Google Algorithm: Exploring the Gap Between Speaking and Searching

By  |  Published 

As SEOs, we grapple with new (and not so new) questions about search engines on a daily basis. One of the most common quandaries we come across: “Why can’t I just ask Google a question?”The answer, of course, is that you can, but you have to articulate your question in such a way that it makes sense to the search engine. In doing so, you can maximize the effectiveness of the algorithms powering the search. Simply put, this means using keywords when you type in your query. If I wanted to know whether or not it was raining in Dubai, for instance, I would type something along the lines of “weather Dubai,” rather than “Is it raining in Dubai?”

Natural Language Search – and Why It Works 

The use of organic sentences, as opposed to strategic keywords, is called natural language search interfacing, and it’s by no means a new concept, but it hasn’t yet been adopted by any major search engines for mainstream use  (although Bing introduced a natural language search interface for shoppingand retail in March, 2011).

What difference does keyword search versus natural language search make? Well, when I input “weather Dubai,” I get the following:

searchcore_dubai1

However, if I type in “is it raining in Dubai,” then my top results are these:

searchcore_dubai2

The search engine is searching for instances of the word “raining” instead of picking up on the semantic clues that would indicate to a person (or, arguably, a natural language search interface) that I want to know whether or not it’s raining, (and so, would need to see the most current weather report, because that’s where the information would be).

Google Swims Against the Natural Language Search Current

Google has repeatedly shot down the idea of transitioning their interface towards one that is more based in natural language, but over time, they have reinforced a commitment to helping users find the information they’re looking for. In some instances, that information doesn’t manifest in the search algorithm, but is accomplished by processes that bring up widgets when words are structured in a specific fashion.

For instance, type in “Where is Hawaii” in the Google search bar and a Google Maps image pops up at the top of the SERP, showing you the islandof Hawaii. Type in “What is Hawaii” and you are instead shown a short, encyclopedic entry on Hawaiithat strongly resembles a dictionary listing. The search results beneath, however, do not directly address either question. Thus far, this is the closest Google has gotten to natural language search interfacing. 

Peter Norvig, Google’s director of research, explained more about why Google has ignored natural language search in a 2007 interview for the MIT Technology Review. “Typing ‘What is the capital ofFrance?’ won’t get you better results than typing ‘capital of France.’ But understanding how words go together is important,” he said.

To give some examples, Norvig explained that “New York” is different from “York,” but “Vegas” is the same as “Las Vegas” and “Jersey” could either mean “New Jersey” or something else entirely. “That’s a natural-language aspect that we’re focusing on. Most of what we do is at the word and phrase level; we’re not concentrating on the sentence,” Norvig said.

Natural Language Search Interfacing in the Future – Maybe

Despite Google’s stance, people prefer natural language searching over the use of keywords, and as people begin to interact with search in more diverse ways – 2011 heralded voice-search with Siri – natural language expression becomes more and more the norm of users.

According to a study on search query length, searches between 5 and 8 words increased by 10% between 2009 and 2010, while searches of 3 words or less were down by 2%. This suggests that users need to more thoroughly explain their queries, an inclination that makes sense as more and more services and data become accessible via the Web.

Google’s emphasis has always been on statistics. The leading minds at the ‘plex have insisted that a statistically-centered approach to search is preferable to a grammatically-centered one. Time will tell if they stand firm on this assertion.

As high technology becomes more of a consumer product, mobile devices have flooded the market, enabling fast access to the limitless knowledge bank of the Web. For most of these users, search is just a means to an end, so change to the dominant paradigm of input-keywords/read-drop-list-of-results is difficult to come by. Such a change would have to be usable and learnable by a massive, and massively diverse, population.

But aside from Bing’s venture into natural language search for retail, which will allow you to specify, for instance, that you want “cashmere sweaters under $100,” the competition hasn’t provided any reason for Google to adapt. It is possible that voice interfaces, like Siri, may force a change, but Google has already released its own Google Voice actions, which have enjoyed widespread appreciation by Android users, although it lacks the natural language interfacing that Apple fans have touted and embraced.

Based on the last year of algorithm changes, it’s reasonable to conclude that Google will not put much effort into natural language search developments until one of its competitors successfully implements it. Then, we can probably expect Google to swoop in, develop a $1 billion product, and leave everyone speechless. But for now, it seems as though they’d rather not.

Contact We Do Web Content

When your business is looking for quality, optimized web content for your site, newsletters, and social media marketing look no further than We Do Web Content. Having unique, organic web content is the leading way to increase search engine rankings and boost your Internet marketing. For affordable, quality content exclusive to YOUR website, blog and clients, contact us today!  1-888-521-3880.