Over the last few months, observers of American media have watched several prominent media outlets “shift to video” (and, accordingly, fire writers) embracing video content that is popular with advertisers, if no one else.

Some consumers of online content (not to mention creators of it) have been confused by the trend, given how many people seem to consider videos on the top of articles to be somewhat burdensome. The idea that those pesky autoplay clips at the top of articles are the future of media has been tough to swallow, as has the fact that Fox Sports, as a notable example, thinks going all-in on video of Colin Cowherd and Jason Whitlock is its wisest course.

Well a new report from video marketing company TwentyThree.net has leant some insight into which kinds of video are successful and which are not. There are two big takeaways, as reported by AdWeek: 1. Longer video gets more engagement than shorter video. 2. Video gets more engagement when it’s on an outlet’s website as opposed to on Facebook or Twitter.

The statistics about long vs. short video are pretty jarring. Per AdWeek:

The “State of Online Video in 2017” report found that videos can, and should, last longer that 90 seconds if publishers want to see higher engagement rates. While 80 percent of videos are under five minutes, they drive less than a third of overall video engagement. Mid-form and long-form videos, which are at least 15 minutes long, drive over half of all video engagement despite encompassing just 8 percent of all video.

“If you make longer videos, there’s more time to engage with them,” said Steffen Christensen, chief technology officer for TwentyThree. “People are worried about what amount of impressions they can get, but tracking video success needs to be different.”

While media outlets seem to have flocked to shorter videos in pursuit of as many views as possible, in doing so they’re sacrificing the engagement that can be much more valuable to advertisers. Maybe you “view,” the autoplay video that comes along with many articles on Sports Illustrated or NBC Sports, but you’re probably not focusing on it. This report suggests it might be worthwhile for companies to invest in more thoughtful, pre-planned video as opposed to throwing together quick-hit clips, for the sake of keeping viewers from pressing pause.

Then there’s the question of social media vs. on-site video. Per AdWeek:

The study also found the difference in video plays per platform, though TwentyThree intends to do further research. On average, people only watch 20 seconds of a video they clicked to play on Facebook; if a video is hosted on a company’s site, people are likely to watch the video for at least four minutes.

It makes sense that video would get better engagement on an outlet’s site, where someone is more likely to have sought it out, not just come across it. Maybe this finding lends some credence to Fox’s all-video digital strategy, though it’s worth noting that most of the video on FoxSports.com is the type of quick, re-packaged content that reportedly gets less engagement.

Clearly, there’s still a lot we don’t know about what kind of video drives what kind of viewership. More importantly, there’s a lot advertisers don’t know about what kind of video drives what kind of viewership, which means incentives could continue to change for outlets as time goes on. We seem to be in a “shift to video” era—now we’ll find out what video we’re actually shifting to.

[AdWeek]

About Alex Putterman

Alex is a writer and editor for The Comeback and Awful Announcing. He has written for The Atlantic, VICE Sports, MLB.com, SI.com and more. He is a proud alum of Northwestern University and The Daily Northwestern. You can find him on Twitter @AlexPutterman.