One of the larger dilemmas facing search engines is developing an understanding of what users want. At first glance, it would appear the task of understanding user intent is easy. Each time search queries occur, the searchers tell the search site exactly what they want. Or do they?
Billions of websites married to user queries combined with few industry standards for sites have created a mish mash of search results. Said mass of listing data is the cause of users' consistent complaints that search is too inefficient for them. Often, searchers say they can’t find what they want fast enough, and if a user can’t find what he wants the odds of an advertiser’s listing appearing in the right place at the right time are significantly diminished.
All of this talk about inefficiency has led to a search race to solve the user intent deciphering problem. Every week we hear about a new and improved search technology that purports to offer a newer and better search only to find the technology doesn’t perform quite as well as anticipated -- if at all. Fear not dear advertisers and commerce hopefuls, new technologies are arriving that are making strides forward in the race for smarter results.
One man’s search result is another man’s search result
Search activity is at an all time high and despite conflicting reports about a pending search advertising bubble-burst scenario, leading search sites are investing billions in developing new technologies to help facilitate an efficient interaction between searcher and result page.
A short time ago, MSN unveiled its plans to learn more about search activity by integrating psychographic, registration and third-party research to learn more about the searcher in an effort to better understand what the user means when he enters a query. Google is also researching next generation targeting methods as well.
Meanwhile, back at the entrepreneurial ranch, convoluted search results have led to specialized vertical search providers jumping into the fray with their own spin on search. Many of us are familiar with travel or automotive vertical search sites but if blogs are your thing, Feedster, might just be what you have been looking for. GuruNet, the makers of the original smart search tool, now offers answers.com, which provides, you guessed it, answers to questions in a search engine format. Other downloadable options like Grokker (more on this in a moment) have gone an entirely different way with turning search into a visual learning experience.
The focus of the most recent developments in search center around targeting or going vertical. While search sites have yet to provide the ultimate tool for users and many simply serve ads with existing platforms into self designated new and different results pages, very few have tried to address the query confusion problem at the point of search until now.
Where’s your mind?
Yahoo! is offering another answer to the user intent question. On May 27, the search/portal giant introduced a research labs project to the internet world that seeks to separate information and commerce search. Yahoo! Mindset uses machine learning technology to separate commerce and information searches.
The machine learning concept as it is applied to search is an interesting one. This technology uses algorithms that are designed to observe, mimic and learn from human behavior. According to Yahoo!’s definition, the learning began with a seed set of data created by humans. In this case, seed developed by Yahoo! engineers.
The dividing line between commerce and information search may be the biggest point of frustration for searchers. Simply put, one user entering the query, “fishing tackle” might be looking to buy the latest and greatest fishing supplies while another user might be researching the equipment used to catch the big ones.
Searchers on Yahoo!’s mindset are offered a slider tool at the top of the page to help sort the natural or organic results into “research” and “shopping” results. Sponsored listings remain unaffected by the slider. In the “fishing tackle” example, when I moved my slider over to the “researching” side, the dominant portion of results included informational content on why I should buy the site owner’s fishing tackle. On the “shopping” side of the slide, I was treated to much more overt examples of commerce sites.
Keep on Grokking
Grokker’s graphical search interaction has been in the press a great deal lately due to its expanded relationship with Yahoo!. You can still download the application or try the newer Java-based web interface. Simply put, search queries are integrated into actionable spheres that can be refined to help find the exact information searchers seek.
Opponents argue that Grokked results complicate what should be a simple interaction. I would point out the search interaction is anything but simple at this point. While Grokking your query might take a few seconds longer than performing a simple Google search for example, I found the navigation within these graphic spheres much faster than the all-too-familiar practice of changing keywords and phrases to perform several queries.
Grokker grabs the top 200 results from Yahoo! and places the searcher in control of shaping the result. According to a Grokker spokesperson, in short time we will see not only directive search integration but content channel mapping from Yahoo! as well. Moving forward, the other big G hopes to continue expanding relationships with other search engines as they are not working with Yahoo! exclusively. Grokker was also quick to point out that clickthrough rates in Grokked results are about four percent -- double that of average search clickthrough rates.
Yahoo!’s Mindset is by no means a perfect solution and they don’t pretend it is. “At this point in time, we’re just looking for feedback and input from early users of the tool to see how we can enhance the product,” says a Yahoo! spokesperson. My tests revealed both commerce- and research-focused results in spite of the slider’s location. Unfortunately, the very nature of the internet -- the ease at which almost anyone can create a web presence and the lack of standards in doing so -- will undoubtedly complicate and skew any results-sorting technology.
Despite the promise of better interactions and more targeted results pages, efficiency still leaves a lot to be desired. Yahoo! is on the right track with Mindset and the ever expanding partnership with Grokker, but we have yet to see how this evolution of search technologies will affect advertisers. As each of these technologies is introduced, advertisers will need a great deal of patience and tolerance in order to weather the series of beta storms to come.
iMedia Search Editor Kevin Ryan’s current and former client roster reads like a “who’s who” in big brands; Rolex Watch, USA, State Farm Insurance, Farmers Insurance, Minolta Corporation, Samsung Electronics America, Toyota Motor Sales, USA, Panasonic Services, and the Hilton Hotels brands, to name a few. Ryan believes in sound guidance, creative thought, accountable actions and collaborative execution as applied to search, or any form of marketing. His principled approach and staunch commitment to the industry have made him one of the most sought after personalities in online marketing. Ryan volunteers his time with the Interactive Advertising Bureau, Search Engine Marketing Professional Organization, and several regional non-profit organizations.
Ryan is the principal of Kinetic Results, Inc. a New York based online presence management firm.
Meet Kevin Ryan at AD:TECH Chicago, July 11, 2005