|
SUMMARY
DEFINITIONS
INTRODUCTION
METHODS
RESULTS
ADDENDUM
Eyetools has helped clients raise conversion rates by 42% - 2000% -- call us to discuss your needs, or check out our blog for other interesting eyetracking.
916.792.4538
Greg, CEO & Founder
Eyetools, Inc. |
 |
Results:
Right Hand Rail: Initial viewing of search results page
When a user first views a search result listings page, the probability of a link being seen is inversely related to its distance from the first page listing. The farther a listing is located down the page, the lower its probability of been seen by users. Previous reports have discussed this effect in great detail. In this study we observed that 62% of participants looked to the right AdWords area while conducting their product searches. The ad in AdWords position 1 and 2 had a 47% probability of being seen by users. Ads in positions 3, 4, and 5 were seen by 30-40% of viewers. Ad 6 was only seen by 21% of all users.
Table 1 – Number of participants who viewed each area of interest along the right rail. Table 1 shows the number of participants who looked to the right rail area during their test session. A steady decrease in viewing of ads can be observed for ad positions 3-6.
Although we were not surprised to find that link position along the right rail to have a significant effect on ad viewing (p= .0022), we were surprised that the presence of the Google Checkout icon did not alter this viewing pattern significantly (Figures 2A-D) (p=0.496).
   
Figure 2 – Full First View Heatmaps. Above are full aggregate heatmaps for first time presentations of search results (N=100). The Google Checkout icon was either not displayed (A), or shown at AdWords position 1 (B), position 3 (C), or position 5 (D).
Figure 3 shows what percentage of right rail viewers saw each ad position. For example, when no button was shown, 61 users looked to the right rail. Forty-seven of those viewers fixated the area of the top most advertisement (Ad 1). The result can then be read as “77% of people who looked to the right saw Ad 1”.
Figure 3 – AdWords viewing on first presentation of search results page. The plot shows the percentage of right hand rail viewers who saw each AdWords area (ads 1 through 6). Although the presence of a button does appear to result in a small rise in viewing for the parent ad, the change was not significant.
Similarly we found no change in number of fixations per ad or in looking time due to button treatment. The button itself received very little focused visual attention from users. Few users fixated the button directly. Direct viewing of the button itself appears to be tied to ad position. None of the 100 participants clicked directly on any Google Checkout button in our test.
These results indicate that the Google Checkout icon has little to no effect on increasing AdWords viewing. If the icon does alter common viewing patterns, a much larger sample would be needed to reliably pinpoint such a small effect. This does not, however, mean that the Google Checkout icon is uninformative. Right hand rail viewers do see the icon, and may make a decision to click on one link or another based on the availability of this service. We did not see any qualitative difference in click-throughs to sites with the button and without, but the sample size used in this study is too small to detect small or medium sized click through effects due to button placement.
If a large increase in click-through rates is observed by icon displaying advertisers, the higher click-through rates are not likely due to an increase in the number of users viewing the ad. Other possible reasons for increased click-through rates may include (1) novelty (if the increase was only seen at the time that Google Checkout was introduced), (2) Google brand/product confidence, or (3) increased frequency of the advertisement presented in the top 2 listings.
The timing of this study, 1 year after the launch of Google Checkout, means that most internet users have probably become accustomed to viewing Google Checkout icons as a part of AdWords listings. The Google Checkout icon does not tell them about the ad’s relevance to their search and therefore is not directly fixated or read. Users may then be using the icon as a source of information about the store, just as they may use the site URL text.
Right Hand Rail: Return views of search results page
Much attention in previous research has been given to the first presentation of a results listing. However, in our study we found that 40% of searches involved at least one return visit to the original search listings page. Heatmaps for all button presentation positions are shown in Figure 4A-D. Our results confirm that right hand rail ad position does not play a significant role in determining if an ad is seen on these return page visits (p = 0.288).
   
Figure 4 – Full Return View Heatmaps. Above are full aggregate heatmaps for return visit presentations of search results. (N= 60, 58, 68, 59 for figures A-D, respectively) The Google Checkout icon was either not displayed (A), or shown at AdWords position 1 (B), position 3 (C), or position 5 (D).
Our results did not show a statistically significant difference in viewing on these return visits when the icon was located in the lower ad positions 3 and 5 (p=0.847). However, a slight increase in viewing at ad position 3 was observed when the icon was present (Figure 5). Further research would be needed to verify whether or not this observed increase in is due to icon placement or user variability. If the increase is the result of icon placement at position 3, the effect is likely to be extremely small.
Figure 5 – AdWords viewing on return visits to search results page. The plot shows the percentage of right hand rail viewers who saw each AdWords area (ads 1 through 6). Here ads in the 3rd position which displayed a Checkout icon received a marked viewing increase. However, this difference was not statistically significant. If the effect is real, it is very small and a much larger sample size would be needed to identify it.
Top Sponsored Advertisements
The secondary goal of this study was to look for qualitative changes in viewing of top sponsored ads due to the recent background color change from blue to yellow. Raw data from previous studies was not available to us, and so we used heatmaps provided in the 2005 Enquiro-Eyetools product search study.
Our results show that 100% of participants viewed link titles for the 2 sponsored links contained in yellow background. Previously studies have reported a marked decrease in viewing of the top most non-organic link. However, those studies did not control for sponsored-link content as we did. When comparing our current heatmaps with other reports containing only 2 text-only non-organic links, we did not find a notable difference.
Although we did not see a large increase in the already heavily viewed top level advertisements, Google users may have increased their willingness to interact with elements located above the organic text listings in the past two years. The common inclusion of “integrated results” such as images or maps above organic text listings may encourage users to venture higher onto the page than before. Google may have also improved the accuracy of its search algorithm such that the sponsored link area may be perceived to return results nearly as relevant as the organic list during product searches.
Summary
We were surprised to find that the presence of an icon in right hand rail listings did not produce a measurable change in viewing patterns. This implies that a number of other factors (including pre-existing individual search habits, vertical page position, and right hand page position) currently play a much stronger role in ad visibility. This study was only able to reliably examine user viewing patterns. We were not able to measure small increases in click-through rates by users who prefer stores which provide a Google Checkout option.
|
|