First published February 9, 2006 in Mediapost’s Search Insider
We’ve recently done a lot of testing on how people interact with search results, both on the general engines we all use, and vertical search engines in a few industries. We discovered a number of things, but one finding in particular surprised us. The user interaction with search results has been defined. A standard has been established. And until a discontinuous improvement in the search interface comes along, we will expect all search to be the same.
Google: The User’s Definition of Search
Google’s interface has become the de facto standard for search. Even now, all three of the major properties have very similar search results layouts, with only slight variations to distinguish them. It’s in those variations, the nuances of design and layout, where the differences in the user experience can be found. Everything is measured against Google, and at this point, Google’s interface defines the ideal search experience.
Information Scent in Search
First, let me weave together a few theories to give some background to how we retrieve and interpret information on a search results page. First of all, information scent. Almost every interaction we have with a Web site is to find some type of information. We have intent, we have a goal, and when we interact with a site, we want to get closer to that goal. This is especially true on a search engine. Here, our quest for that information is intensified.
Information scent says that most cues on a Web page have an inherent information scent about what could lie behind the cue. Every hyperlink or navigation option offers some “residue” of what we will find when we click on it. We assess all the cues on a page, and typically go to where this scent is the strongest.
On a search engine, we have been conditioned to believe that this scent will be strongest in the top organic listings. We naturally move towards these. The top sponsored ads happen to be in the path between where we typically orient ourselves (upper left corner) and where we want to go to pick up the information scent. Because of their position, they have a good chance of catching our attention. This behavior creates the Golden Triangle we identified in our first eye tracking study.
So, is position enough? No, we do want to verify this by confirming the scent on the individual listings. And here is an important point to remember. On the average, we take about six seconds to scan listings before we choose one on a search results page, and in that time, we scan four or five results (this is based on our previous research). But it takes about six or seven seconds just to read one listing. So we’re not reading them. We’re scanning them, and this is a crucial difference. In scanning them, we’re looking for patterns of words that seem to offer scent. This is the semantic mapping I talked about in a previous Search Insider (I’d Love to Search but Words Get in the Way). We’re spending no more than a second (or less than a second) to pick up whether there’s a pattern of words that offer the information residue we find most closely matches our intent. It’s a split-second decision.
So, how do we pick up these patterns? Here’s where Google has created one of its de facto standards. The first place we look is the title of the first listing (toward the left side), and we expect to see our search query bolded. This immediately reinforces that we’re on the right “scent.” From there, we quickly scan to pick up other words. The more hit bolding there is, the stronger the subliminal confirmation that this result offers strong scent.
Google does the best job of query hit bolding. Their use of fonts, the size, and the relative strength of the bolding quickly reinforces a relevant pattern. MSN, in contrast, doesn’t do any hit bolding on the title. Don’t be surprised if you see this change in the near future, as MSN draws closer to the Google standard.
Google also has a slightly different page balance. There tends to be more white space separating organic listings from the sponsored right side rail. The page looks a little less crowded and more usable at a glance. And as I mentioned in last week’s column(The 50-Millisecond Judgment), this split second judgment will affect our entire interaction on the page.
White space aids in the assimilation of word patterns. It causes them to stand out a little more. Have someone run the same query on the top three engines, then show you the three results for a split second each. Which page tends to offer the greatest chance of success? For the majority of us, I’m guessing that’s Google.
Another point on page balance. As I’ve said, our destination is typically the top organic listings. The biggest difference between Yahoo and Google is in how far down the page those organic results are pushed. They are significantly lower on Yahoo. Again, this runs against the standard of what we expect.
Implications for Enterprise Search
Finally, a quick word on enterprise search. For vertical engines and other sites in which search results play a major role, take the emerging standard defined by Google to heart. Understand that when people interact with your search results, they’re expecting a Google-like experience. The further you take them from that, the less ideal the user experience will be.
Google has done a lot right (and a few things wrong) but perhaps the smartest move it’s ever made is to pay meticulous attention to the search user experience. Whether the company designed an ideal interface by intention, or whether we’ve just been conditioned to accept it as the ideal interface, it works for us.