A few months ago Google quietly rolled out the biggest change to their search algorithm since the
early 2000s. They followed it up with the recent announcement that search data is being moved
entirely to secure search – meaning all the keyword data around which many search teams would
base their marketing efforts is no longer available (at least, not as easily – we have found some
workarounds to gain some arguably inferior insights_.

When the update, duly named Hummingbird, was officially announced on Google’s 15th birthday on
September 27th , it raised an important question: Why is ranking data no longer available? Because
Google has always pushed to deliver the most relevant and highest quality content to its users. But
that still doesn’t answer the question, or does it?

In the past, when looking into how to best optimize a webpage, amongst the first considerations were
“which keywords should we be targeting?” You’d have a look at the volume of searches for a
particular keyword, and ensure that your site is optimized for the most relevant phrases. By taking
this data away, Google is further emphasizing the fact that they don’t want you expending energy
optimizing for keywords. You should be concerned only with optimizing your website to have high
quality, relevant content. The algorithm follows suit, delivering results that reflect a user’s intent,
rather than results that match specific search queries by keyword.

A Simple Explanation of Semantic Search

But how can Google gauge a user’s intent? It’s called semantic search. Let’s say you were searching
for the best Thai restaurant in Toronto. In your mind, you know you are looking for the best Thai
place, but the first search on your mind may be “where is the most delicious Thai place in Toronto”.
Previously, Google may have returned results that have been optimized for that phrase, or a portion
of it. But let’s look at the results I get for my search:

First of all – Google understands that I am searching for a restaurant where I can devour some Pad
Thai and fresh rolls – even though the word restaurant was not in my search. You can see it has
presented me a list of reviews and geo-targeted restaurant options, because my intent is to find a
place to go eat, and these are the tools I need to make my decision.

Long-tailed searches like the one I made above are becoming more and more popular with mobile
devices integrating search via voice options. Searching is becoming more conversational, and Google
continues to expand their knowledge graph in order to be able to understand complex searches and
turn them into relevant results. Semantic search has been around for a while now – the new
Hummingbird algorithm is taking it to a new level.

If you are looking to add a new member to your family, you might start with a search “Rottweiler vs.
Lab”. Take a look at the result you will get now:

Google’s Inside Search blog has more details on some of the new features – a few of which, at this
writing are still not triggering on Google.ca.

Hummingbird’s Effect

Hummingbird will continue to provide smarter, more relevant search results for our queries, but the
big question right now is – how will this affect SEO strategy going forward?

With the advancements of semantic search, there is going to be a stronger focus on optimizing pages
of your website to certain topics, rather than focusing on keywords. The key has always been to
provide the best experience for the users on your website; putting the content they are hoping for in
the places that they expect it to be. It used to be essential to use the appropriate keywords on a
given page, but as time goes on, and recognition gets stronger, pages on relevant topics will become
the important focus.

Each page and the content of said page will be the central unit of analysis going forward. Ranking
metrics will become more comprehensive to account for the rising number of local, , mobile and
conversational searches. As long as websites continue to provide relevant and useful content, and
their users continue to digest that content, rankings on Google shouldn’t be affected dramatically, if
at all.

Paid Search

Considering the fact that Hummingbird deals with conversational search, we anticipate that
campaign structures, use of match types and even ad copy may eventually need to adapt to more
complex search queries in order to be able to capture that traffic. In fact, we already see many of
those types of conversational/question-type queries in our search query reports. Using these types of
results for keyword expansion however, may be difficult as the current Google AdWords’ limit for a
keyword’s length is 80 characters and 10 words. Based on Media Experts’ SEM data, keywords over
40 characters in length currently receive very little search volume. Whether Google AdWords is
willing to change those AdWords limits remains to be seen.