Posted by Jonah A. Berger, SEO Operations Director, EMEA
- Google has made a significant algorithm update called “Hummingbird” that is said to affect upward of 90 percent of search results
- Unlike the Panda and Penguin updates that preceded it, Hummingbird focuses less on the ranking algorithm and more on conversational search and how queries are understood
- So far the update hasn’t made much of an SEO impact in terms of search visibility and ranking, as Hummingbird has been live for more than one month, according to Google
- SEO efforts shouldn’t be altered too much right now. If anything, future keywords and content should focus more on longer queries and providing searchers with answers.
Hummingbird is in Flight
Precise, fast and conversational—apparently this is what Google has in mind with its latest algorithm update known as Hummingbird. Rolled out more than one month ago but unannounced (and somewhat unnoticed by the SEO community) till 26 September 2013—the eve of the search engine’s 15th birthday—Google has made what it has said is its biggest overhaul to the ranking algorithm since Caffeine in 2010. That algorithm update focused on a new search indexing system that allowed for the fresher, more relevant and real-time results we see today.
So what exactly is Hummingbird?
Said to affect 90 percent of Google search results, Hummingbird is part of what’s fueling the search engine as it attempts to move away from displaying results based primarily on keywords and more toward providing answers. Think of it as a more intelligent understanding of search requests and the semantics behind a search query.
Danny Sullivan, editor-in-chief of Search Engine Land, explains how Google’s handling of search queries could change with Hummingbird in the fold.
“Google said that Hummingbird is paying more attention to each word in a query, ensuring that the whole query—the whole sentence or conversation or meaning—is taken into account, rather than particular words,” Sullivan said. “The goal is that pages matching the meaning do better, rather than pages matching just a few words.”
In a pre-Hummingbird world then, one can imagine a searcher typing “Movies that Leonardo DiCaprio starred in besides Titanic” into Google and getting results that are a combination of Leonardo DiCaprio movies and Titanic. What aren’t likely to be displayed are results that answer what the searcher actually seeks. Yes, this may be a very specific example, but one of the drivers behind Hummingbird is that Google will answer a query in a human-like manner; it will understand that the searcher wants to see results that in this case have nothing to do with Titanic.
Search isn’t just for Typing Anymore
With Hummingbird, Google is also focusing more on the mobile search audience that’s using voice search to traverse the web. Think Siri, the intelligent personal assistant released by Apple in 2011, and how a hybrid version of her/him (depending on your preference) can be utilized by Google to be part search engine, part personal assistant. Google Voice Search has been in existence for years, first as a Google Labs tool in 2008 and most recently in 2012 in the form of an iOS app to rival Siri. With Voice Search, all a user needs to do is tap the mic icon on the Google search bar and start talking, similar to how Siri is utilized today. But where Google hopes to take Voice Search that Siri hasn’t gone is to a place where results are quicker, spoken more frequently (e.g. sometimes Siri will say, “Here are your results…” and not speak the actual answer), and in the end what’s returned is more factual and targeted because it’s coming from a search engine and not just a personal assistant.
A Move toward Conversational Search
Google has been working on making its interactions with searchers more “human” in recent years, and in 2012 it unveiled its knowledge base known as Knowledge Graph. Overflowing with billions of facts that are extracted from sources like The World Factbook and Wikipedia, Knowledge Graph provides searchers with detailed information that’s displayed within the SERP itself, as highlighted in blue in the following screen shot:
This allows for Google to not only answer queries in a more efficient manner, but brilliantly it also keeps the searcher from having to click through to another site to find what he is looking for.
“It’s [Knowledge Graph] another step away from raw keywords (without knowing what those words really mean) toward understanding things in the real world and how they relate to each other,” said Matt Cutts, head of Google’s webspam team, in a recent interview. “The Knowledge Graph improves our ability to understand the intent of a query so we can give better answers and search results.”
With the Knowledge Graph as part of its arsenal to instantly provide a wealth of relevant information without much more than the touch of a finger, what makes Hummingbird truly unique is that it allows Google to quickly dissect full questions and longer search queries instead of just displaying results that are based on words within them, as the latter is a practice that has always been commonplace to search. As a result, in a perfect Hummingbird world there will be a dialogue between human and machine. Amit Singhal, senior vice president and software engineer at Google, spoke about conversational search at a Google conference in May 2013.
“A computer you can talk to? And it will answer everything you ask it?” Singhal said. “Little did I know, I would grow up to become the person responsible for building my dream for the entire world.”
Wait, Haven’t Search Engines been Providing Answers for Years?
Search engines have always analyzed keywords within a query and not necessarily the semantics behind the query itself. Remember the search engine “Ask Jeeves”? Founded by Ask.com way back in 1996 (two years before Google was born), Jeeves’ fictional butler character was unique at the time because in theory you could ask him “anything” and he’d provide the answers. The unfortunate truth was that Jeeves was just like any other search engine, as it used the keywords in the query to return what it felt was the best match in its index. This is very similar to what the Googles, Bings and Yahoo!s of the world have been doing ever since.
British physicist Stephen Wolfram unveiled his “computation engine” Wolfram|Alpha in 2009 and promised it would “present a new paradigm for using computers and the web.” Wolfram|Alpha’s primary differentiator was that instead of displaying results from a massive index of web pages and links that were crawled by search engines, Mr. Wolfram set out to create an engine that would return “curated data” results, or simply put, a knowledge base of structured data that’s collected from a variety of sites. Does this sound familiar? It should because this is the shift to semantic search that Google and other search engines have been making for years.
While on the topic of semantic search, how about a little fun? Let’s quickly test the search query “Where can I buy cake in London?” in both Wolfram|Alpha and Google.co.uk to see what’s displayed. First, Google:
As can be expected, Google’s results are localized and focus primarily on ordering cakes online from bakeries in the London area. Top natural search listings include The Cake Store and Hummingbird Bakery (apropos for this topic). Now, the same query in Wolfram|Alpha returns results that are anything but similar, if they can be considered results at all:
Whether we stumbled on a query that Wolfram|Alpha couldn’t handle or the engine was just having a bad day, it’s obvious that our query wasn’t answered in an effective manner, or anywhere close to what Google displayed. If I’m a searcher and have to make a choice, it’s obvious that Google is the clear winner here. But how good of a job has Google done in answering our question like a human would?
Let’s build out the query a bit and converse with Google some more. To do this we’ve queried “Aside from Hummingbird Bakery, where can I buy cake in London?” If you ask this question to a family member or friend, they’re likely to provide you with places that aren’t Hummingbird Bakery, knowing full well that you’re trying to avoid it. Is Google this smart just yet?
No, it’s not. Apparently Google’s Hummingbird algorithm hasn’t gone into action yet with this query, as it’s still picking apart the keywords we’ve entered to show Hummingbird Bakery results. As a searcher who has a very specific need, these results are pretty useless. It’s the search experience like this that Google is hoping to avoid with its Hummingbird.
OK, All this Sounds Cool. So how will Hummingbird Impact SEO, if at All?
Here at Performics, we haven’t experienced any positive or negative impact on our client engagements since Hummingbird took flight, and many other respected SEOs in the industry will tell you the same. The fact that Hummingbird had to be announced by Google itself and that the news broke roughly one month after it went live is proof that, for now, SEO without Hummingbird is the same as SEO with it.
Your first reaction, whether you’re an SEO or site owner, when a major algorithmic update happens is to take a deep breath and not panic. Look at your site analytics and reports to see if there have been any dramatic shifts in traffic, visibility and ranking. Without indication of a major change, it’s best to stay on course with your SEO efforts and to avoid the influence of rumor and hearsay. Through all the hubbub, fire and smoke that the SEO community can generate, the fact is that a lot of the theory and best practice that worked in SEO then still works now.
“Google’s saying there’s nothing new or different SEOs or publishers need to worry about,” said Search Engine Land’s Sullivan of Hummingbird recently. “Guidance remains the same, it says: have original, high-quality content. Signals that have been important in the past remain important; Hummingbird just allows Google to process them in new and hopefully better ways.”
Not every Google algorithm update affects the art and science of how SEO is done. If this were the case, as SEOs we’d be changing our best practices and how we best enhance our clients’ engagements nearly every day. Some algorithm upheavals, like Panda and Penguin, have rattled and stirred the SEO cage and forced practitioners and site owners to think twice about their content, and to dig beneath the many layers of a link portfolio to see exactly where the links are coming from and learn if they’re damaging. The Caffeine update of three years ago was considered a big deal to the industry because it “changed the game” by bringing real-time results like Twitter and Facebook content into the SERP mix, and it made site speed a household term that today can affect visibility and ranking of a site’s pages if they’re “heavy” and slow to load.
Peering into the Future
Whether you’re new to SEO or have grown some grey hairs with it through the years, you’ve probably figured out by now that what keeps SEO exciting and sometimes even slightly unsettling is that Google and other search engines are relentlessly pressing levers and knobs, pulling and pushing, altering, cutting, tweaking, copying and pasting their ranking algorithms all the time.
Time will tell with Hummingbird, as early indications are that it’s much more of an improvement to what Google is doing behind the SERPs and not in front of them. If anything, the future of SEO remains challenging, demanding and ever changing. Mobile devices and user experience are hot topics, social media and its influence on search is still being decoded and local search has made its way into everyday SEO conversation. If you’re spinning your wheels on the “new what’s next” in search, it’s these areas that should garner your attention.