The Hummingbird Update was announced by Google in September 2013 and was considered to be the most in-depth change to Google’s algorithm since Caffeine in 2010. The update is linked to the development of Google’s Knowledge Graph and their ability to understand the intent of a search query rather than focussing on the exact keywords. Think of it as a mix between the Knowledge Graph, latent semantic indexing and intent matching. In a nutshell the Hummingbird update overhauled Google’s algorithm, making it far more effective at understanding the intent behind search terms and providing results that match this intent, instead of keywords. For example, if you were to search “Why can’t I get Windows 10 on my laptop?”, Google will understand that you have a problem downloading or reserving Windows 10 without you having to use those specific keywords. As such, the results are articles or information which answer the question, i.e how to download or get the upgrade option on your laptop.
How do they do this? By being very, very clever and having access to an unfathomable amount of data from searchers. Or by sorcery. Either way, it can give you a headache very quickly.
Why the big change?
Google has to continue evolving as the internet evolves. Not only are new websites springing up by the second across the web, but the amount of content on the internet is increasing at a frightening rate. As such, Google has to make sense of this content and continue to be the best at giving the most relevant information to a search query.
Simple right? Not quite. There are a couple of very compelling factors that have somewhat forced this update:
Firstly, Google has been around for long enough now that it has an assumed authority and complexity amongst the majority of its users. In fact you can probably notice from your own usage of Google that you are starting to treat the search engine like a human being. Gone are the days where people do not understand how search engines work – an amusing clip released by Google when they rebranded in September 2015 and wanted to show how far they had gone said: “work the Google on the internet machine”. Everybody understands Google now and more importantly, searchers now trust Google to understand them if their search queries are more complex than 2 or 3 keywords. What does this all mean? It means that users with a question are now typing very long searches into Google expecting a highly specific result to be returned. This expectation may have been one of the driving forces behind Hummingbird.
Secondly, the adoption of smartphones has been mindblowing. They now rule our lives. But with smartphones comes voice activated search, something that throws the proverbial spanner in the engine. Couple voice activated search with this assumed complexity and suddenly Google receives searches that do not mention particular keywords – yet the user expects the same result. Yes, whilst we are still typing our search queries may remain reasonably short for the time being, but voice activated search is a completely different ball game. It is a very strong argument that Google have seen this coming and made the necessary preparations.
What does this mean for SEOs?
At a basic level it should not mean much change. Thin content will suffer, but truly valuable content will excel. When Hummingbird first came out the advice was to answer questions. This theory still stands because if you are answering questions then you are providing real value to the user and Google claims to be able to understand this value. Again one of the tactics that was popular when Hummingbird was released was to include more questions as headers of pages or paragraphs. This certainly helps you provide more answers within the paragraph and can help structure the webpage more effectively but can be overused. Probably not by accident, a Search Engine Land that comments on the update had 17 paragraphs and answers!
In general though, it means that you want to focus on a very clear information architecture for your site and to focus your content on providing real useful information for the user. If it doesn’t add value, it won’t be valuable for Google to include in the SERPs.