Mobilizing Social Search

There is room for innovation and improvement in mobile search. This message came through loud and clear as I prepared a workshop on Mobile Search Future Prospects for the JRC Institute for Prospective Technological Studies of the European Commission. Content (what the individual is searching for) and context (the where, when, and why that may have prompted the query) change all the rules, diminishing the importance of universal search based on keyword queries and with it the chances of Google and Microhoo to dominate the mobile space.

Building on my own mobile search research, an ongoing project that began back in 2004/2005 when I researched and wrote the industry's first report on mobile search and content discovery, I currently track 80-plus search engine companies and providers across 10-plus categories of mobile search.

I group these tools and technologies into three main categories: Interface Search, based on how people interact with the search engine (via text, voice, and schemes that harness the camera phone to input queries); Actionable Search, focused on the interplay between what the individual wants and the results delivered (i.e., vertical mobile search services that index downloadable content and local search providers that give individual answers rather than lists of links); and Social Search, approaches that infuse human preferences and judgments to pinpoint truly relevant information and better answers.

Currently there is little reliable data on the magnitude of social search use, but the explosion of services over just the last few months is impressive. They range from Germany's hiogi, a community-based Q&A service that taps members to answer mobile queries, to Heystaks, a browser toolbar for both Firefox and Internet Explorer (with an iPhone app on the road map) that adds collaboration features and a host of social networking functions to your favorite search engine.

So does social search deliver a superior search experience on mobile phones? I explored this question in "Pump Up The Volume: An Assessment of Voice-Enabled Web Search on the iPhone," a performance analysis of voice-enabled mobile search services. (Note: While the report was sponsored by ChaCha Search, Inc., the opinions expressed in it are those of the researchers and don't reflect those of the organizations referenced.)

In our report, the first collaboration between myself and Albright Communications' Peggy Albright, the study's lead analyst and principal author, we found that ChaCha, in fact, "proved superior" to two other voice-enabled search options for the iPhone: the Google Mobile App with voice and Vlingo for iPhone, an application that allows users to direct their spoken queries to Google or Yahoo!. For the purposes of this study, Vlingo provided a spoken interface to the Yahoo! search engine.

Indeed, ChaCha provided exceptional results, with its human guides interpreting the search query accurately in most cases. According to our study, ChaCha interpreted natural language search queries (queries asked as questions) accurately in 94.4% of the tests and delivered an accurate search result in 88.9% of cases. Google's voice recognition technology interpreted queries accurately in 16.7% of tests and delivered accurate results in 22.2% of tests. The Vlingo for iPhone voice recognition technology correctly interpreted queries in 72.2% of cases and delivered accurate results (via Yahoo!) in 27.8% of tests.

To evaluate the overall performance of voice-enabled mobile services offered by ChaCha, Google, and Vlingo for iPhone with Yahoo!, we asked a series of 18 queries representative of six typical mobile search categories (Navigational, Directions, Information Local, Information General, Social, and Long Tail). For each query, nine performance characteristics were evaluated including response time, results accuracy, voice recognition accuracy, number of results received, key taps required, relevancy of the result, location awareness, use of advertising, and presence of other value-added features.

Our study found that the use of human agents to help interpret queries positively affects the quality of results delivered when compared to traditional algorithmic searches. In mobile search, social search and variations that tap the wisdom of crowds to deliver the right mix of answers and entertainment have a competitive edge over the plain-vanilla search we know from the PC.

Social will have a huge impact in mobile because the personal touch it adds to search results represents a perfect fit with the mobile phone-an intensely personal device. What's more, it benefits from the rise of mobile social networks, mobile-only communities that pave the way for us all to have more input into our search results.