Wednesday, January 25, 2006

Searching and Browsing

The word browsing has a popular meaning that for most Westerners brings to mind the idea of going through a book store, or a retail store, looking around, interested, but not exactly shopping for anything in particular. You're looking, but you're not feeling goal-oriented.


Searching, on the other hand, is goal-oriented. You have something in mind, and you're looking for it as specifically as you can.


These two concepts are interrelated and co-dependent, but they are two distinct concepts. It's hard to search for something without doing some browsing, but you can in fact browse without really searching.


Consider the use of a popular search engine, like Yahoo or Google. You're combining these two concepts. You type in search criteria, and what you get back are search results, which is a misnomer. There is no such thing as a single result to a search. What you're really getting back is a subset of the information that's available, and you're invited to browse that subset of information.


Apply the same thought process to, say, people on Match.com, a popular dating site originally built by some friends of mine, Sterling Hutto, Thede Loder, and Piyush Shah (and others, of course, but I'm name dropping so I get to pick the set of names)....


In Match.com you know you're not looking at all of the world's population, you're looking at the subset who have registered with the site. So you're already browsing a subset of the world's population, just by entering the site. But you probably will narrow the search a little bit by constraining the browsing to male or female, and perhaps apply an age range.


This is where it gets interesting. Are you searching, or are you browsing a filtered list? I claim they're the same thing. That the concept of search is actually just the act of applying criteria to the process of browsing. How you traverse the data, how you actually do the browsing, is independent.


For example, Google gives you search results, a set of short descriptions of the pages it has indexed, and link to each. It presents them to you in "pages" or sets of 10 or so at a time, with a set of links to additional pages across the bottom. This has become a standard paradigm.


But the pages of search results are not themselves anything but a subset of the pages that they've indexed, although they are arranged in an order that makes them (according to Google) probabilistically more likely that you'll find them interesting.


Now consider a "rating" site like Hot or Not , which I didn't even know existed until I started researching this stuff. Hot or Not has a different mode for browsing: they show you each member of the site one at a time, and you click, effectively, a "Next" button to go to the next one in the list. It's amazing how addictive this is, behaviorally, since you don't know what (or who) is coming. They disguise their Next button as "Rate this person on a scale of 1 to 10" or "Yes" or "No" in terms of whether or not you want to meet the person. But rating someone as a 5, or clicking "No" is effectively "Next".


Now go back and think about Google search results, and imagine if they presented the search results to you, not in a list, as they do today, but with the actual page of the first result, with a "Next" button at the top, so you could literally look at each of the sites in sequence, one at a time, rather than poking through the list of results.


I am sure there are reasons that Google didn't choose this approach, but I use it as an example to point out that the acts of searching (or constraining a set of data by specifying criteria) and browsing are independent.

No comments: