Voice search to replace traditional methods for searching has a long way to go.
Data from a study released Tuesday shows that less than 20% of respondents use a smart speaker to search for information more than five times per week. “Never” and “less than twice per week” made up 56% of all responses to the question of how often it is used to search for information.
Perficient’s fourth annual study focuses on how people use voice commands with their devices.
More than 1,000 participants answered 30 questions. In addition to how they use voice, respondents were asked to describe when and why they use voice on their device. Analysts then compared responses to the same questions in the studies that took place during the last three years.
advertisement
advertisement
"The study data shows us the 'voice revolution' is not here yet," wrote Eric Enge, general manager at Perficient, in an email to Inside Performance. "People are not ready to use voice as their primary mode of input to devices."
Perficient also wanted to also explore the reasons why. Some 59% of those participating in the poll agree or strongly agree that it’s frustrating when the device cannot understand the request.
Nearly 36% of respondents said they want to use the device less when it doesn’t understand them. Some 16.9% said they don’t use voice commands. More than 41% said an incorrect answer will reduce their use of voice.
Mobile phones are the most common devices for using voice commands, at nearly 40% more than use with smart speakers like Amazon Alexa.
When participants to the study were asked to cite how they ask questions on their smartphone, the rankings were as follows:
Personal assistants also have a difficult time understanding voice commands, according to the data. When asked how well their built-in personal assistants on their phone -- Google Assistant, Cortana, and Siri -- understand them:
While the study focuses on the choice between keyboards and voice, Enge doesn’t believe that consumers -- many of whom now prefer non-touch services -- will have an effect on voice search, because speech-recognition technology has a long way to go from becoming a natural-language conversational tool.
Voice recognition for individual words may be quite high, but understanding full sentences is weak, and the ability to understand chains of sentences together or the ability to hold a sustained conversation is virtually nonexistent, Enge explains in the email.
What’s the message to marketers? "Marketers need to find and consider specific opportunities rather than think the whole world will rapidly become voice-centric,” Enge said.