At the company’s 15th birthday celebration last week, Google announced its largest algorithm update to date. As part of that announcement, the company revealed to search-industry professionals that the update has been live for a month.
What? How could a massive update—one that affects 90% of search queries worldwide—have escaped the notice of SEO geeks everywhere?
The answer is simple, but profound. Those of us who have been around the SEO industry for more than a decade think and monitor Google by using traditional thought processes and queries. But the world has changed. The way people search for information isn’t the same as it was fifteen years ago.
Google’s Matt Cutts foreshadows this algorithm change in a video released on July 8, 2013, in which he introduces an alternate way of thinking about context-based queries. The concept he reveals in this video suggests that Google is challenging the status quo as they research user expectations around voice search.
As Google focuses on its mission, “To organize the world’s information and make it universally accessible and useful,” the company continues to innovate in ways that make it easier and faster for consumers to connect with relevant information. The Hummingbird Update is just the latest in a multi-year series of updates, designed to ensure that Google supplies relevant, timely information in the fastest way possible.
This update feels a bit like the Google Instant update in 2010. Both were significant updates that affect user experience; however, the changes feel like a logical, natural progression in search engine user interface design.
In the near-term, Hummingbird will be most noticeable to SEO industry professionals who play with contextual search technology. We will start to search for information by defining a topic, and will then iterate deeper into search with pronouns instead of explicit searches with all necessary keywords.
For example, I did some experimenting with my Android device by performing the following list of search queries:
Q: Who is the president of the United States?
A: Barack Obama
Q: How tall is he?
A: Barack Obama is 6’1” tall
Q: Is he married?
A: Barack Obama’s spouse is Michelle Obama since 1992
Q: How old is she?
A: Michelle Obama is 49 years old
At this point I was pretty impressed. So, I decided to push Google’s version of Siri a little further and found Google Voice Search assembling queries for me:
Q: Do they have children?
A: Google Query: “how old is michelle obama children”
Q: Do they have pets?
A: Google Query: “do they have pets”
In both cases, the contextual thread was broken and the results started falling apart. I then proceeded to test a few other searches. Google was able to provide me information about the CEOs of some public-facing companies. I also found querying for details about movies works well.
Over the next year, we will surely see this technology’s effect on the way people search. As consumers learn that they can search more efficiently, and understand the power (and added safety) of voice search, it will become a commonly used tool, at least on Google devices.
The more important thing to consider is the trajectory. The Hummingbird update gives us an interesting piece of technology that we can show off to friends and family. I’m sure that my wife and I will begin a new round of Android vs. iPhone battles in our home. However, as online marketing professionals, we need to start planning for where this technology will be a year or two from now. Websites with thorough, well-organized product information are going to win.
Today, Google doesn’t impress me when I ask, “What is the top-rated washing machine?” But in the next few years, as we see Hummingbird v2 and beyond, and as developers figure out how to build for this type of search, the following series of queries will be realistic.
Q: What is the top-rated washing machine for large families?
Q: How heavy is it?
Q: How do users rate its reliability?
Q: Is it quiet?
Q: Will it eat my socks?
Q: How much is it?
Q: Where can I buy it today?
Q: Do they use high-pressure sales techniques?
Take note of the type of data that will be required to properly answer these questions. Product reviews need to be available and evaluated with natural language processing to determine how large families feel about their washing machines. Next, manufacturer data is a reliable source for product weight. Then, to address reliability, it is probably best to consider all reviews for the product, but maybe even reviews for similar products from the same brand. After that, a combination of manufacturer data and interpreted user-generated content will help accurately answer the quietness question. Regarding socks being eaten, this question will most likely be a distraction that will result in a search query.
However, it will be important for search engines remember context from previous queries, unlike Google when I asked, “Do they have children?” while researching the Obamas. Then, for the cost and location queries, current geo-coordinates need to be compared with those found in local store pages, which hopefully contain accurate price and inventory information at a local level. Finally, local store reviews must be considered, so that consumers are directed to stores offering an amazing shopping experience.
The future of voice search will feel something like this within a couple years. Consumers will expect accurate answers to questions like these as they use Google Voice Search and Siri to guide their shopping, entertainment, and service provider decisions. To reap the benefits of this technology trend, it will be vital to provide extensive structured markup (schema.org code) and various types of user-generated content. Those who embrace this trend will be the winners as we enter the latter part of this decade.