Universal Search is not Exploratory Search

In a recent response article, Danny Sullivan takes Forbes CEO Spanfeller to task on the whole Google vs. The Newspapers issue.  There are a lot of things I agree with Danny about, and an equal number of things that I disagree with.  But I feel compelled to propagate one nugget from Spanfeller:

Spanfeller: Search is not really all that great at the moment, a comment repeated time and again by much more astute folks then me. This is especially true when looking for high-quality professionally created content. This is not to say that user-generated content or ecommerce options or product specs should not be returned in search results, simply that there is clearly a better way to showcase the different paths an end user might be pursuing. The idea that everyone is forced into trying to “game” the system so that they get their “fair” (or sometimes not so fair) share is testament to how terribly wrong this entire process has become.

This excites me because I see in this statement an acknowledgment and realization that Exploratory Search and HCIR (“showcasing”) is necessary.  Sullivan, however, completely misses the point:

The reality is that we have no dependable stats on whether search has improved or gotten worse. But I do have my suspicions when a publisher who believes he has “high-quality professional created content” tells me that it’s “especially true” that their publications aren’t being well served…As for different paths to showcase different type of content, here’s a newsflash. Google has what’s called universal search, where they actually segment things into groups like news, products and user-generated content (blog). Sort of what you’re wishing for, except that wish came true back in 2007.

Universal search as an aggregation of different content types (text, images, video, etc.) into a single ranked list.  It is not, however, a “showcase [of] the different paths an end user might be pursuing.”  A data type or information representation (ascii vs. numeric vs. pixel) is absolutely not the same thing as an information seeking path.  Using my ongoing example of off-beaten-path cafes in Prague, the fact that universal search mixes in one or two images, videos, patents, finance listings or snippets of Google Code into the results list for my [hidden prague cafe] query does not (necessarily) help me with my need to find off-beaten-path cafes.  Given that most of the web results already show the popular cafes, any found images are likely going to be of those same popular cafes as well. My desire is for a different kind of cafe, and universal search does not help me find those cafes. Sullivan is incorrect in saying that Spanfeller’s wish came true in 2007. Universal search is a mixture of data types, not of information seeking pathways. It is a mistake to conflate the two.

Spanfeller’s is spot-on when he says, “The idea that everyone is forced into trying to “game” the system so that they get their “fair” (or sometimes not so fair) share is testament to how terribly wrong this entire process has become.” This is because, universal search or not, Google still does not give the user any way of refactoring or examining results in an alternative, exploratory manner. The user is forced to follow the ranking that the search engine gives, and the only way to explore the 1.3 million results returned is to step through it, one results page at a time. Even personalized search does not address the problem, because the user is still given only a single pathway through the 1.3 million results. As a result of this undifferentiatable unidimensional ranking an artificial scarcity is created, and everyone is forced into trying to game the engines through SEO.

I, like Sullivan, do not know if search has gotten worse.  I only know that it hasn’t gotten much better.  I don’t need any stats to show me that. I can see it every time I have an exploratory-oriented information need, and am forced to crawl through pages and pages of universal results, one link at a time.

This entry was posted in Exploratory Search, Social Implications. Bookmark the permalink.

Leave a Reply

Your email address will not be published. Required fields are marked *