Is Mobile Voice Search Hot?

16 Comments

If you have tried to do search on non-qwerty mobile phones, then you are all too familiar with the frustrations that go along with it. Piper Jaffray projects the global mobile search revenues will generate $11 billion by 2008. While that number might seem a tad overoptimistic, the lure of the billions has most mobile search start-ups and many big companies looking for alternative ways to conduct mobile search.

Start-ups like 4INFO are using SMS for search. Others like PromptU are using voice-based search technologies, where you speak into the phones and the search results are sent back to your handset. Another company that is making a lot of waves in this space is San Diego-based V-Enable. The company showed off its search capabilities at CTIA and claims that it was getting 90% accurate results and was sending them back to the mobile phone in less than a second.

But will it be enough for these tiny tots to beat giants like Google, which just today filed a patent for a voice-based search technology.

What are your thoughts on mobile search and which technology would you bet on in the mobile search domain? Will Google be the Google of mobile or do you see an opportunity here for another vendor? Send me your feedback, and post comments.

16 Comments

geek

In a Google study published at CHI 2006, users entered on average ~ 2 keywords per query (internalized as mental model of web search), across all new search mediums: phone, PDA, web speech interface. Unfortunately, this indicates that even for limited number of words per query, speech recognition systems don’t convert speech to text accurately in all cases.

For now, Google is advertising SMS mobile-specific-search on the forefront of its XHTML interface.

Robby

I think the voice search offers a myriad of new possibilities that text-based searches cannot. Products like AskMeNow.com and Free411.com will emerge as major players in this area. It can be immensly easier to call-in a request versus texting it in, especially when dealing with something like driving directions or directory assistance.

Jouko Salonen

My GMAIL knows how to read Finnish language inflections and is able to bring correct Ads even if the mail includes pretty difficult inflected words. So I think Google does good R&D work in this area. Mobile Voice Search includes difficult language specific issues but I think Google has here a clear first mover advantage. and, most importantly, a massive, existing, global user-rhizome who instantly starts providing real life data for pattern recognition analysis, as soon as Google launches first BETA.

Neil Edwards

Mobile search will be all about accessing sources of mobile-friendly content. Many companies are playing with new search clients for the phone like SMS or Java search clients. Search on the phone will be driven by the browser at the end of the day just like it was driven by the browser on the PC. The fundamental problem is discovering or finding “real” made for mobile content on the Internet. Most sites format for a PC and a broadband connection. Many browsers, search, and rendering software solutions have popped up guestimating on dotCom content and sending the whole thing to the phone. Try it. It is not a good user experience and you do not get what you want. It comes back to having a good source of content where the content author followed some type of open standards that guaranteed their mobile site worked for a large audience. The key to good mobile search will be to finding the source of good mobile content.

Lalit Sarna

The fundamental issue with access to information through the mobile device is the attention span of the users. Most activities on the mobile are on the go activities. Ergo, impulsive needs for access to information quickly and efficiently. User s do not have the time or patience to click through several results and do comprehensive research.

While Different UI’s take different approaches on the most optimal way accessing information through the mobile device, the inherent problem is largely ignored. The quality and relevance of the information. Generic approach to search works well if the user knows exactly what they want, like the name of the store or directions. However generic search is not very conducive for discovery of new products, services, media through a cell phone. For mobile high specialized vertical solutions personalized to user’s demographics, tastes and location are desirable.

Our approach to personalized mobile search is enabling community aided vertical social search. We enable users to discover products, services media based on reviews, recommendations and rankings provided by trusted peers and like minded people. Further enable filtering of peer promoted search based on the contextual relevance of the reviewer to the content and to the consumer of the review. If I am searching for music then I want to find top rated music from people who have similar tastes as me and who have an existing reputation for providing good reviews. The perfect solution to mobile search would be to pair up optimal UI with a vertical social search engines that enable device specific customized access to data.

Disclaimer : Following information is description of our approach . If you want to know more contact me at [email protected].

As far as creating these vertical knowledge pools, we have taken the approach of creating user traction by enabling brands who have an existing user base to create interest based communities addressing the needs of their users across varied demographics. Thus enabling users to interact with and around the brand and attracting the attention of their peers to the brand. Much like Myspace was able to co-brand its community with the 18-24’s focused around music. However, One size never fits all and you don’t always want to wear red. Today you are a music fan; tomorrow you are a parent, day after you are an avid traveler. Users have varied interest and needs to interact with various communities, what they need is to be able to take their reputation, profile, content to the community that best fits their need at the time.

We enable brands to consolidate the resources on the back end while build highly targeted individually branded front end communities. These communities can be all owned by one brand or brands can pool resources with strategic partner communities. Thus created an ecosystem of connected consumption and increasing their interconnected reach. Through structured communication, interaction, collaboration and contextual reputation Index communities evolve into vertical social search engines.

An appropriate analogy would be an individual shop Vs a shopping mall Vs Wal-Mart. Myspace = individual shop, individually branded interested focused, group of inter connected communities with consolidated back ends = shopping mall, Google = Wal-Mart. Wal-mart provides a limited selection of low quality cheap goods in various categories. Individual shops are spatiality boutiques providing large selection of high quality higher priced goods specific to their ctaegory. Wal-Mart can offer lower price primarily due to its ability to buy in bulk for the masses. We enable the individual shops to pool resources on the back end cut cost and gain a collective large volume of consumer flow thus enable them to provide high quality goods at competitive prices of Wal-mart.

C-SPAN geek

Mr. Schmidt’s comment:
‘We’ve looked at [voice enabled search] and it’s hard because of the ambiguity of language. You can do very good answers on trained sets with specific categories. It’s very hard to do [voice enabled search] in general for everyone, all the time’ (World Economic Forum, 2006). Or, simply put ‘No’ that futuristic AI research stuff won’t be coming out in the next year or two.

Russell de Sa

The patent was actually filed in 2001…maybe it was granted yesterday?

Danny Sullivan

Considering they launched voice search back in 2001, I vote not hot any time soon. It’s been down for well over a year, if not longer. Not only do people not realize it already exists, but no one complained when it disappeared. Cool they have a patent; sad they lost the product, I suppose.

Ugo Di Profio

Automatic Speech Recognition (ASR) may be good enough in controlled acoustic environments, like your home, but it is not robust enough considering the quality of hardware and the mobile everchanging environments where you may want to make a spoken search. In other words, baseline recognition quality would be poor.

On top of that, and a bigger issue is that current search engines are optimized for key word(s) search and those keywords can be very specific and very very difficult to recognize, even for an ASR system with a large vocabulary in an ideal environment.

The merit of ‘asking’ rather than ‘typing’ a query, is that one can be wordy; the drawback is that we are limited in the choice of words we can speak (and reasonably expect to be recognized).

The lure of Spoken Search (and, more in general, spoken interaction with computers) will drive R&D in speech recognition to a point where it may be usable, but the real challenge for Spoken Search is to develop new search engines and methods based on connotative description(s) as opposed to the current denotative keyword(s) approach.

User: “I need the recipe of a French dish of boiled meats and vegetables”
System: searching for “Pot-au-Feu”

Ben

Could TellMe be an interesting partner/acquisition candidate for GOOG?

skibare

If Google figures out “”Mobile Speech”, please notify Fidelity Investments cause Fidelity must have the Dumbest VOICE Recognition on the planet!
skibare

Nicole Simon

In order to gain “global revenues” one has to start making business globally. So far, I don’t see a lot applications or business models which truely embrace the fact that there is a world outside the borders of the US.

I notice a lot of people getting tired of the fact that even in this world of the Net there still are so many just US based companies. The Asian market is seperated fromt the European through language issues.

But Europe – at least the part I am talking about – is capable of doing business in English. If you would let us.

Mobil info, for example combining routing information with Google Maps? Good luck. We are happy to have country borders for main Europe. SMS services to receive information? Yes, it is very complicated to actually take the mobile providers and say “guys, we would like to have a global campaign and market world wide ‘send text to xxx’ and do response”.

It is the basics which are not covered, but then again: This lays a lot of ground for interesting European startups which will start earning money on a market ‘you’ could have had easily.

Jackson West

I’ve used Google SMS for a while now, and while it didn’t always give me the info I needed, it worked pretty well. Now I have a web interface on my phone, so I just use that. But I’m a lot more willing to put up with the numeric-alpha keypad than most folks.

While Promptu certainly has a good chance to carve out a niche with carriers, I would love something more generic. A neural network that ‘learned’ phrases by listening to voices and gathering user input about the relevancy of the results would be an interesting solution. In other words, the search engine is built by the user queries (though obviously it would need to be thoroughly seeded first).

isb

While I think the UI issues in mobile search are interesting, contextual information like location will be highly influential. In the future, I see mobile search evolving into a collection of “context based” search verticals.

scott shaffer

Mobile search isn’t the correct term, it’s mobile info .

What and how you query will be completely different using a phone than a PC.

No, Google will not be the “Google” of mobile search. Unless Google can find a way to utilize the camera on the phone, and a RFID reader, they will be missing out on the biggest database for mobile search.

The camera and RFID reader WILL BE your “mouse” for mobile search.

Physical actions and your location will replace typing and keywords for search.

Benjamin

Surely the first step in this market is for mobile phone manufacturers to optimize their handsets for full voice operation. So far this functionality is unreliable and limited.

Voice search is still a long way down the road.

Comments are closed.