Yahoo finally and officially joins the group of search companies getting on the voice search bandwagon, and announces that it has launched voice-enabled oneSearch for the Yahoo! Mobile iPhone app.

While Yahoo comes to the party more or less six months later than rivals such as Google, there is some indication that the wait was worth it if we consider that this service extends beyond allowing people to conduct keyword searches (for flight numbers, locations, Web site names, local restaurants – the works). People using the app can also use voice to customize the ‘My Interests’ tab. The procedure (according to the press release): “Simply click on ‘add anything’, speak the topic you’re interested in, then select the relevant content and add it to your page.” The Yahoo! oneSearch with voice application is currently available on more than 80 different devices and across platforms including Blackberry, Nokia, Windows Mobile, and now the iPhone – with support in eight languages.

Various bloggers have tried out the voice app, which harnesses speech recognition technology from Vlingo, and reported mixed results. But it’s difficult to judge the user experience based on random road tests. (This is why MSG has pooled its resources to produce mobile search research that, like my own mobile advertising white papers, offers readers a balanced assessment based on first-hand experience and solid methodology.)

The Yahoo app, however, comes in too late to be included in Pump Up The Volume, MSG’s own assessment of Web search on the iPhone. But that won’t keep us from conducting our own road test of the Yahoo app soon. Regular readers and Twitter followers (@peggyanne) may recall I announced the project a while back (a teaser before we had further refined our methodology to account for fundamental differences between natural language and keyword search, an important improvement that makes the results all the more compelling).

The white paper, researched and written in collaboration with Peggy Albright, MSG Associate and founder of Albright Communications, will be released next week. By way of background, our work assesses the overall performance of the voice-enabled search services offered by ChaCha, Google, and Vlingo in a typical range of use cases and scenarios. (Vlingo for iPhone converts queries into text and submits them to one of two search engines, Google and Yahoo. We chose Yahoo.)

A special highlight: A foreword by Bill Meisel, Editor of the specialist publication and voice technology knowledge destination Speech Strategy News. I’m honored to have him on board for the voice search white paper, and look forward to showcasing his analysis/columns on MSG soon.

I won’t divulge all the white paper results and stats here. However, I can say that ChaCha’s  results proved superior to both the Google Mobile App’s voice feature and Vlingo for iPhone.

ChaCha uses human agents to transcribe and interpret/answer search queries (spoken as questions). However, we found the performance of voice recognition alone does not determine nor predict the accuracy of a search result. Indeed, one search provider exhibited high voice recognition accuracy but still had difficulty delivering the intended search results regardless of query format.

Some background on the methodology of this defining work, which will be available for free download.  We created 18 queries representative of mobile search usage and trends. The queries covered search categories considered common in the mobile environment, such as navigation (to a specific Web site), directions, local information, general information on timely topics, and specialized or unusual long-tail topics (sometimes referred to as “dinner table” questions). We also included specific queries that represent the most popular mobile search terms in 2008, based on mobile search data publicly reported by AOL and Yahoo. Recognizing that the search engines used in the Google Mobile App and Vlingo for iPhone services are built from the ground up to handle keyword search – matching documents/information on the basis that they contain one or more terms (keywords) – we conducted a second set of tests using keyword queries.

Peggy and I are proud of the research and look forward to collaborating together on future mobile search assessment reports and work contracted by our various clients. I will keep you posted of our progress on MSG.

A key takeaway I want to leave you with: This white paper provides evidence that mobile social search – which harnesses human judgment, thus restoring balance in a model that tends to promote search engine optimized websites over destinations the user may find genuinely relevant and useful – has significant advantages over algorithmic computer-centric search approaches. As I have pointed out in this earlier analysis, in the case of the mobile phone, an intensely personal device we have with us at all times, a more people-centered approach represents a perfect fit with our search behavior and our expectations for a more personalized service. Indeed, the rise of mobile social networks further underlines our increasing requirement for search services that effectively inject human preferences into the equation. This, itsmy.com CEO Vince Staybl, recently told me was the primary motivation for the tie-up between his mobile social networking service with socially-assisted search engine Taptu, and I fully expect many more such partnerships to follow.

Disclaimer: ChaCha is an MSG supporter and white paper sponsor. The opinions expressed in the white paper are those of Peggy Albright and Peggy Anne Salz, and do not reflect the opinions of organizations referenced in the paper.

Tags: ChaCha, Google, iPhone, itsmy.com, Mobile Search, Mobile SEO, Taptu, Vlingo, Voice Search, Yahoo

This entry was posted on Thursday, May 21st, 2009 at 5:58 pm and is filed under Local Search, Mobile Research, Mobile Search, Mobile Social Networking, Multimodal Search. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.