What is Google Hummingbird?

Google Hummingbird is a new search algorithm announced by Google as part of their 15th birthday celebrations at the end of September 2013.  The salient part of the previous sentence is a new algorithm; this is a totally new departure and potentially affects over 90% of searches. Interestingly, the algorithm was changed about a month before the announcement and received no noticeable negative feedback from users – a sure sign of success!

Why a new algorithm?

Users’ searches have been getting longer and more specific over time. As search queries became longer and more conversational in style Google had to understand the meaning of those queries and to adapt search results accordingly. As always it was about giving the user the information that they wanted.

Give me an example

Google wants to understand what you mean when you type in a search term. First off an example: Think of some interchangeable search queries – say ‘the best ice-cream shop in my area’. You could also mean ‘the best ice cream parlour in my area’ or ‘where to get the tastiest ice-cream in my area.’  These are all long tail conversational style queries, and Google needs to know the best pages to return as a result. Hummingbird is the new algorithm that Google hopes will provide the solution. First off they need to know where you are, which they do if you have set your location.

How does Google Hummingbird Work?

To put a very complex process in simple terms, Hummingbird uses Google patents to re-write queries using synonyms, and by paying attention to all the words in that query and how they relate to one another Google Hummingbird hopes to provide the most relevant search results based on that particular query.

In other words, ice-cream and ice cream mean the same, tastiest and best probably mean the same, and shop, parlour and where to get all mean an actual place.

Here is an excerpt from one of Google’s patents:

A search query for a search engine may be improved by incorporating alternate terms into the search query that are semantically similar to terms of the search query, taking into account information derived from the search query. An initial set of alternate terms that may be semantically similar to the original terms in the search query is generated.

Hummingbird ties together the query re-writing patents that Google has filed and also builds on the latent semantic indexing that is the discovery process for finding related terms and phrases. LSI plays a key role in what is returned on search engine results pages (SERPs), learn about LSI in my earlier blog – Why content is king, the science behind the statement.

What does Hummingbird mean for me?

It seems clear that Google is working towards returning the most relevant content to any given search query. In practical terms this means that Google will continue to reward good quality content. Having a content strategy based around long-tail keywords, and paying attention to the 5 W’s (what, where, when, who and why) will pay dividends in the long term.

Bill Slawksi, Seobythesea.com.  Rand Fishkin, The Moz Blog.

Google Patent Examples:
Search queries improved based on query semantic information
Synonym identification based on co-occurring terms