This post summarises some of the major Google algorithms that we know of. You could write thousands of words on each one of these algorithms. However, the aim of this post is instead to give you a good idea to the basics of them. In the future, I will be looking into writing more about some lesser known algorithms that do not get spoken about quite as much.
- Hummingbird is a rewrite of the Google search algorithm.
- Hummingbird improves “Conversational Search”, “Contextual Search”, as well as “Semantic Search”.
- The platform rewrite was released in the middle of 2013.
What Is Hummingbird?
Hummingbird is the Google algorithm. Hummingbird is the main algorithm, while the other algorithms, such as PageRank, Penguin, Panda and Rankbrain are add-ons that improve Hummingbird. Previously Google’s main algorithm for search did not have a name, in the middle of 2013 Google did substantial work and overhauled a large portion of the engine. At the same time naming it Hummingbird. It was called Hummingbird as it is “precise and fast”.
What does Hummingbird do?
The biggest feature we know of from the search engine rewrite is that it now supports “Conversational Search”, this is mainly for people searching using voice search. Google previously was doing this with it’s Knowledge Graph answers; this upgrade applied this to all search results.
More reliable third-party information has come out since its release from the likes of Marcus Tober and the Searchmetrics team; they revealed information at SMX West in 2014 from a study they undertook. Its findings were that the context of a search as well as previous searches both influence search results post-Hummingbird, they also revealed that Hummingbird attempts to understand search entities (places, people, and things) and how they are all semantically related to each other. Since this rewrite and Google’s improvements on the way the semantically understand related terms, the Searchmetrics team is reporting that Google is now delivering similar search results for queries that are similar to each other; as if they were all the same query.
How can I optimise for Hummingbird?
With this change, your SEO strategy should begin to optimise more for a user’s intent behind a query. Think about what the user is going to expect when they visit your page and make sure you are producing high quality, unique content that fulfils what they are looking for. Keyword research and including primary keywords in things such as titles, H1s, etc.
Keyword research and including primary keywords in things such as titles, H1s, etc. is still just as important, but it is less important than ensuring you are fulfilling the searchers need. On top of targeting primary keywords, try and think about related keywords in the content of the page that would usually naturally co-occur.
What is RankBrain?
Rankbrain is an artificial intelligence algorithm (machine learning in this case) that is the newest addition to the Google, algorithm family. It was announced in this video interview with Bloomberg which I suggest you watch.
As with all other algorithms, it is part of the main Hummingbird algorithm and is just one input of many algorithms that decide where a page should rank. Google has announced that Rankbrain, behind content and links, is the third most important signal for ranking.
What does Rankbrain do?
RankBrain is a query processing algorithm that returns better search results for more complex phrases that Google previously could not understand quite as well. Rankbrain is especially useful for long tail queries that Google has never seen before (currently sitting at around 15% of all queries). One way it provides better results is by taking long and more complicated queries and matching them with a shorter query, that it believes in answering the users search.
It also better processes negatives. For example for the query “Can I fly a plane without training”, previously Google would not handle the negative word “without”. With the release of RankBrain Google will now understand what you mean. Gary Illyses had this to say:
“Our old query parsers actually ignored the ‘without’ part. RankBrain did an amazing job of catching that and instructing our retrieval systems to get the right results.”
How can I optimise for Rankbrain?
I do not recommend you try to optimise specifically for RankBrain due to it targeting long-tail keywords that are difficult to understand. If you are following best practice and writing naturally, considering search entities, keyword co-occurrence and best practice for semantic SEO, you will already be well on the way to being well optimised for RankBrain.
- https://moz.com/blog/rankbrain-unleashed – One of many great posts by Gianluca Fiorelli, if you are interested in RankBrain, read this.
- http://www.bloomberg.com/news/articles/2015-10-26/google-turning-its-lucrative-web-search-over-to-ai-machines – Initial announcement and video.
- Panda is a Google algorithm that aims to show high-quality sites in search results, while demoting low-quality ones.
- Panda launched on February 24, 2011, since then is has had many iterations improving the algorithm.
What is the Panda algorithm?
Panda is a well know Google algorithm that had a large impact on search results (11.8% of all search queries) when it was initially announced and released on February 24th, 2012. Since then it has had many iterations that have continued to demote and promote websites. Currently, we are on version 4.2 of the Panda algorithm. See Google’s original announcement for this algorithm here.
What does Panda do?
The Panda algorithm aims to demote low-quality sites that do not add value to users; this can be because the content on the website is thin, or it is duplicated from other sources. See this excerpt from Google’s original announcement:
“This update is designed to reduce rankings for low-quality sites—sites which are low-value add for users, copy content from other websites or sites that are just not very useful. At the same time, it will provide better rankings for high-quality sites—sites with original content and information such as research, in-depth reports, thoughtful analysis and so on.”
How does Panda determine quality?
Google has not given much information on how the Panda algorithm determines quality other than some advice from Google employee Amit Singhal. You can see a full list of things to take into consideration here.
The list from Amit Singhal can be clarified into the following points:
- Thin content
- This is simply pages with very little content or content that does not add much value for the user.
- Unoriginal content
- This is when content is copied from elsewhere and put on another website, this is usually due to it being scraped, duplicate product descriptions or just where someone has taken some content and slightly rewritten it (Google is good at detecting this).
- Nondifferentiated content
- Nondifferentiated content is content that has already been covered in-depth across the web. Unless you have something to add, or you are going to present the content in such a way that is better than what is currently available, from Google’s perspective, it does not need any more articles. You should always be aiming to make content that is 10x better than what is currently available.
- Poor-quality content
- Low-quality content is content that is factually incorrect or poorly written. An indicator of poorly written content could be content with lots of grammar or spelling mistakes.
- Curated content
- Lots of sites that tend to curate content tend to get hit by the Panda algorithm. This happens when content is curated by no extra value is added via commentary. If your page just includes an extensive list of links, there is a good chance that it will not perform well.
- Thin slicing
- Creating pages on a site that cover very similar topics so you compete for variations of a search phrase.
How do I know if I have been penalised by Panda?
The Panda algorithm used to update periodically; recently Google has said that the Panda algorithm is continuously updating. The algorithm constantly updating means Panda is frequently re-scoring sites and pages. After it has scored the pages on your site, it then rolls changes out over a matter of months. In the future, it will be harder to tell if you have in fact been impacted by this algorithm.
If you think you have been penalised historically by the algorithm check for significant drops in traffic and compare it to known updates to the algorithm.
How can I avoid being penalised or recover from Panda?
Write high-quality, unique content! Follow the list Google has suggested to keep in mind when creating your site and writing its content. If you are penalised by the Panda algorithm audit the content of your website and ensure you are following the list Google has provided. Also, make sure you have not failed any of the points mentioned above in the blog post. After finding any low-quality content, invest time into either improving it or removing content that is not valuable.
- Penguin is a spam algorithm created by Google that penalises sites with a backlink profile that contains a high percentage of links that were created to improve rankings unnaturally.
- Penguin shook up the search landscape on the 24th of April 2012. Just like the Panda algorithm, there have been many updates since it first launched.
What is Penguin?
The Penguin algorithm is Google’s response to webmasters trying to improve rankings in search results in a way that it deemed unnatural and spammy. It targets various types of “bad links” that Google can detect algorithmically without a manual review of the site. The Penguin algorithm updates around every year. When that happens sites either have a drop in traffic, an increase in traffic from a recovery or they remain as they were. Be aware, if you are penalised by the Penguin algorithm, you may be waiting a while before it updates again and you recover.
What type of links does Penguin target?
Some examples of link types Penguin targets and penalises sites for are:
- Directory sites
- The very first release of Penguin targeted low-quality directory sites that did not have the best reputation. While there are many great directories, you will want to be listed in, before submitting your site to them consider the quality.
- Article directories
- The first release of Penguin also targeted this type of link. Avoid submitting articles to article submission sites with links back to your site.
- Repeated used of exact-match anchor text
- One way to influence rankings is by having a lot of backlinks with the keyword you are trying to rank for. Doing this nowadays is a sure fire way of getting penalised by Penguin. You should try and avoid using undescriptive anchor text such as “click here”, but at the same time be wary of having an excessive amount of the same keyword-rich anchor text.
To avoid getting penalised by the Penguin algorithm, I recommend checking your backlink profile with tools such as Ahrefs and Majestic. Ensure you do not have an excessive amount of links that fit the above categorisation. If you do have a lot of spammy links, follow the suggestions below to ensure you do not get penalised.
How do I recover from Penguin?
To recover from the penalty or avoid being hit by the penalty in the future I recommend periodically checking your backlink profile. I highly recommend the previously mentioned tools, Ahrefs and Majestic. If you find any links that do seem to look spammy or fit the above categorisation, you may want to look into disavowing them.
There is a lot to learn about all the different Google algorithms currently out in the wild, and even more algorithms that we can only speculate about. Re-read the above blog post and try and let as much as possible sink in! Learning search algorithms and how they work is one of the first steps to becoming a great SEO.