All that cash being poured into voice-activated search is beginning to look like a good bet.
A recent study from Usable Products Co. indicates speechrecognition technology finally may have become user-friendly. The New York-based market research firm tested four mobile search offerings-three text-based services and one voice-controlled application-with 80 consumers, allowing users to evaluate each offering for an hour.
Participants test-drove solutions from InfoSpace Inc. (a WAP service), JumpTap Inc. (Java), Nuance Voice Control and Yahoo Go. While “none of the four search solutions was a clear winner,” according to Usable Products President Scott Weiss, consumers preferred talking over triple-tapping.
“Users predicted voice search would be the worst of the four search products, but in final usability, it performed better than expected,” said Weiss. “We were surprised that participants enjoyed voice search, and how much more they liked it than searching via phone keypad.”
Ad-supported search relevant
Perhaps more importantly, nearly four out of five participants found that ad-supported search delivered more relevant results than searches without commercial messages. While consumers generally found sponsored, text-based ads annoying, more than one-third of participants said banner ads actually enhanced the user experience.
That finding provides a welcome contrast to other studies, most of which have discovered that consumers are loath to receiving any sort of pitch on their phones. And the study seems to validate the millions of dollars being thrown around in the speech-recognition space: Microsoft Corp. earlier this year acquired TellMe Networks in a deal rumored to be in the $1 billion range. Nuance has seemingly outspent Microsoft lately, shelling out more than $1 billion as it swallowed up competitors and other developers in the last year or so. Google Inc. also offers a voiceactivated search service, and several smaller players such as VEnable and Promptu are also finding some space on the playground.
The players are vying for a piece of a U.S. mobile search market expected to reach $1.4 billion in 2012, marking a compound annual growth rate of 112%, according to figures released last week from The Kelsey Group.
Voice vs. thumbs
How much of those revenues will be driven by voice-rather than thumbs-is the subject of much debate, however. Unlike many new technologies, speech-recognition offerings have been on the market for years, giving consumers plenty of opportunity to try them-and, oftentimes, to be infuriated by them.
“People don’t approach speech from a position of innocence. Everybody’s been in voicemail jail, everybody’s played speech-recognition guessing games,” said Opus Research founder Dan Miller, a longtime observer of the space. “A lot of people’s first experience was, ‘Wow, this is kludgy.’ But it’s less so now.”
Less-than-perfect technology can still make for a painful user experience, especially during searches for obscure content. And advertisements may prove more difficult to deliver through speech-enabled offerings, which can require an awkward back-and-forth while a user views a screen, listens for voice prompts, speaks into the phone and then repeats the process. But one of the keys to a successful voice-driven application is minimizing the calisthenics, Miller said. A voice command could launch an application, perhaps drilling down through a layer or two, before the thumbs take over.
“Once you’ve entered your query, you’ve gotten a result and you have that in front of you in a navigational format, that’s a form of a good outcome,” said Miller. “That’s a positive experience for the user.”
There are plenty of situations where voice is an inappropriate way to navigate an application, of course. Triple-tapping can covertly allow consumers to access information during a business meeting, for instance. Speech-recognition software probably won’t help much in the front row at a deafening rock concert. And consumers will certainly be unwilling to speak aloud when it comes to entering data such as Social Security numbers or PIN information.
Which is why speech should be just one part of the equation, Miller said. Consumers should have the flexibility to employee as many different data-entry avenues as possible, and should be able move easily from one to another.
“The idea is to be able to reach the broadest audience possible in a way that makes sense for the user,” according to Miller. “Speech could figure into all of them, but it’s only one of many modalities.”
