indexable URL
Crawl the site and join it with Search Console and analytics data. Put each URL into one of four groups: keep as is, improve, merge with another page, or remove/noindex. Start with templates (tag pages, archives, filtered URLs), where most thin pages hide.
Step 2: Improve before you delete
Google’s later advice is worth taking literally. Gary Illyes said in 2017: “We don’t recommend removing content in general for Panda, rather add more HighQ stuff.” So the first move for a page with real search demand is to make it better: answer the question fully, add first-hand detail, examples, original photos or data, and a named author with relevant experience.
Step 3: Consolidate overlapping pages
Singhal’s questions asked whether a site has duplicate, overlapping or redundant articles on the same topic with slight keyword variations. If you have five posts on nearly the same thing, merge them into one strong page and 301 redirect the others.
Step 4: Noindex or remove what has no value
Pages with no purpose for a visitor (empty tag pages, auto-generated location pages, scraped descriptions) should be noindexed or removed. This is where technical SEO matters: robots rules, canonical tags and parameter handling stop the problem coming back.
Step 5: Fix ads and layout
Reduce excessive ads and boilerplate so the main content is the first thing people see. Raise editorial standards: proofreading, sourcing, clear authorship.
Finally, ask Matt Cutts’s question: “How compelling is my site?” If a section exists only to catch search traffic, it is a liability.
| Panda-type problem | Fix |
|---|---|
| Thousands of near-empty tag and archive pages | Noindex or remove; keep a few well-curated hubs |
| Product pages using manufacturer text | Write original descriptions for best sellers first; add specs, sizing, photos, FAQs |
| Several short posts on one topic | Merge into one in-depth guide and redirect |
| City pages with swapped place names | Keep only locations you serve, with genuinely local detail |
| Ads above and inside the content | Cut ad units, move them below the main content |
For shops, the product page problem is the most common one I see, and it is the core of my ecommerce SEO work. Expect recovery to take months, and do not expect it if the work is cosmetic.
Is it still relevant today?
Panda no longer exists as a separate system. Google’s ranking systems guide lists it as retired: it became part of core ranking systems in 2015, confirmed publicly in January 2016. Gary Illyes said the core-integrated version is “not real time”, which matches what we see: quality reassessments still tend to show up around core updates.
What lives on is the idea that Google judges a site partly as a whole. The helpful content work of 2022 to 2024 and every core update since carry the same message Panda did in 2011: lots of pages made only to rank can weigh down the pages that deserve to.
Frequently asked questions
When did the Panda update start?
Panda 1.0 went live around 23 February 2011 in US English. Google’s announcement was dated 24 February and said it had gone live “in the last day or so”.
How many Panda updates were there?
About 28 confirmed versions and refreshes between 2011 and 2015, ending with Panda 4.2 in July 2015. Trackers number them differently.
Is Google Panda still active?
Not as a separate algorithm. Google says it became part of its core ranking systems in 2015. Its quality judgements continue through core updates.
Should I delete thin content to recover from Panda?
Improve first. Google said adding more high-quality content was better than removing content in general. Remove or noindex only pages that have no value to visitors.
How long does Panda recovery take?
Before 2016, recovery usually waited for the next refresh. Today, most sites that do substantial work see movement over several months, often around a core update. Nobody can promise a date.
Sources
- Search Engine Land: Google forecloses on content farms with Farmer update
- Search Engine Land: Google Panda update guide
- Search Engine Land: Panda is now part of core ranking signals
- Search Engine Land: Panda 4.2 is here
- Search Engine Journal: Panda update history
- Wikipedia: Google Panda
If your traffic dropped across the whole site after a core update and you suspect a Panda-style quality problem, request a free audit. I will tell you which sections are weighing the site down, and my SEO audit services turn that into a page-by-page plan.
More Google algorithm updates
- Previous update: Attribution (Scraper) Update (28 Jan 2011)
- Next update: Freshness Update (3 Nov 2011)
Other content quality updates:
- May Day Update (28 Apr 2010)
- Quality Update (Phantom) (29 Apr 2015)
- Fred Update (8 Mar 2017)