Posted on January 5, 2009
Attention span for Web video
Peter at Video 2 Zero wrote a provocative post, with data — Ideal length for web video — a couple of weeks ago. The gist is, fewer people stick with an online video to the end if the video is longer. Longer than what, you ask? Just longer. In general, people open a video, start watching, and then begin dropping off.
This shouldn’t surprise anyone. It’s the same way you act when you open any Web page — you look or scan or read until you feel like you’ve had enough. Then you leave.
I don’t think this is any different from the way people read printed newspapers. You start, and eventually your attention wanes, and then you quit.
Now, before you start arguing — yes, sometimes I have searched for exactly that information (that video, that Web page, that story, etc.), and then I am likely to stick with it to the very end. So your fascinating video, or text story (about the cause of a plane crash, or the rescue of some avalanche survivors, or the future of our 401(k)’s) might be long, but I am already committed to it, and I am predisposed to watch (or read) the whole thing.
Most things that we encounter online do not meet that criterion. Most things we encounter online are not the exact thing we have been searching for.
So we sample. We take a little taste and move on to the next thing.
A 17-year-old recently told me that sometimes she starts looking at YouTube and somehow, before she knows it, three hours have gone by. “Really? Three hours?” I asked her. “Seriously,” she replied. But you know darned well she did not watch every video she opened to the end. YouTube is a place for sampling, tasting, surfing. It makes it really easy (and irresistible) to have another, and another, because of that “Related” box on the side.
I take two lessons from this:
- A shorter, extremely tight and fascinating video has more chance of being watched; it is more likely to succeed in communicating its message, because more people would finish watching it.
- News sites need to put more effort into facilitation of this sampling behavior, and production of long items (videos or text) is counterproductive to that effort.
Don’t take that to mean I want to dumb things down. I do not. But the array of information has to become smarter (like YouTube’s arrays), and the complex (long) pieces must be reconfigured into smaller pieces, to adapt to the sampling behavior of the audience.