More than half a year ago Jeff Weiner, head of search at Yahoo, took me aside and showed me something he was personally passionate about. He called it Search Spots – and the idea was really cool. It’s been done before in various contexts, but the idea was very well implemented and totally integrated with Yahoo and your browser in general – basically, anything you might be reading can become fodder for an inline search. From the email I received from Yahoo PR:
Last Christmas Jeff saw a news headline for the #1 single in the UK and wanted to learn more about the artist, watch the video, maybe buy the CD. It took him about 20 minutes to search for all this information. In a staff meeting he addressed this issue and asked folks around the table how long they felt it should take to access this type of related information – the consensus was it should only take seconds – but the current search experience was so linear it was prohibitive.
In parallel, one of our top engineers (and top search innovators) was off in his spare time creating a prototype to address the exact same pain point that Jeff was sharing with his staff. Y! Q. This product was not on the roadmap for product development but demonstrated such tremendous value to the user experience that it was greenlighted and set into production. With limited resources, he and his team went to work to create what is going to be pushed live late tonight on (next.yahoo.com) as a test product.
The URL is here. I plan to play with this, as it is Mac compatible (yippee!). More to come…Yahoo’s blog take is here…
Excellent new feature from Yahoo. This is not new innovation, but it is an excellent execution. Other implementations of contextual search have in the past been more “black box” in nature, hiding from the user which terms are helping to define the context, and hiding from them the insights this can provide. Yahoo’s approach here strikes a good balance between capturing the richer context that a page of text can provide, while opening up the black-box such that the relevant terms can be seen and manipulated by the user. This is a solid step forward. Hope to see this feature rapidply make it to the Yahoo! search mainstream.
I’m impressed with the pace at which Yahoo! is moving forward with their Search business, innovating and executing very well indeed.
Nice execution, indeed. And, as Jason says, no real innovation, here.
As a matter of fact, a highly similar approach to provide context-sensitive search capabilities has been widely used in the realm of academic and research libraries over the past 5 years. It has taken off to such an extent that a standard has been created that specifies how requests for context-sensitive services can be conveyed to a system that can deliver them. The standard has been formulated in a generic way, allowing many different uses by many different communities. It’s called the OpenURL Framework Standard for Context-Sensitive Services, developed by the National Information Standardization Organization . The standardization work is finalized, and release of the specification is expected very soon.
In the academic and research library community, systems that can deliver context-sensitive services are called linking servers, and many commercial offerings exist. SFX from Ex Libris is the market leader.
I just downloaded it and i have to be honest and say i just dont really like it that much. its a bit to manual, yuo have to highlight the search and then you can put in another subject? It didnt even pull up what i was looking for.
I agree. It is a great idea, however; the execution does not really reflect the ideal. It seems like instead of typing in key words, you highlight them. I also was having a hard time with the results. For the most part, I did not get relevant results. Wasn’t there something already like this out during the summer?
Y!Q is proving effective for me. After the first set of results appears the context can be further refined using the ‘more like this’ link alongside the most relevant result.
I also like the implementation, leaving me in control. Some other contextual search tools assume the entire page of text you’re viewing defines the context, often with poor results as the specificity is watered down by considering too much input text. By highlighting the most relevant paragraph or section I get better initial results that I can then refine further.
Seems Yahoo has some effective functionality here. It’ll be interesting to see how they tie this into search.yahoo.com and the class leading X1-based Yahoo desktop search – they’ve a very strong search offering developing. With Microsoft having its search platform in place and ‘now free to innovate’, I wonder what response we’ll see from Redmond?
The ‘more like this’ link would be helpful, but I don’t get anything relevant back. Maybe you have more experience with searching or are doing something different, but I can’t get the thing to work. The idea is great, but right now, i don’t feel the functionality is there.
Y!Q deserved a lot more attention than the launch of MSN Search, but is utterly poor at producing any decent result.