Page 1 of 4

Eye Tracking Report: St. Joseph Channel Website
 
Sujoy Kumar Chowdhury
Victoria Rodgers
Missouri Western State University
May 12, 2010

Overview:
 
We evaluated which advertisement panels within the St. Joseph Channel homepage were viewed the most.  Eye-tracking equipment was used to measure where and how long each participant viewed different areas of interest, we refer to as “lookzones”.  The lookzones were divided into two groups, content and advertisement (ad) zones.  The content zones have information relevant to the website; whereas the ad zones contain advertisements from other organizations.  We found that the ad zone named R12A, which is a tall ad-panel located at the right side of the webpage, was viewed the most in comparison to other ad zones.

Procedure: 

Nine users participated in this study.  These participants were asked to sit at the eye-tracker, while a researcher calibrated the machine.  The participants were asked to look at the screen images of the website as if they were ‘surfing the web’.  Each image was displayed for five seconds. The images were broken down into lookzones with a three-part naming scheme. The identifiers of the lookzones began with either L, R, or C to indicate left, right, and center column respectively.  It was followed by a number that represented the sequence in which lookzones are located in each column.  The numbers started from one at the top of the webpage and increased downwards.  The last letter indicated whether the zone is a content zone (C) or an advertisement zone (A).  In addition, lookzone R1A included two lookzones (R11A and R12A).  This is because the R1A lookzone was split into two different lookzones in the some of the screenshots.  Although the website contains a number of dynamic elements, only static screenshots were used for this study.  Four screenshots were retrieved on February 4 and February 12 for the two top and two bottom views, while the full view screenshot was retrieved on February 15.

Figure 1. Layout and ranking of lookzones (content and ad zones). This image is the full screenshot of the website with blue lookzones drawn over key areas of interest. The yellow stars or circles refer to the ranking of the advertisement and content lookzones respectively.  The zones are ranked with number one having the highest average viewing time and number eleven having the lowest.  

Figure 2. Average time spent in lookzones. This table displays the average time that participants were engaged in viewing each lookzone.  The viewing time was measured in seconds.  R1A* contains lookzones R11A and R12A within. The highest average time spent in any lookzone was for C3C, which is the central content lookzone with media clips and headlines.  

Figure 3. Average time spent in combined and split versions of R1A lookzone. The R1A advertisement zone is often split into two different panels, which are referred to as R11A and R12A lookzones. Users spent more time in the same area when the R1A lookzone was split (0.44 second) than when it was combined (0.11 second).

Results:  

We primarily measured the average amount of time each participant spent across different lookzones.  These values are displayed in Figure 2. Users spent more time on the content zones than the ad zones in general.  R12A, L4A and C1A are the top three ad zones based on this metric.  Amongst the content zones, C3C, C4C, and L1C ranked higher than others did.

The C1A ad zone was situated at a “prime location” (center-top), which is often posited as one of the most lucrative spots for banner ads.  It is interesting to note that users did not look at this ad zone for the longest period.  This can be attributed to a behavior called banner blindness[1].  It is commonly believed that users gradually get trained to overlook banner advertisements in common locations, such as the top-center ad zone of C1A.

The R12A and L4A ad zones are located beside the content zones that received the highest attention.  It can be argued that some of the attention was directed towards the adjacent ad panels due to its placement.  The bigger and more visually rich advertisements, such as L4A, received more attention than smaller, text-only ads such as L3A. Zone R11A follows as the third position despite its small area.  The other ad zone, L5A, may receive less attention in a live website than our current study due to the common tendency of users to not scroll below the fold.

The content zones received different degrees of attention based on their location, content, and size.  The largest amount of attention was directed towards C3C and C4C.  The users were curious about the St. Joseph Channel logo (L1C) and the weather box (L2C).  The ad zone (L3A) failed to draw much attention, although close proximity to the larger content zones seem to have increased viewing times for other ad zones.

Recommendations:

Although current study found some differences in viewing times across ad panels, it was not significant enough to establish a differential pricing-model solely based on attention.  However, higher pricing can be justified per unit area when an advertisement panel is split into multiple parts.  This follows from the observed trend that users spend more time when there are more ad panels in the same area (Figure 3). Again, one of the split ad zones- R11A received the highest amount of attention per unit area amongst all lookzones.  This observation may also justify any premium pricing for this otherwise small-sized spot.

The viewing time of the ad zones were consistently less than that of content zones.  This emphasis on content should be maintained to ensure an acceptable level of user experience.  The visual treatment of L3A lookzone may need minor enhancement so that users who make use of the weather content do not miss the sponsor’s name.


[1] Pagendarm, M., & Schaumburg, H. (2006). Why are users banner-blind? The impact of navigation style on the perception of web banners. Journal of Digital Information, 2(1).