Humans have short attention spans. And thanks to the great distraction known as the Internet, it’s getting harder than ever to focus on one thing at a time.
Attention span has a pretty direct impact on content and video marketing. SumoMe looked at 650,000 sessions and found that only 20% of people actually read articles start to finish. (Yeesh—we’ll be keeping this post short to avoid losing 80% of you.) We assumed there’s a similar drop-off in video engagement, so we dove into our customer data to find out how long videos should be, when you should go long, and if there are any situations where seconds matter.
We confirmed our theory that humans have really short attention spans, but we found some other useful takeaways too.
Because every video is a different length, we decided to explore length vs. engagement. Do people take length into consideration when they decide to watch or leave? We wanted to find out.
We expected engagement to decrease with video length. In other words, we would expect people to watch more of a 1-minute video than a 2-minute video. But we also imagined that increasing a video from 1 minute to 2 minutes—a difference of 1 minute, but 100% longer—would have a much bigger impact on engagement than the jump from 10 minutes to 11 minutes—a change of one minute, but just 10%.
In mathspeak, we expected that the relationship wasn’t linear. We predicted we’d see an exponential decay like the one below:
Luckily, we have plenty of data to draw from—we looked at 564,710 videos, and more than 1.3 billion plays for this research, so we feel comfortable that the results are representative of wider video engagement trends. So what did this data actually tell us?
Drum roll please...
2 minutes is the sweet spot. After that, the drop-off in engagement is significant. Here are a few important points from this data.
Engagement is steady up to 2 minutes, meaning that a 90-second video will hold a viewer’s attention as much as a 30-second video. This is surprising and actionable information for video marketers. If you’re making short videos, you don’t need to stress about the difference of a few seconds. Just keep it under 2 minutes.
Takeaway: Short and sweet is a safe strategy. Assume that your viewers are busy. But don’t sweat over a few seconds here and there if you’re still going to come in under the 2-minute mark.
After 2 minutes, every second counts.
This is where the exponential decay kicks in. This isn’t to say that you should cut a 10-minute video down to 2 minutes—some content warrants longer videos—but it does suggest that you’d be better off cutting 30 seconds from a 2-minute 30-second video to keep viewers’ attention.
Takeaway: If you’re new to video, start with short videos under 2 minutes. This will help you get a feel for what resonates with your audience before you invest in longer content.
The engagement decay really starts to level off after the 6-minute mark. Every second counts between 2 minutes and 6 minutes, but there’s hardly any drop-off between 6 minutes and 12 minutes. If your video ends up being 8 minutes instead of 7 minutes, it shouldn’t have a significant effect on engagement.
Takeaway: Beyond 2 minutes, focus entirely on the content. If the viewer would benefit from brevity, go short. If they’d benefit from depth, don’t cut just for the sake of cutting.
Just like writing, a video needs to be as long as it needs to be. Videos beyond 12 minutes are likely either narrative-driven stories or intense tutorials. In both cases, the viewers expect the video to be longer.
Still, every minute beyond 12 means you are losing people. And while the drop-off from 12 minutes to twenty minutes is less steep than the drop-off from 2 minutes to 6 minutes, it’s important because longer videos are often more expensive and time-consuming to make.
Takeaway: Focus on making the video great, and the engagement will follow. You may also want to give viewers the option to a link emailed to them to watch later.
It’s safe to assume that your viewers have short attention spans and plenty of distractions. The ideal length depends on the content, the context, and the viewer. So study your engagement graphs and keep it interesting!