If you’re not sure whether your video should be long or short, which thumbnail, or what call to action to use, why not setup a controlled experiment to see which one works best?

By setting up different versions of a landing page, each with a different version of your video embedded, you can then compare the results. You might find that one version has a better retention rate or is better for conversions. You never know until you try!

Here are 5 ways to use A/B testing to boost the performance of your videos.

5 Ways to Use A/B Split Testing to Boost Your Video’s Performance

1. Thumbnail Image

Video hosting platforms will automatically generate a thumbnail for you, but that image won’t necessarily be relevant, so you’re much better off designing and uploading your own.

Will an image of a person work better than a close-up on your product? How about text in your thumbnail explaining what the video is about? Thumbnails act as a poster for your video and can make a huge difference in play rates, making them the best place to start with A/B testing for your videos.

2. Testing Your Call-to-Action

Believe it or not, the difference between “Download a free ebook” and “Download your free ebook” can result in tangible differences in click through rates. The timing of a call-to-action within a video is also an important consideration, as it won’t be seen if it’s too late in the video.

This one may require a re-edit of your video, unless you’re using a professional hosting solution like Wistia which allows for custom CTA’s during the embedding process.

3. Testing for Video Length

It’s tempting to simply say that “shorter is better”, but this isn’t always the case. In fact, videos that are longer than 2-3 minutes are shared more often, possibly because it takes time for your audience to establish a strong enough emotional connection for them to forward the link to other people. The Kony 2012 documentary is a good example of this.

Studies have shown that average viewer retention rates tend to be roughly the same for online videos around 3-4 minutes in length as 5-10 minutes, but that viewers decide within the first 30 seconds whether they will continue to watch the rest of the video – a clear indication that expectations matter.

So how long should your video be? The answer is probably something like “long enough, and no longer”. But one way to know for sure may be to make two versions of the same video and put them to the test.

4. Testing Placement and Size on Page

You’ve invested a lot in this video, so let’s make it the focus so that it can do its job, placing it ‘above the fold’ and free from other links or distractions.

Then again, the video may require some setup with the use of copy and a strong call to action (eg. “Watch the video below to learn more..”) in order for visitors to hit that play button.

5. Testing With and Without a Video

By setting up a ‘before and after’ test to compare the same landing page both with and without a video, you might actually find that the absence of video is better for conversions – although this tends to be rare.

Being able to quantify any increase in metrics such as conversion rates can help you accurately measure ROI for your video marketing efforts, hopefully allowing you to justify investing more in video in the future.

Aren’t video metrics fun? Learn more about measurement and optimization in our video production Knowledge Centre.

Using A/B Split Testing to Boost Your Video’s Performance