As profit margins from voice become razor-thin, operators are hoping to push lucrative content and applications to bulk up the bottom line. Ironically, they’re looking to voice-driven technologies to help.
Consumers’ difficulties in accessing content and services from a wireless phone are well documented. A study released last week from Harris Interactive found that nearly one-quarter of consumers said network speeds were too slow to use their phones for anything other than voice, and 19 percent cited confusing menu options as barriers to data use.
ChangingWorlds, an Irish artificial intelligence company, and Swedish consultant firm Mobile Metrix recently found that 65 percent of mobile content and applications on WAP decks required more than 12 clicks for users to access-enough to keep most would-be buyers from completing transactions.
To make matters worse, the exponential growth of data offerings available increases the number of menu options and deck layers, pushing more content outside the acceptable “click range.”
A number of speech-recognition technology developers are working to create easier shopping and messaging experiences on the mobile frontier. Instead of slowly drilling down one click at a time through a half-dozen deck layers to find the latest Eminem ringtone, for instance, a user could just say “new hip-hop ringtones” and jump directly to a menu of the latest music clips.
“Carriers have a variety of revenue-generating services that frankly have been quite a disappointment” in terms of sales, said Rich Geruson, chief executive officer of Voice- Signal, which sells speech-recognition technology to handset manufacturers including Motorola Inc. and Nokia Corp. “Basically, with each click or tap on your phone, you start losing customers-10 to 15 percent per click. One of the things among many that our stuff provides is the ability to eliminate the need for all those clicks.”
Like many wireless technologies, speech-recognition software has a decidedly mixed record. Earlier versions-most often used for voice dialing-were unintuitive and inaccurate. While voice surely is the most natural way for consumers to use their phones, uptake for speech-recognition technologies has been slow.
Many early versions were voice dependent, meaning users had to “train” their phones to understand commands. When calling, the few consumers who tried the technology often had to speak a name into the handset several times, and even then sometimes would find themselves on the phone with the boss when they tried to phone home. But developers say the technology finally has both the accuracy and the speed to gain mass-market traction.
“It’s only been in the last six months or eight months that the technology was really there,” said Samuel Poole, CEO of V-Enable, which integrates speech-recognition software with a mobile-search service. “One of the things we had to do was improve the technology so we got the access time down to three to four seconds on a (CDMA2000) 1x phone and one second on an EV-DO phone. … The other thing was to improve the accuracy. Right now, we’re consistently at the 90-plus percent (accuracy) level.”
Speech-recognition software developers aren’t the only ones hoping consumers will warm to the idea of using voice for needs other than conversation. Wireless carriers, content providers and messaging companies also are working to get users to talk to their phones instead of through their phones.
VoiceSignal technology is being deployed on 30 million phones, the company said, through Cingular Wireless L.L.C., Sprint Corp., Verizon Wireless and more than a dozen other operators around the world. The company works with phone makers and carriers to develop a user interface specifically for customers who want to use voice to operate their phones.
“A few carriers are really starting to understand that the next trend in mobile devices is simplicity,” said VoiceSignal President Dan Roth. “One of the things that historically companies have been slow to understand is the importance of the user interface as a differentiator. … Without a doubt, it’s the most important thing a user identifies with on a device.”
V-Enable has inked deals with eight carriers that expect to launch voice-friendly phones and services later this year. Revenues from content downloads and other potentially lucrative offerings are the driving force, said Poole.
“All the carriers right now have the same issue: They have so much content they’re trying to offer to consumers it’s a real challenge for them,” Poole explained. “Carriers have said to us they have so much content now they don’t even know what they have.”
Of course, the bigger problem is that consumers don’t know what’s being offered or how to access it. A recent study from U.K. game-maker I-play found that 30 percent of wireless users didn’t know whether their phones were capable of downloading content. Speech-recognition providers hope the technology is flexible enough to allow even the least tech-savvy consumer to explore everything a phone a can do with just a few words.
Other developers are using carrier-based voice technologies instead of embedded software to push wireless data offerings as they work to provide speech-enabled customer relationship management solutions. BeVocal Inc., a California-based speech-recognition company, powers a “Voice Store” for Cingular Wireless.
Instead of talking to their handsets, subscribers can call in to the automated store to buy insurance for their phones, upgrade messaging packages or check out other premium services. Such offerings can be coupled with speech-activated CRM services, allowing a carrier to segue from a customer-care call to an upsell opportunity. A user who calls in to check his text-messaging charges could be offered an assortment of mobile text- and picture-messaging packages through an automated speech-recognition service.
“The carriers are learning today that they actually need to increase the amount of time spent on the phone with the end user,” said Steve Tran, co-founder and vice president of marketing for BeVocal. “This is counter-intuitive to what the carrier did in the past,” which generally consisted of wrapping up a customer care call as quickly as possible.
For embedded speech-recognition technologies to take off, though, developers agree that usage must integrate seamlessly with other popular mobile platforms, including WAP, BREW and J2ME. Even the most efficient voice-driven applications will fail unless they can work hand-in-hand with other technologies to access information and download content. And carriers must find out not only how consumers might use speech technologies, but also must make users aware of what the software can do.
“Many of the carriers are in the pilot stages (of using speech-recognition software to drive revenues); they’re just rolling out and testing them to get a sense for what’s going to work in the marketplace,” said Peter Mahoney, vice president of worldwide marketing at ScanSoft, a longtime player in the speech-recognition space. ScanSoft offers both handset-based and network-based speech-recognition software and has deployed its technology with T-Mobile USA Inc., Verizon Wireless and Virgin Mobile USA L.L.C., among others.
“I think the most important thing we’re seeing is that there’s this cooperation between handset-delivered capabilities and capabilities delivered over the network,” Mahoney said. “For the user, we have to get to the point where it’s transparent where (technologies) start and stop.”
