On this page
Finishing a video project is something to celebrate but it doesn’t always mean your work is done. There are simple steps you can take post-production to significantly improve the performance of your videos. Our pick of the bunch? A/B video testing.
What is A/B video testing?
A/B testing or split testing is the simple technique of testing two variants of the same thing (like a video ad) to see what performs better. In an A/B video test, the variant could be as simple as different CTA copy or as complex as an entirely different thumbnail – it’s up to you.
What’s great about A/B testing is it allows creators to make informed decisions about their videos – data doesn’t lie. With concrete results, you can rely more on evidence and less on chance, and ultimately improve your video marketing strategy.
How to A/B test videos
Here’s a list of things to test in your own A/B video testing experiments.
Title
What’s in a name? Quite a lot when it comes to video. The title is the one of the first things viewers see when searching for something to watch. Names affect where your video appears in search results and significantly influence your click through rate, so it’s an important element to test.
Thumbnail
They might sound small, but thumbnails are another element that has a big impact on your video’s CTR. These images set viewer expectations and entice them to watch. A/B testing an image this impactful is an excellent idea, so much so that Netflix has made it a key part of their strategy. Try changing up your thumbnail with different images and fonts to see what resonates best with potential viewers.
CTA
Test variants of your "call to action" to see which one triggers more of the behaviour you're after. It’s incredibly easy to change one or two words in a CTA to improve engagement with your video. The difference between “Shop Now” and “Explore More” could be as much as 90%. Test out a variety of CTAs within your video or on your own website if your video is posted there. Max Benz from HörbuchHeld can confirm this: "In our tests, we found that people definitely underestimate the CTA in a video (or even on a website). Even small changes can have a big impact. In one test, we were able to increase our conversion rate by 67% by adapting the CTA even better to the target group."
Length
How long should my video be? It’s a question all marketers and creators ask themselves when putting their content together. Depending on where you post your video – YouTube, Facebook or Instagram – the ideal length varies. To be 100% sure what works best for your video, test different lengths – the results might surprise you.
The first 10 seconds
10 seconds – apparently that’s all the time it takes for a viewer to decide whether they want to keep watching your video. It makes sense then, to test multiple video intros to see what captivates viewers. Make sure you do this armed with the knowledge of what makes a great video intro.
A/B video testing tools
There are a variety of platforms ready to enable your A/B video test that fits your requirements.
YouTube
YouTube doesn’t currently offer any built-in A/B testing tools, luckily there are programs like TubeBuddy. You can install it for free on Chrome to start testing and access all of the data you need to learn what works for your channel.
Facebook enables you to perform all kinds of A/B tests on video ads posted to their platform. You can choose to create an A/B test using the Ads Manager Toolbar or set up your own from scratch by duplicating an existing campaign and changing one or more variables.
Google Analytics
If you’re posting your video on your website, Google Analytics can be a great tool for A/B testing. Google Analytics can track the performance of two or more variants of the same webpage. If the sole variant you change on your webpage is your video, there’s your A/B video test!
A/B testing example
When it comes to examples, this is a big one. During the 2008 US presidential election, the Obama campaign raised $60 million thanks to a successful multi-variant test. This was more of an A/B/C/D test than an A/B test, and it was a landing page, not a video but you get the point.
This example uses a few of the tests we’ve mentioned above – title, image and CTA. In the end, it turned out audiences responded best to the title “Change we can believe in”, overlaid on a black and white image of the Obama family with a “Learn more” sign up CTA.
The original Obama campaign
The Obama campaign after multi-variant testing
Bonus Tip [2022]: A/B Video Testing With Real People
And now you can A/B test your videos even before showing them to your real audience. Similar to how movie studios test their ads and clips on focus groups, you can A/B test your videos with real people using a live on-demand testing platform. This lets you get valuable user feedback within an hour, without the hassle of focus groups or live testing.
