When all the planets in the solar system are properly aligned, the task flow for a search is simple: the user enters a search term into an input field, clicks an accompanying button, often labeled "Search," and is taken to a results page that lists possible matches to her query. The user clicks the first result, which, naturally, goes to the exact content she wants to find, and is taken to the content page for that item.
In the article "The Power of Defaults" (http://www.useit.com/alertbox/defaults.html), in fact, Jakob Nielsen describes a 2005 study performed at Cornell University in which users clicked the top item in a set of search results forty-two percent of the time, regardless of whether or not it was the best result. Eight percent chose the second result. Even after the researchers switched the top two results, users still chose the top result thirty-four percent of the time. In short, users really want to believe the search engine always offers the correct result first.
So what happens when it doesn't? Do users click the Back button, read through the rest of the results, and make a second, more appropriate choice, perhaps sifting through several pages of results to find the right one? Hardly.
Rather than go through ten pages of results—or even more than one page—a user is far more likely to modify her search term and run a new search. Only a small percentage of users will continue on to a second page of results, whereas most users, most of the time, will opt to modify the search term.
Clearly, this simple three-step task flow—enter search term, view results, click—is far less reliable than it may appear. We continue this discussion throughout the rest of this chapter.