 |
|
(Amit Singhal, senior vice president of search at Google, introduces the new ‘Hummingbird’ search algorithm at the garage where the company was founded on Google’s 15th anniversary in Menlo Park, California September 26, 2013.
Credit: Reuters/Stephen Lam)
|
Earlier, Google used to pick words from the search query and find out web pages that used those words based on over 200 different signals. But, the Hummingbird search algorithm has provided a fresh new approach to how Google predicts the search query. Google will now predict the context of the search query and try to return more precise results often based on searchers intent, relationship between words, use of pronouns, searchers location, underlying meaning of the words of the search query etc.
The figure below shows the timeline of Google search. Features like did you mean, synonym search, autocomplete, universal search, Google instant, Knowledge Graph etc have all contributed for a better and a more precise Hummingbird algorithm.
 |
| The Timeline for Google Search (Image source:- googleblog) |
Google had launched Knowledge Graph in May 16th, 2012 and expanded it in Oct 19th, 2012. The Knowledge Graph is a basic model working on the concept of semantic search. It helps Google to find the relationships between search entities and categorizes people, places and things for the betterment of the search. For a more deeper understanding of Knowledge Graph, click here.
The Hummingbird algorithm will explore the power of relationship between entities for finding out the context of the query. Apart from Knowledge Graph, it will use features and concepts like voice search, natural language processing, Google now, comparisons and filters in Knowledge Graph, comparison tool, Google reminders etc.
All the above features and updates will work behind the scenes for Hummingbird update.
Some Questions Related to the Hummingbird Algorithm Update
Why did Google updated to Hummingbird algorithm instead of its old algorithm?
Google needs to handle long and complex search queries that demanded precise and accurate results. Due to the amazing capability of Google to return accurate results, users have started to rely more and more on the search engine with the result that longer and complex queries were entered. It is for this reason that Google had replaced the old algorithm and updated to Hummingbird for accurately predicting the users intent behind the query and to present them the closely matching results.
When did this update happened?
The exact date is not announced but it may have happened around the end of August or the first week of September.
Who announced this update?
This update was announced by Amit Singhal, senior VP Google, on the company’s 15th birthday. Amit mentioned about this update at an event hosted in the garage that Larry Page and Sergey Brin (Google founders) had rented to start their own search company which is today known as “Google”.
What percentage of search queries are affected by this update?
It affects nearly 90% of the worldwide searches. It would normally get reflected on long tail queries.
What about Penguin and Panda?
Penguin and Panda are two of the most successful web spam updates and they will continue to work alongside Hummingbird.
How is Hummingbird different from the old algorithm?
The old algorithm worked on word by word basis and did not interpreted the combined context of the group of words used in the search query. However, the Hummingbird update finds out the real meaning behind the search query and returns more precise results.
Amit Singhal Sharing His Views on Google Search After Unveiling Hummingbird Algorithm
Here is a short video transcription to what he said:-
” When it comes to technology behind search, the fact is that computers still don’t understand language like you and I do. We at Google had made tremendous advances in understanding language…Knowledge Graph has been foundation into that, the new algorithm that we launched today called Hummingbird has been a great leap forward, but still we are far from human level understanding of language. Still having more computers helps, but its really not the bottleneck that we are facing these days, its all about understanding better what users expect from Google, what they are asking and then fulfilling it via algorithms…”
The Backbone of Hummingbird Algorithm
The backbone of Hummingbird algorithm is built upon few main technologies as described below:-
Conversational Search
Google has moved a step forward from being a search engine to a QA engine. A machine that can interact with you in the same manner as you would indulge in conversation with your family or friends. You can ask Google “How old is Justin Bieber?” and Google will return the current age, instead of just returning references to web pages that may contain information related to Justin Biebers age. This is made possible with the help of conversational search (Google also returns Knowledge graph information related to Justin Bieber which means Google is using the Knowledge graph to find out the age related information). You can now directly interact with Google and find direct answers to most of your queries. But still, this works on particularly long tail queries, something that can hint Google that the user is demanding an answer instead of the conventional search results.
Semantic Search
Semantics is “Science of meaning in languages” and it forms a base of the Google’s relevancy factor of the main search algorithm. As SEO’s we all know that the main components of search algorithm is divided into 2 parts namely the relevancy factor and the popularity factor. Relevancy is measured taking into account Semantics, NLP, LSI, TF-IDF, QDF etc. and popularity is measured on the basis of backlinks.
Semantics forms the base working behind the main relevancy search algorithm. The Knowledge Graph powers the semantic search model for Google.
Natural Language Processing (NLP)
This is one of the important areas which takes into account and makes use of machine translation, sentiment analysis and question answering. It works on the principle that search engines should process the natural language queries in the same way as humans do. This involves detecting parts of speech within the search query and relationships between them and other wide array of search technologies for better parsing of the main search query.
Knowledge Graph
A huge database of objects and their relationships have been covered in what is known as the “Knowledge Graph”. This database stores data related to people, places and their relationships. It helps Google to parse the query which includes the name of people and places and helps it to return more precise results. The Knowledge Graph is increasing day by day and by making use of trusted resources like Wikipedia, this huge database of facts is helping Google power today’s Hummingbird algorithm.
Google Now
Think of it as your own personal assistant. It helps you to stay connected with your interests and also lets you set reminders or restaurant reservations. This is a great way to judge the fact that Google is going to interact with our lives at a very personal level. Hummingbird is an extension towards this approach, i.e. to specifically cater to your demands instead of showing irrelevant search results that bears no relation with your location or interests.
Also See:-
