Spending on digital video in Europe has increased by 42% over the past year, with agencies shifting budget from other digital channels to support their video activity.
The survey, conducted by Adap.tv among 175 ad buyers and publishers, found that 48% of agencies have shifted adspend from display, 33% from print, and 10% from search.
However there were still concerns around the ability to properly measure how many people are exposed to video ads online.
Ad viewability emerged as the top concern for more than two-thirds (67%) of European agencies, while ad verification and ad fraud were cited by 57%.
These concerns are fuelled by the fact that the industry is still struggling to come up with an accepted standard for video viewability measurement.
While this problem persists advertisers will be paying to serve ads that will never be viewable to users.
To shed some light on this issue, Kellogg’s and BrightRoll ran a study that evaluated how different companies tracked the viewability of different ad units.
For more on this topic, read our guide to viewability in online advertising and this post on why measures of digital marketing ROI are meaningless unless ads are viewable.
Each viewability measurement company (VMC) was asked to provide the following measurement for every video impression:
The VMCs involved in the study were narrowed down to those that offer solutions based on the IAB VPAID standard, are compatible with programmatic buying, and were independent measurement company with no media buying capabilities.
Kellogg’s then ran 40 different tests that varied based on different scrolling behaviours, browser type and the presence of iframes.
This table shows the accuracy with which each of the VMCs was able to measure the viewability of the different ad scenarios.
Some clearly performed better than others, but none of them achieved 100% accuracy across the board.
The results show that the consistency associated with display viewability measurement does not hold true in video viewability measurement.
The accuracy of various measurement partners varied primarily based on user behaviour (whether the user scrolled during playback) and based on browser/iframe combinations.