After introducing non-skippable ads and finally bumper ads in 2016, Now YouTube is alpha testing a new tool which is based on machine learning to automatically pull out a six-second video version from a langer video ad.
To an extent, it seems a little ridiculous and also beneficial for others.
Debbie Weinstein, Google’s vice president of YouTube and global video solutions, acknowledged that there was some skepticism when Bumper ad type was announced, advertisers remained wondered, ” How can we tell our story in merely six seconds?”
However, Weinstein argued, “We have learned over time that creative love constraint. They’ve historically been constrained to 30 seconds, and then 15 seconds, and constrained by whatever dimensions of a particular media format.”
She added a Bumper ad type could be a short teaser for a longer ad. For others, the format could allow a way to cut down a 30-second length video ad into a sequence of six-seconds smaller ads
Right now, this Bumper Machine is in alpha testing and sooner will lead to beta testing and finally will be available for advertisers, which will allow advertisers to create a Bumper ad by scanning a longer ad for “key elements,” like a tight focus on branding, products, human beings, and voice-over. The result always ends with “the final call to action in the last two to three seconds of the video ads,” Weinstein Said.
Here for an example, as an initial test, GrubHub took a 13-second ad and used Bumper Machine to create the six-second ad type below.
six-second ad type here below
Here Weinstein suggested that Bumper Machine will be beneficial for different advertisers of all shapes and sizes – some small advertisers who want to create six-second Bumpers ads could create them with little time and effort as possible, while big giants or brands and agencies can treat them as more of a “jumping off point,” which can be refined as inspiration.
Weinstein doesn’t expect that advertisers to start posting machine-created Bumpers willy-nilly. The idea behind it is always had some level of human review.
Here the advertisers will get three to four executions, and a human advertiser will choose the best guesses that machine is going to make on the bases of his wisdom. And decide which of the three or four is best, or decide all of them are good or do some light editing required on top of that.