Classic leaderboard (728×90) display ads are 7.2% more likely to be served in-view than medium rectangle (300×250) ads (74% vs. 69%), and 12.1% more likely to be delivered than wide skyscraper (160×600) ads (74% vs. 66%), details comScore [download page] in a March 2012 study of 12 national premium brands. Although the classic leaderboard delivered the best in-view rates, there was a large amount of variance across the sites measured, ranging from 7-93% using the size. comScore insight suggests that a possible cause of the variance among ad sizes is the typical placement of these ads on a webpage.
The definition of in-view used by the study was that at least 50% of the pixels of the ads must be in-view for a minimum of 1 second. The average in-view rate of the campaigns was 69%. A February report from AdSafe also found a significant proportion of display ads not being in-view to their intended user: according to that report, more than 38% of display ads in December 2011 were either not in-view at all to their intended user or in-view for less than half a second.
The comScore study also found significant variation when breaking down the data by content types. Coupon sites delivered the best in-view rates (89%), followed by directories (81%), entertainment – radio (73%), entertainment – TV (72%), and technology – news (70%). By contrast, just 27% of pets site ads were served in-view, easily the lowest of all content types. comScore points to common layouts among sites of a similar genre as a potential reason for the variance.
Meanwhile, large sites within a content category appeared to fare better than their smaller counterparts in delivering viewable ads. Using comScore Media Metrix rankings within specific content categories as a proxy for site size, the study revealed that in-view rates rose alongside higher site rankings. Top 50 sites delivered 77% of ads in-view, compared to 74% for top 100 sites, 70% for top 500 sites, and 61% for the remaining sites.
The study looked at the effect of page position on in-view rates, with some intriguing results. The percentage of above-the-fold ads in-view ranged from 48-96%, while for those below-the-fold, the range was 3-67%. comScore insight indicates that some above-the-fold ads went unseen as users quickly scrolled past them before they loaded, while many placed below-the-fold delivered high in-view rates.
Of note, when the study took a site-level view across campaigns, it found one site with a 100% in-view rate. That particular site had all ads placed in the center of the homepage, and scrolling was not required to reach the remaining content.
Examining cost data provided by 8 of the campaigns, which totaled 300 unique ad placements with accompany CPM data, the analysis discovered that there was almost no correlation between the CPM paid for the ads and whether it was in-view, with a correlation coefficient of just 0.19. This indicates that marketers must understand the actual delivery of ads by both site and placement in order to value media based on the ability to reach a real user.
About the Data: The comScore US-based validated Campaign Essentials (vCE) Charter Study involved 12 national brands, 3,000 placements, 381,000 site domains and 1.7 billion ad impressions. Select advertisers from the charter study include Allstate, Chrysler, Discover, E*TRADE Financial, Ford, General Mills, Kellogg’s, Kimberly Clark, Kraft, and Sprint. All of the impressions analyzed in the study were delivered in iframes and none required publisher site pixels. The study notes that because vCE Charter Study participants included major branded advertisers, who inherently buy more premium inventory than the average online marketer, the study findings are not necessarily representative of the overall online advertising market. In fact, because these advertisers generally engage in high- end, premium campaigns, the findings may represent “best-case scenarios,” rather than the norm.
Subscribe now to receive more charts and articles like this in your inbox. A fast read in a clean, mobile-friendly design.