Not content with taking away the little keyword data we had left this week, Google has again surprised the online marketing industry with a brand new algorithm.
There isn't an awful lot of detail out there right now, though Search Engine Land has some details, including the fact that this this new algorithm has been running for the past month.
I've been asking some of the finest minds in SEO for their thoughts on Hummingbird...
What is Hummingbird? How big a change is it?
Andrew Girdwood, Media Innovations Director at LBi
Hummingbird isn’t an algorithm update. It’s a new algorithm and that’s incredible. It’s incredible because it’s a huge change and no one noticed.
Sure, there was some odd behaviour and some ripples over the last month but they’re all too common in search and I don’t know a single expert who looked at them and came to the conclusion Google had dropped its old algorithm entirely in favour of a new.
In fairness, the Hummingbird algorithm uses many of the old rules, weights, filters and systems of the old.
Joe Friedlein, Browser Media:
Hummingbird is an evolution of Google caffeine, which rolled out in 2010 and was focused primarily at finding fresher results. Like a hummingbird, the new algorithm is intended to be 'precise and fast'.
It is too early to really tell whether it is going to be a massive game changer but it is claimed that it will affect around 90% of searches, so it is not something to be ignored!
Alex Moss, Co-Founder at 3 Door Digital:
This is going to be a bigger change than people may realise as, especially when it comes to mobile SEO, long tail keywords are becoming more commonplace and are also generally of higher quality than that of someone entering a short tail phrase.
Dan Thornton, Founder at TheWayoftheWeb:
Hummingbird is a major overhaul of the algorithm used for Google's search results to improve results for longer 'conversational' search phrases, more widely known as Semantic Search.
The idea is that Google can now understand more of your intent, rather than the keywords being used.
Julia Logan, SEO consultant (better known as Irish Wonder):
From an outside point of view, I think it will take us some time to fully estimate the entire impact this change has on what we see as Google users in the SERPs.
This change, however, seems to be in line with the other latest developments in Google, so we probably should not expect a radical 180 degrees turn.
Teddie Cowell, Director of SEO at Mediacom:
The fact that Google chose to announce it from the Menlo Park garage where the company started out shows it is a very significant change, and symbolically the beginning of a new start.
Jimmy McCann, Head of SEO at Search Laboratory:
Google is trying to give more semantic search results so that the intent behind any given search query is met with results that match that intent and ultimately provide users with a better experience.
For example, if I were to search for ‘’how do I install blue widgets in my house’’, Google would return a set of results that maybe included the words ‘’install’’, ‘’blue widgets’’ & ‘’house’’.
As a text reader, Google could return any number of pages that contain these phrases but do not necessarily relate to the actual query ‘’how do I install blue widgets in my house’’.
With Hummingbird, Google will be placing greater emphasis on the other words in the search query such as ‘’how do I’’ and ‘’my house’’.
Rather than just matching individual words and phrases to the pages in Google’s index, it is looking at all the words in the query and the relationship between each to understand the intent behind the query more. This in turn should lead to users finding the specific results they need for their query.
Malcom Slade, SEO Project Manager at Epiphany Search:
Hummingbird is basically a change in how Google interprets the intent of a user’s query to ensure the returned results are appropriate.
Assuming that pre Hummingbird (pre August 2013) Google was using pattern matching to understand intent, post Hummingbird it is using much higher level NLP (Natural Language Programming) concepts to ensure that the full query is answered by the returned sites.
Kevin Gibbons, UK MD at BlueGlass Interactive:
The biggest impact appears to be towards 'conversational search'. At such an early stage of an algorithm update you normally can't go too far wrong with reading what Bill Slawski has to say.
Why does Google feel the need to do this?
Search is changing. Google’s called out Voice search in reference to Hummingbird and as it happens I asked a question of Google about 'voice PPC' at ad:tech this month.
Voice search is already possible and impressive at Google. You don’t just search without typing; you search without keywords as the computer remembers the context of your conversation.
[How old is David Tennant] could be your first search and you can follow that with [Who is he married to] and you’ll get the answer for David Tennant’s wife.
I wanted to know if and when Google would make 'voice search PPC' and its adjustments available. You never get a straight answer in a situation like that but I got the impression that the search engine was not yet convinced the quality of the experience was good enough to monetise.
Hummingbird, in part, is designed to get the voice search experience up to the point where Google is happy with it. That means more ad options.
To be honest, there is nothing new in what Google is trying to do. It has always had a desire to return the 'best' results for their users.
'Best' can mean different things to different people, but relevancy is always going to be vital and more recent results tend to be more relevant, so Google will always be interested in fresh content that is successfully attracting social signals.
Hummingbird will be in its infancy but once ironed out completely I have confidence it will help both searchers and site owners alike, especially smaller businesses with niche products or services.
Semantic Search has been growing for several years, and the intended goal of producing highly relevant search results ties in with Google's longterm strategy.
But it also gets more traction when you consider the massive growth in search on mobile devices, and the arrival of Google Glass.
It's also a potential weapon against competitor social networks, which in some ways provide a semantic rival, as the answers of a trusted friend can infer meaning which a search engine doesn't.
I particularly like the emphasis on intelligence, and the 'conversational' reference Google uses to describe Hummingbird and how it would work more efficiently in mobile environments.
Hummingbird is the natural convergence of two roads search has been on, firstly providing more intelligent, accurate and useful information responses, secondly making how we request and access that information as natural to us as possible.
With an extremely open mind on the future, it's not unfeasible that interacting with Google will at some point feel so natural and intuitive that it will effectively pass the Turing test.
To be more natural and provide more accurate results, and to keep up with the naturally evolution of how we use search and the internet. As queries get more complicated, Google needs to provide more specific answers.
The key emphasis is on ‘answers’ rather than ‘results’, with Hummingbird paying particular focus to conversational search. Personally it appears like this is a change to reflect and accommodate the ever increasing use of search on mobile devices.
Query types and user intent can be quite different on mobile devices because they often take place in a different setting. Amit also launched a post yesterday that highlighted the development of Google and the importance of conversational search on mobile devices.
This is a massive step towards making Google results better for all users on all devices. Mobile technologies like 4G, Voice Search and Google Glass have changed the way we interact with Google and the results we desire need to be much more targeted and accurate.
Historically Google has always had issues with really long tail informational queries. "Where can I get a pair of size 11 Nike Airforce 1 trainers in the next 12 hours?” would only return a good result if someone had asked that very question on Yahoo Answers or similar and would have very little value to the current user.
Hummingbird implies that Google is now drawing a lot more meaning from the query prior to attempting to return results.
What effect do you think it will have on some site's rankings, and the role of the SEO?
Interestingly, this comes at the same time Google’s reduced keyword referral data. Voice search has less emphasis on keywords.
I argue that the most important rule of SEO is that people matter and I believe that modern SEO is very much now about interacting with people. Hummingbird seems to support this approach; moving to humanise search further.
Food for thought: is the predicted rise of voice search and conversations with Google one of the inspirations around the removal of keyword data from natural search?
Ranking will definitely fluctuate over the coming months, but what we could also see is the increased use of structured data by Google in order to serve answers directly within the SERPs.
SEOs will have to ensure that, although you can provide better information within a SERP, that it is enough but not too much that the searcher does not convert into a visitor.
As the search journey will be more a conversation with Google taking place within its ecosystem, we will have much less of an idea about which keyphrases and what the overall search chain was that lead to a website.
The role of individual keyphrases in their own right will continue to diminish, because the entire query journey to an item of content becomes more important.
This gives real context of the recent removal of all the organic keyphrase referral data, because seeing the final referrals coming through to websites for things like “what’s the address?” and “where can I buy one?” would be pretty pointless.
Less locally-targeted sites may suffer, but hopefully Google will eventually learn to feel the line between truly local queries and those just looking for information regardless of location.
Considering that there is pretty much a different version of the algorithm for different verticals, it all depends on the actual implementation of the Hummingbird in different verticals, we have seen lately that Google can be pretty flexible.
SEOs will probably become more specialised in certain verticals to be better at them but we have seen this happening already so this is not earth shattering.
The key point for us as SEOs is the Hummingbird is still based on links, which still play one of the most important parts. They serve both navigational and discovery purposes, and had been doing so long before Google has come to exist.
The jury's still out! For competitive, transactional queries, very little. One thing we've been doing a lot of recently is building informational content hubs with clients, so I'm very interested in seeing if these will be rewarded from hummingbird.
Right now I think it's a minimal change though, it was rolled out a few weeks ago and no-one really noticed until they announced it - so I think the question is more about if this will be dialled up as a more important factor in the future.
Have you noticed any changes over the past month as a result of this?
There are a few ranking changes that now make more sense to me. It’s the benefit of hindsight. By and large this change has been impressively seamless.
One of the reasons why so few SEOs noticed the big change is that ranking reports and traffic tend to watch the dominate keyword phrases. Hummingbird is very much about making the long tail searches a better experience and few brands or agencies track those.
So far, I can't say that I have seen any concrete evidence of big changes as a direct result of the Hummingbird update but there is an ongoing trend of content sites performing well and sites that have been thin on content but relied on bulk link building have suffered.
It's always important to avoid mistaking correlation for causation, particularly when the change was made without any announcement, and all other potential reasons need to be ruled out.
We haven't noticed any particularly large fluctuations with any client websites, but we have seen a couple of internal test and project sites which suggest they were affected by the algorithm change.
I have noticed several things. First: this is not a panacea against search spam. In fact, some of my recent experiments have shown that it has become easier to rank spammy sites, at least in some niches.
Second: I have no direct proof but it looks very much like Google has started putting the database collected via its disavow tool to use in order to detect sites getting links from known sources of 'unnatural' links before the sites in question even get any benefit from those links.
This is quite a scary development as
- We have no way of telling which sites have been disavowed.
- It is not known in what way exactly Google is using this database.
I have posted about this issue in detail earlier this month. As for search results relevancy, including some of the 'conversational' queries, often they still raise a brow, like the one below but, as with any algorithm being rolled out these days, this is of course work in progress, and one thing for sure is that we are living in very interesting times.
Yes, we have been seeing very positive results over the last month.
Sites that put the effort in to bringing unique content into all of the travelled areas of their site and are genuinely aiming to be the thought leaders in their market are doing well while sites still following the online brochure backed by loads of link acquisition are losing ground.
It is early days but it would seem that quality and user experience have gained importance while links may have been dialled down slightly.