Mobile phones are, after all, phones, so it’s natural that they respond well to voice commands. So it’s no surprise that voice search has grown dramatically given advances in the technology.
According to Google, speech inputs have grown 6X in the past year, as shown in the graph below. As voice technology continues to improve, voice search is quickly becoming more important in mobile content discovery and site optimization.
Source: http://www.blindfiveyearold.com/search-trends-to-watch-2011
The fanfare associated with Siri, the voice-activated virtual assistant on the most recent iPhone 4s, is the main reason that voice search has more in the spotlight, but voice search has actually existed for a number of years.
Google has offered a speech-input search app since 2009, available for download on iPhone, BlackBerry and Nokia s60. This functionality is also now standard on Android phones, and can be downloaded from the Android Marketplace if it is not already pre-loaded. Before these apps were available, Google even offered Goog411, which was a phone number that searchers could call to submit a search, and have the results read to them or sent directly to their phone via SMS.
Big Strides Have Been Made In Voice Recognition
Historically, the technological interpretation of voice commands has been less than accurate, which made early adoption and enthusiasm for voice search weak. Variations of different languages, dialects, pitches and accents must be cataloged, and understood, even in noisy conditions. It’s not an easy task.
Since those days, there has been significant improvement in the voice interpretation technology, which has emboldened both Apple and Google to pursue voice search more actively. Google switched to a cloud-based processing of voice searches, which builds understanding of each local language and dialect based on commonalities among queries, and the clicked results. Google says:“While initially we may not accurately recognize words spoken in every regional accent and dialect, one of the major benefits to Google’s cloud-based model is that the more people use Voice Search, the more accurate it becomes.”
While it was not widely publicized, the voice understanding technology that Google uses is probably relying on the same technologies that drive Google Instant and Google QDF and Google Quality Score, in which cloud-based analysis of a consistently strong CTR (and low bounce rate) for a lower-ranking site will naturally gain it credibility and rankings over time, as it gets more clicks than the listings above it. This kind of long-term reinforcement with clicks is probably informing and altering voice search results, or it will soon.
Here is what Google is saying about Google’s Voice Search:
“Our goal is to bring Google Search by voice to speakers of all languages. We follow a rigorous process to add each new language or dialect. Working directly with native speakers in each country, we spend weeks collecting spoken utterances to create the specific models which power the service. Our helpers are asked to read popular queries in their native tongue, in a variety of acoustic conditions such as in restaurants, out on busy streets, and inside cars. We also construct, for each language, a vocabulary of over one million recognizable words. It’s no small feat, but we love doing it.”
Similar Queries For Text And Voice Mobile Search
The good news about voice search, in terms of content discovery, is that it appears to simply translate a voice query then submit it thorough the same mobile search algorithms as if it was typed in. This means that there is no algorithmic change, and the only major change for a mobile SEO strategy would be to anticipate the longer and more casual, question-based queries, which are more common in voice search.
Google Mobile Voice-Based Search Categories are Similar to Mobile Text-Based Search Categories
As a whole, voice based search is most common for local searches, but this is likely because most early Android devices included Google Maps with the voice search functionality. Other than that, the categories that are commonly searched for in mobile voice search appear to be very similar to mobile text-based search.
The mobile ad network Chitika reports that the biggest category of voice search is “News” at 20%. After that is “Reference” at 17% then “Other” at 15%, and these are all largely similar to the text-based searches that they measure from in the Google Search Application.
Virtual Assistants Drive Data Usage
Interestingly, access to a virtual assistant like Siri appears to drive lots of engagement with digital information in the form of mobile data. The Huffington Post reports that: “iPhone 4S users transfer on average three times more data than users of the older iPhone 3G model….”
Some of the data consumption can be attributed directly to the heavy requests that come from Siri, as she scours multiple primary sources to get your answers, but another component is the reduced barrier to search. Without having to type, search and access to digital information becomes much easier and more casual.
While Siri has only launched in the US, Apple Insider reports that this will soon change:
‘Siri is still in beta mode, with limited functionality outside of the U.S. Apple is, however planning a rapid international expansion for Siri this year. The company is actively hiring iOS software engineers to help develop the Application Programming Interface for Siri and port the feature to other languages.’
Siri and her other virtual assistant competitors can complete a variety of tasks and access a variety of different information, beyond simply inputting a search query. They can pull information from a number of sources directly and only default to a web search if they do not have access to more specific information.
Create Feeds To Assist Assistants
This adds to the complexity of optimizing for voice search. Beyond anticipating what keywords people will use to search, we are now forced to anticipate what sources the virtual assistant will prefer or recommend. The good news, this may be less of a black box, and more influence-able if someday you are able to work directly with Apple, and submit qualified XML feeds.
All of the sources that Siri pulls data from are just websites that are sending Apple specialized feeds, so that Siri can access the information more quickly.
While it cedes a lot of information and power to Apple, it seems like the biggest win in voice search would be for Apple to work directly mobile resource sites to accept their information feeds. This would mean that users have quicker access to information about show times, character bios, episode summaries, celebrity bios and maybe even movie trailers and music news.
With this strategic move, you would become a primary source of information for your niche, and would not have to rely so heavily on Google to help people find your mobile content.
コメント