Local Search Potential takes ‘Quantum Leap’: Yahoo!, Cornell Study of Local Search Capabilities Released

Okay, it’s not quantum computing….yet. However, a study released this week by Yahoo! Research in conjunction with researchers at Cornell University highlights the increasingly refined local search capability and capacity that search engines are developing.

Using a model based on probability theory to track the most frequent queries from Yahoo!’s logs, the research team led by digital search pioneer, Lars Backstrom from Cornell, found that the ‘probabilistic’ methodology they utilized (please forgive my tantalizing allusion to quantum theory, I couldn’t resist) significantly outperformed the simpler geometric techniques currently utilized by Google, Yahoo!, MSN et al. Using probability theory, the research team was able to pinpoint the natural geographic “hot spot” for any particular search query, as well as its ‘dispersion’ – how quickly the relevancy of a particular search term falls off as you move away from the geographic locale naturally associated with the search term. (The researchers used the names of baseball teams as an illustrative search term. The search term “Yankees” quite naturally had a wider geographic dispersion than the term Royals did, reflecting the national following of the New York team as opposed to the more localized following of the Kansas City ballclub.)

Even more interestingly, the authors of the study were able to track and predict how the natural epicentre, or “hot spot” of a search term as well as its dispersion can change over time. (The illustration they used was queries about “Hurricane Dean” as that storm blew in from the Carribean. Search intensity and dispersion increased and widened in real time, dispersion spreading widely just before the hurricane made landfall. In contrast, queries about the “Grand Canyon” – a national monument carved by time, but relatively unaffected by its passing – pretty much reflected the population distribution of the U.S. as a whole.)

As search engines begin to adopt this next generation of local search methodology it will “affect both search-based marketing and advertising efforts by region,” and will “also be useful as a component for search engine rankings themselves.” The methodology outlined in the study appears to be eminently applicable to everything from marketing political candidates (ugh…enough!) to focusing one’s online marketing efforts geographically and in real-time to changes in cuisine tastes, music preference, and even consumer responses to changing weather patterns.

The focus that search engines and leading academic researchers are putting on local search capabilities underscores how important a segment of online advertising and marketing local search has become – an importance that will only increasingly rise over time. The good news is that thanks to Yahoo! Research and Cornell we will soon have the ability to track that rise in importance in real time, as well as its dispersion across North America and the world!

No related posts.

Related posts brought to you by Yet Another Related Posts Plugin.

This entry was posted on Tuesday, April 29th, 2008 at 1:41 pm and is filed under Local 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.