The perfect search engine would be a small microchip implanted in our brain. It would act as an instantaneous connection between the vast complexity of our brain and the vast complexity of the Web. To find something, we would just have to think about it and the chip would match that concept with the most relevant destination online.

Unfortunately, such a development hasn't rolled out of the Google Labs yet. So for now, we have to shoehorn our thoughts into a small quarter-inch by three-inch box on the search engine's home page. We have to distill our thoughts into a few choice words and hope this provides the search engine with enough to go by. And there lies the ultimate vulnerability point of search. Often, our ideas are too big to capture in one or two words.

Small Words, Big Searches; Big Words, Small Searches

We all have different intentions when we go to search. As I've mentioned in previous columns, many of us turn to a general search engine when we're mapping out unfamiliar territory online. When we define the boundaries of our concept, we often leave them vague and inclusive, because we don''t want to rule anything out. So, perhaps I'm at the beginning stages of considering a trip to New Orleans. I haven't done any research yet, so I'm looking for options and alternatives. My mind is open. This particular canvas hasn't been painted on yet. So my search is likely to be broad, i.e. "New Orleans." By keeping it broad, I know I should include everything on New Orleans.

We also use search as a navigation short cut to get to the most appropriate page on the Internet. We want to go directly from point A to B (again, the topic of a previous column) without a lot of detours to get in the way. Often, these types of searches happen well into the research phase. For example, let's say I had done a lot of research into New Orleans and in a previous session I remember seeing a page on upcoming events on the New Orleans''s Chamber of Commerce Web site. I don't have the URL and I didn't book mark it. So I go to the search engine and type in "New Orleans Chamber of Commerce Events." It's a very specific search that should take me right where I want to go. I don't want to see everything on New Orleans. I just want to see this one page.

Mapping Our Thoughts to Words

The challenge comes in the search engine trying to interpret my intentions based on my key phrases. Let''s go back to the first example. Although I've kept the search broad ("New Orleans") I obviously have a concept of the type of sites I'm looking for. They could be restaurant directories, accommodation guides, lists of things to do, official visitor sites, or other rich research sources. This is my concept, unstated to the search engine but residing in my mind.

So, when the search results come up, I'm looking at them through a "semantic map" that contains many words that flesh out my concept and might catch my attention. I''m trying to match the ideas in my mind with the results I see on the page. While I searched for "New Orleans" I'm actually looking for anything that might give me valuable and trusted information on how to make my trip to New Orleans more enjoyable.

The Eyes Have It

We've just recently completed two studies that show the impact of semantic mapping in the search process. One was an eye tracking study and one was an analysis of the importance of different factors in precipitating a click through. Based on these two studies, here's what seems to happen. The eye looks for a visual cue, generally the phrase we just searched for, in the title. Starting on the top of page on the left hand side, we start scanning down the page in an "F" pattern. While we're focused on the visual cue, our peripheral vision is open to the appearance of words that might match our semantic map. Even though we didn't search for any of these words explicitly, their appearance in the title and description has a strong implicit impact on which link we start reading. When there seems to be a match based on a quick scan including both where our eyes are fixated and the extra detail picked up by our peripheral vision, we switch to more traditional reading behavior, reading first the title and then the description from left to right. This lateral activity creates the horizontal arms of the "F".

As an example, we saw that people searching for digital cameras were presented with two listings from the same site, with almost identical titles. The listings were first and second in the organic results. Both listings promised "unbiased consumer reviews" in the title, after the query string "digital cameras." We saw fixation points on both of these visual cues. The difference came in what was shown in the description. In the second listing, there were recognized brands mentioned, including Kodak and Nikon. The vast majority of searchers quickly scanned past the first listing and started active reading of the second. It was a better match for their semantic map.

So, what does this mean? Well, it means that it's not enough to be No. 1. It's not even enough to make sure you have the query string in your title. To maximize the potential for click through, you have to understand what might be in your target customer's semantic map and match this through careful crafting of both title and description text. Bidding and organic optimization can put you in the right place, but you''d better have the right message too.