61% of directly placed ads stayed in-view for at least 1 second during the fist half (H1) of 2013, up from 56% in the second half of 2012 and 49.9% in H1 2012, according to [pdf] a new study from Integral Ad Science. Viewability for directly placed ads was better than for networks (43%, slightly up from 42.2%), exchanges (40%, up from 37.6%), and hybrids (40%, down from 41.1%). Longer engagement with viewers was much harder to come by: just 31% of directly placed ads remained in-view for 15 seconds, although that was a significant rise from 22.7% in the previous 6-month period.
Details from Integral’s “Semiannual Review” indicate that viewability continues to be better for vertically oriented (160Ã—600 ”“ Skyscraper) ads, with an average of 67% in-view for at least 1 second.
Suspicious Activity Lowest for Directly-Placed Ads
Directly placed ads not only had the highest engagement, they also were deemed to be lowest risk. While more than 1 in 5 impressions overall were suspicious of being fraudulent activity, only 2% of those placed directly exhibited suspicious activity. By contrast, exchanges were far riskier, with 20% of impressions deemed questionable (though that was down from upwards of 30% in H2 2012).
While the Integral study suggest some improvements in fraudulent activity, a recent report from Solve Media indicated that levels of suspicious activity continue to rise. Solve Media also pointed to rising fraud within the video ad space: in its study, Integral reveals that 3% of impressions on pre-roll ads were suspicious, with suspicious activity more concentrated on in-banner video ads. In fact, 40% of impressions through exchanges on in-banner video ads were estimated to be suspicious, according to the report.
- Food sites boasted the highest level of engagement, with an average in-view time of more than 2.1 seconds. Education sites fared worst on this level, with an average in-view time of about 1 second.
- Ads on shopping sites were again deemed the least risky in terms of suspicious activity, while family and health sites had the highest rates of suspicious activity.
- The overall proportion of high-risk inventory (impressions that represent a low degree of brand safety) stood at roughly 6% in Q2, and was relatively consistent across channels.
- Risk content continues to be mostly the realm of illegal downloads, drugs, offensive language and alcohol. Risky adult and hate speech inventory declined between Q1 and Q2.
- About 13% of ads collided with another ad from the same campaign during H1, down from more than 20% in H2 2012.
- Canada received the largest amount of the US’ non-geo-targeted content, at 16.4% in Q1 2013 and 15.8% in Q2. Both were improvements from 34.5% in Q4 2012.