Google Hummingbird - Reconstructing Google’s Algorithm


On its 15th Birthday, Google broke the news that its original search algorithm has been retired. No need to panic, Google still works. The new algorithm, named Hummingbird, is a combination of new and existing components with the overarching goal to better organize the world’s information. Hummingbird quickly delivers more relevant content based on user intent. Amit Singhal, Senior VP and Google Fellow, told Danny Sullivan that the last time the algorithm was rewritten to this extent was over a decade ago, in 2001.

Unlike past Google algorithm updates such as Florida, Panda or Penguin, Hummingbird wasn’t released in a single session. It was gradually released over time, much like Caffeine. According to Google’s press team, Hummingbird has actually been in action for a few months, leading industry experts to reasonably assume the Google search team has done extensive testing and analysis on the new algorithm. Because of the manner in which the algorithm was released, Hummingbird will not be a single inflection point in search traffic, but rather a gradual change. It also aligns with recent increases in “not provided” traffic (coincidence, I think not), making it even more difficult to pinpoint whether a site “won” or “lost” with this update.

The other major difference between Hummingbird and other updates is that Hummingbird is more of a replacement of the algorithm than a specific filter or adjustment. Hummingbird improves the way Google interprets queries and matches them to search results.

Query Intent

Search engines look at a variety of factors to determine the intent of user search queries. The most visual factor tends to be location. Searching for “restaurants” will return dramatically different results based on the location of the user. Users with IP addresses or network connections in Atlanta, Chicago, and Boston will see remarkably different search results pages.

Another factor that Google looks at is historical search behavior. Suppose the user loves Zappos. They often search for “zappos” or “zappos shoes” and regularly interact with the site. If that same user searched instead for “running shoes” or “dress shoes,” Google would likely display a relevant page from the Zappos website in the search results based on the user’s history of interaction with Zappos.com. Search engines may also want to deliver search results for women’s or men’s running shoes based on the user’s historical search behavior.

With the Hummingbird update, Google has continued to refine how it processes queries based on search context. Users will eventually be able to string together a series of queries to narrow down the search. For example, you could start with “Where is the Eiffel Tower?,” then proceed with “How tall is it?” and “How much are tours?” Each query depends on context from the previous question, and the result is conversational search.

Relevant Content

With Google’s Hummingbird algorithm, content is still king. However, relevant content is increasingly important. Hummingbird goes beyond simply examining the words in a search and matching them to Web pages that include those same keywords. When a user types in a query, Google is now processing the query to examine and identify the intent and meaning behind it.

Google representatives have mentioned a few times that approximately 20% of the search queries entered each day are brand new. At one point, it was common for a webmaster to create content optimized with long-tail keywords. Whether it be the 20% of queries users have not yet searched, or the other 30-50% that get very few searches each month, a given Web page was previously able to rank quite easily for a few targeted long-tail search terms. If a site employed long-tail optimization tactics sitewide, it could potentially recieve tens of thousands of visitors each month from long-tail searches alone.

With the introduction of Hummingbird, webmasters would be wise to stop chasing the long-tail synonyms for each general search query, but should instead focus on variants or similar questions. Instead of creating pages for “What”s an easy way to relieve allergy symptoms?,” “How are allergies treated?” or “How do you treat allergy symptoms?,” a website could have a single page about relieving and treating allergy symptoms and additional pages describing alternative treatments or treatment side effects (and other similar, but distinct topics).

It’s no longer enough to create pages relevant to a specific keyword phrase or search query. Content should be relevant to search intent, and the surrounding architecture should also support that theme. Relevance can extend beyond the page itself to a specific section of a site and the domain as a whole. Long-tail optimization is still key, but in a different way than before.

Structured Data & Entity Search

In order for a search engine to understand the intent behind search queries and filter through billions of Web pages to return the most relevant results possible, it needs a common ground for comparison. How can a search engine understand that a chain of hotels stretching across the world are all related to each other? What about a local burrito chain with a dozen locations? Is there some kind of database it references to connect these words? Well, yes, in fact, there is.

Hummingbird relies on a fundamental data layer of entities in order to organize the Web, although structured data isn’t a new thing for SEO. Google and Bing have been pushing webmasters to use Schema.org markup, a common format of structured data, in order to label entities and attributes within Web pages. Structured data can be used to markup reviews, ratings, movies, tv shows, restaurants, apartment complexes, law firms, hotels, and even people. Google has been incorporating an increasing number of data elements into search results via rich snippets, including recipes, sports scores, and, most recently, TV episodes and showtimes.

A year ago, Google’s knowledge graph was tracking 500 million entities, but, more importantly, 3.5 billion relationships between those entities. The more Google expands its entity graph, the better it will be at combing through the Web and selecting the perfect content to match each search query. If you’re not already using structured data on your website, now’s as good a time as ever to escalate the priority of those stories in your development queue.

Is SEO Dead, Again?

With every search engine update, many ask the question, “Is SEO Dead?” The answer depends on one’s approach to SEO. Are some “SEO” tactics dead? Probably. But, to be honest, they most likely weren’t helping very much in the first place. Buying links, link sculpting, hidden text, and keyword stuffing are not good ideas today, but there was a short period of time when they did “work,” if short-term success was the goal.

The way we approach SEO is sustainable. We focus on the user and what is best for connecting our clients’ websites with people who are looking for their content, services, and products. As SEOs, we are chasing the same goal as search engines and forgo short-term tricks or secrets for projects that support a complete long-term strategy for success. Google Hummingbird is a great thing for users and for our clients, and we look forward to future updates that continue to improve the way search engines organize the Web and improve user experiences.

by Jordan Silton, SEO Technology Lead


About Search Discovery

Search Discovery is a digital intelligence company that empowers organizations to make better, dynamic and transformative business decisions. We're focused on delivering results for our clients through digital analytics, SEO, digital media, and technology.


Comments

Let's Talk.

You're excited. We're excited. Give us a shout so we can get this party started.

!
!