- November 8, 2016
This post was originally published on Street Fight on November 8, 2016.
Technology evangelists have been predicting growth in voice search for years, but new advancements in natural language processing are creating an acceleration in the market that few could have imagined. In just two years, voice search volume has grown from virtually nothing to accounting for 10% of online searches. That figure is continuing to rise, spurred by technology improvements that are making it easier for smartphones to accurately capture complex commands.
Voice search is changing the way people interact with their mobile devices, as well. Sixty-percent of smartphone users say they’ve started using voice search within the past year. These increases are easy to see when you look at search query volumes in Google as well. When consumers search with voice commands, they’re more likely to use natural language than when they type in search keywords. For example, rather than typing “dentist Brooklyn, NY” into a search engine like Google or Bing, people are more likely to use conversational phrases like, “My son chipped his tooth and he needs a dentist right away”. Natural language phrases tend to be longer and more nuanced but they also reveal greater intent, which is useful for marketers trying to capitalize on voice search technology.
Part of this shift away from the keypad is the result of improved word recognition accuracy. Back in 2013, Google’s voice platform accurately recognized just 80% of words. In two years, that rate has increased to 90%, with some technologies, such as Baidu, reaching accuracy rates of 95% and above. Experts believe that improving word accuracy rates will play a key role in the continued adoption of voice search technology. Speed issues are also likely to be a factor, as humans can speak 150 words per minute, while they can type just 40.
All of these improvements have made it easier for people to use voice search technology in their daily lives and as it’s gotten easier, voice search has become more ubiquitous. In September 2014, just one-in-10 queries came through speech (via Baidu). Less than one year later, in June 2015, Apple’s personal assistant Siri was handling more than a billion requests per week through speech. As of May 2016, one-in-five searches on mobile apps in the U.S. are voice searches.
At the same time, there’s been an increase in localized search volume. Local searches consist of 22% of all queries and according to Google, the number of “near me” searches has increased 34-fold since 2011. Eighty-percent of those local “near me” searches now occur on mobile devices.
What do these numbers really mean for the future of local search? For one, the acceleration of voice search queries opens the door for a continued rise in more wearable devices, such as Google Glass and Android Wear, as well as GPS devices and smart objects, like the Amazon Echo. Driverless cars will also rely heavily on speech search technology. By bringing together natural language processing with databases of business and location information, technology providers are helping to make some of these ambitious projects a reality.
In the coming years, technology influencers like Mary Meeker and others expect voice search to surpass all expectations as far as consumer growth is concerned. In fact, within just five years, Meeker anticipates that half of all searches will fall into the voice search or image search categories.
In order to meet those expectations, technology vendors will have to continue ramping up their voice search offerings. Already, vendors like Soleo, a local search and digital media company, are combining intuitive search technologies with natural language processing to better understand search context and intent. To do natural language correctly, Soleo came up with 10^22 possible search query combinations for the U.S. market. Its latest Local Search API version brings together natural language processing with a database of more than 20 million business listings.
Business listings are an important part of voice searches, particularly when “near me” searches are involved. Localized searches may see an even more rapid shift to voice than other searches, based on consumers’ increasing comfort with using hands-free devices in the car or when their eyes are otherwise occupied while outside the home. When this ultimate shift takes place and voice searches match or surpass text, it will be even more important for businesses to have a firm grasp on local and natural search, to avoid being left behind.