A while back, we got an offer we couldn’t refuse. The good folks at Mirametrix asked if we were interested in custom eye-tracking data (which traditionally costs a small fortune) for any Google searches. Um, does Matt Cutts like cats?

Since I once worked down the hall from an eye-tracking lab, I was the obvious choice to lead this shopping spree at the nerd candy store. So, we picked 5 different Google SERPs, representing the diversity Google has created in the past couple of years, including the newly expanded site-links. This is the story of those SERPs. They’re all about pizza, because I’m from Chicago and was apparently hungry when I made the list.

The Equipment & Methodology

First, a little bit of background. Mirametrix produces affordable, portable eye-tracking systems for researchers. Our data was collected using an S2 Eye Tracker (shown to the right), which looks a little bit like an Xbox Kinect. Each SERP was shown to 8 subjects between the ages of 18 and 30 for 30 seconds. Subjects were told the search term of interest and then were allowed to view the full-screen SERP freely. All SERPs were de-personalized and localized to Chicago, IL.

Heat maps were created by aggregating the subject data. Subjects saw the full-screen SERP, but I’ve cropped each image below the point that activity trails off. I should note that this is actual eye-tracking data, which should not be confused with “click maps” or heat maps created from mouse movements. These patterns come from people’s direct visual interaction with the SERPs.

(1) Local #1 - "best pizza in Chicago”

I’ll start with a query for “best pizza in Chicago”, because the results are probably closest to what you would expect. I picked this particular SERP because it had strongly integrated Local/Places results, along with maps on the right. The eye-tracking data looked like this:

You can see a modified “F-shaped” pattern here, with most activity centering on the top results and some eye movement toward the map. As you might expect, the top listings attracted the most attention.

(2) Local #2 - "pizza"

Next up is a local search for “pizza”. These results were more varied, with a couple of organic results followed by an integrated 7-pack that more clearly separated Local/Places results. This data got a bit more interesting:

The impact of the Local 7-pack appears to be very powerful, drawing attention from the top 3 organic listings. Even map fixations appear to be toned down from the first SERP. Whether this is because people are familiar with Google’s local results format or are attracted to the distinct formatting, it’s clear that they were biased toward this part of the page.

(3) Video Thumbnails - "how to make a pizza"

This one was a special request from the Big Boss – Rand was interested in the impact of video thumbnails in organic SERPs. I found that the query “how to make a pizza” brought up video thumbnails for the #2 and #3 spots. Here’s what the data had to say:

Although individual results are a bit hard to separate, it does appear that subjects’ eye movements focused on the first video thumbnail, possibly even at the expense of the #1 organic result. Especially with something as visual as a pizza (who doesn’t love pizza?) the attraction of an image could really tip the click-through scales.

(4) Product Images - "pizza cutters"

The next search was for “pizza cutters” – this brought up brand and store searches at the top, along with images for shopping results after the 3rd organic listing. The eye-tracking data looked like this:

There’s definitely some pull toward the product images, although the top organic results still do fairly well. The “Related Searches” seem to get relatively little attention, even though they appear where the first organic result would usually be.

(5) Expanded Sitelinks - "Pizza Hut"

Finally, we decided to test-drive the new site-links. A search for “Pizza Hut” brought up 6 expanded site-links. Not surprisingly, this search also triggered some local results. Here’s the visual:

Although the Pizza Hut listing gets some fixations, there seems to be a strong pull toward the local listings. Even with a full pack of expanded site-links, the main Pizza Hut site got much less attention than I would have expected. When you want pizza, you want pizza, not a corporate history.

Some General Implications

I think the first and most obvious implication is that, as Google moves away from 10 plain listings for more and more searches, it is definitely having an impact on search users. You need to be familiar with your competitive space and take advantage of SERP enhancements, like video thumbnails. Ranking #1 might not be pulling the weight it used to if your competitors down the page have more visually interesting results.

These results also suggest that the in-page Local/Places results are having a strong impact, even if they fall in the middle of the page. In these limited cases, they seemed to pull attention away from the top organic spots. If your query has a local flavor, you need to be aware of how your Google Places page is competing.

Of course, these are exploratory results, and more data would be needed to back up any given finding, but I hope the general observations are interesting. I’d like to thank Ben Yoskovitz for setting up this opportunity, and Anton and Amineh at Mirametrix for managing and running the eye-tracking studies.