Facebook has been a big driver in online advertising innovation. It has pushed native ads to the forefront, and it has also driven sound-off autoplay video to be more generally accepted by advertisers. On this latter point, however, they've recently stumbled.
Linear TV is declining - cord cutters are increasing and consumer attention is shifting to digital. Yet digital has never managed to get a cut of that sweet linear TV budget. While the total digital ad spend approaches $100B, TV is pushing $200B. Naturally, various digital platforms have made the play that theirs is the format that provides the most analogous experience to linear TV advertising, and those declining linear TV budgets should go to them. Facebook and Google (especially YouTube) has been at the front of this conversation.
Facebook was among the first to push the boundaries by showing volume-off autoplay ads and trying to have them count as video views. Advertisers grumbled behind closed doors, but also spent aggressively on Facebook's video ads, comforted ever-so-slightly by Facebook's very aggressive position that they would only count a video ad as a view when it was anywhere on the screen playing for at least 3 seconds. Naturally, 3 seconds of autoplay that people aren't looking at is not the same as a pre-roll ad or, for that matter, a linear TV ad. But such is the power of Facebook that it can get advertisers to play its game.
With Facebook, there are now a few eligible metrics - video ads that simply appeared, video ads that hit the three second threshold, and video ads that hit the advertisers' target (likely completion). Facebook recently came under fire for its Average Duration of Video Viewed metric. One might expect that this is total amount of video ad viewed / total number of times it was shown. Facebook was instead representing Average Duration of Video Viewed as total amount of video ad viewed for users that saw at least 3 seconds / total number of times a video ad was shown for over 3 seconds. Naturally this shows a significantly inflated number - it doesn't include any of the times that a user had it on their screen for less than 3 seconds. This did not impact actual billing, because advertisers were only billed for those ads available for over 3 seconds. But it did impact how an advertiser considers their media allocation among channels - especially between YouTube, which lets users opt out of the ads at various points, and Facebook, which lets users scroll past the ads.
One of the main concerns about Google and Facebook is that they don't play nicely with third party attribution. Whereas TripleLift is subject to oversight from complex multi-touch attribution systems and are thus accountable to its buyers, Google and Facebook attempt wherever possible to use their market heft to avoid such accountability. They will continue to have market power, and likely will continue to be able to play by whatever rules they want. But the more walled gardens make mistakes like this, the more large advertisers will demand third party attribution. It remains to be seen how many mistakes like this happen, and how much market power large advertisers really have - but it will an interesting few years.