Growth Marketing

Are You Giving Your Videos A Chance To Perform?


You’ve probably heard enough from us about why you should always be testing your copy, images, page layouts, and site functionality. “Yes, Joanna & Lance, we get it. We should always be testing. What else do you have for us today?

Actually – despite your skepticism 🙂 – we have a test result that you’ll want to understand, especially if you embed videos on your site.

With all the video hosting options available — YouTube, Vimeo, Wistia — embedded website video has exploded. “Explainer” videos, product demos, and video blog posts are everywhere. It’s more likely that you’ll encounter a website that uses video than not.

The challenge with video is that it’s not easy to optimize. It’s not like changing copy on your website. With video, you’ve got to shoot the whole thing over or have a decent command of Camtasia.

With the time and dollar investment that typically goes into developing a website video, you want people to hit the play button. Once your visitors are watching, they’re hopefully receiving your message, becoming more interested and engaged, and moving one step closer to converting.

With so much riding on the video, it’s critical for you to choose the right frame to show your visitors. That single frame can make a huge difference to visitor behavior.

We’ve been working with a cool company that sells personalized nutrition and fitness coaching programs in a hybrid delivery model. They use a combination of software-as-a-service and 1-on-1 coaching to deliver the program to participants – to ensure that people are successful.

The company has managed to achieve a decent clickthrough rate to their men- and women-focused landing pages with their copy and testimonial video – but they wanted to optimize the home page and direct even more traffic to the landing pages.

So in a recent (and fun!) 10-way split test, we modified headlines, subheads – and, you guessed it, the display frame for their testimonial video.

Here’s the relevant section of the home page hero section we tested (default):


And here’s the challenger (Variation #9), with revised display frame:


As you can see, we isolated the change to a single variable.

Our goal was simple… to generate more clicks on those two blue buttons.

Any guesses as to the outcome? Is it obvious? Perhaps, but the magnitude of the results surprised us (and we’ve been doing this awhile).

The challenger version resulted in a:

1. 64% increase in clickthrough to the “I’m a Man” CTA (99.8% confidence)

2. 109% increase in clickthrough to the “I’m a Woman” CTA (100% confidence)

3. 48% cumulative increase in opt-ins to a free 5-day video course from the 2 gender-focused landing pages

The 3rd outcome was purely the result of sending more visitors to each of the landing pages. We changed nothing on the landing pages themselves.

To what would we attribute the gains?

  • The default video display frame does not evoke an emotional response
  • The default video copy competes with the “other” headline
  • The challenger video skips the “tell” and goes straight for the “show”

What do you think? Tell us below in the comments…

So what can our client do with this test learning?

Changing the display frame on embedded videos is a trivial task. This type of change can be quickly applied (and tested) on other key pages. The results have also re-focused the team on the importance of showing real results whenever possible. It may sound like an obvious thing to do, but there are a lot of things for business owners and marketers to think about. 🙂

If you have embedded videos on your site, this is a test you can have some fun with!

About the author

Lance Jones

  • Rich McDonald

    Good idea….never thought of those splits [even with just the opening frame]!

  • robindotadams

    Good stuff… although I tend to be here at copy hacker for the “copy”, any hack is useful. As I’m involved in Video any stuff is useful on that subject. Like Alexandre I’m coming back to the ratio of clicks on video/click throughs.

    In fact… are the results of this test actually relevant as do you know whether the click thru’ could be attributed to watching the video or not watching the video.

    Without knowing if the video was watched before a click was made, you’ve no real idea of whether the final click thru’ was due to watching of the video or the thumbnail.

    I’m hoping/guessing (there’s an opportunity to test!) that a more involving image = more people watching the video = more click thru’s.

    Is the moral of the story to encourage people to watch the video with an engaging thumbnail which will increase final click thru’s? I’ve seen too many videos that don’t shout WATCH ME in the thumb nail. Isn’t that the point of the video???

    • Lance Jones

      Hey Robin, thanks for your questions! My client doesn’t really care whether someone watches the video or not. Beyond images of people, there isn’t a lot more to the video (perhaps an opportunity there!).

      The only thing they really care about is the fact that this new thumbnail produced a 48% increase in the number of opt-ins — which translates into 1000s of new subscribers per year. The increase was due purely to the change in thumbnail… and you’re absolutely correct, it may have resulted in more video plays… but in the end, it doesn’t really matter because the outcome (true conversion event: subscribes) was dramatically and positively impacted.

      Hope that helps!

      • robindotadams

        Lance, this is the type of fascinating discussion I could spend many an hour on!

        I agree that the results suggest: good thumbnail = more clicks, but that’s hopefully because the video is a good one. If it was a poor video but had a decent thumbnail then I’d suggest that good thumbnail = poor video = less clicks.

        Appreciate this comment is redundant until we get specific results in that area… but if a video sucks, even if it’s got a good thumbnail, it’ll have a negative impact on conversion. You guys are always going on about understanding customer flow and reducing leaks… a video will be part of that flow on a page.

        Ultimately what your research shows is that when you get a decent thumbnail – good things happen 🙂

      • Joanna Wiebe

        The fact is that 1 variable changed: the opening frame of the video. Nothing else changed. Good video, bad video, Coen Brothers-directed video… that’s not the point. The point is this: don’t overlook your opening frame.

  • In a related story, YouTube recently introduced a “Custom thumbnail” button for YouTube account holders. It allows you to upload your own thumbnail for a video instead of being forced to choose one of the 3 YouTube provides by default whenever you post one.

  • Lance Jones

    “Skin in the visual”… can I use that with my client, Aaron? 🙂

    Alas, we were not tracking video plays for this experiment, but we’ll make a point of it next time.

    Thank you for commenting!

    • Aaron Orendorff

      It’s yours.

  • How did you A/B test the thumbnails? Did you upload the video multiple times and embed different videos into each challenger?

    • Lance Jones

      Thank you for asking, Jason. We created a duplicate of the home page that incorporated the modified Vimeo video (which was also duplicated from the default video). In Optimizely, we then set up a redirect for that one variation, so 1/10 of the home page traffic was seamlessly transported to the duplicated home page.

  • Momoko Price

    You know, I wonder if the Challenger display screen was unintentionally communicating *social proof*, which contributed to the conversion jump.

    I realized when I looked at the Challenger display, the first thought that went through my mind was “all these people are doing it” (and look happy), whereas the default screen keeps the UX totally isolated.

    Hilariously, there’s no actual indication that the people in the display screen *are* product customers, you just naturally assume …


    • Lance Jones

      Great point, Momoko. With such a strong resulting lift, I think the new video frame probably encapsulates a number of human psychology principles — including social proof. I agree that one naturally assumes the photos are of actual customers… and in this case, that’s absolutely true!

  • Joanna Wiebe

    I found this test super-fascinating not only because of the results but also because I’ve never, ever seen anyone split-test the opening / freeze / display frame of an explainer video before. (Nice job, Lance!) *Everyone with video on their site should test their display frame.*

    • Lance Jones

      Haha, thank you, Joanna.

      You want a real mind blower? Testing background images on your key pages… I hope to have more on that topic soon!

  • Ramsay Leimenstoll

    Very interesting! My bet is with the “showing” instead of “telling” hypothesis – people are really drawn to human faces and scantily clad human bodies (even if these aren’t specifically sexy, alluring photos), and it’s a visual representation of what their clients want, *immediately* presented to them. The text version plays on curiosity, which is a good motivating human emotion, but desire+envy is even stronger.

    • Lance Jones

      Agreed, Ramsay! I believe this plays into the reptilian brain (versus emotional or rational parts of the brain). I really like how you put it. Just reading the words “desire + envy” makes you (the collective you) feel something.

  • bradfordshimp

    Excellent idea on testing different video thumbnails. While we think of video as hard to change, changing the thumbnail is easy! I’m wondering if the change is thumbnail increased the video views as well, which in turn led to more click-throughs.

    I think another quick video test would be to try out different sizes for the video.

    • Lance Jones

      You’ll also be disappointed in my response to Alexandre below, but it’s nice to see we have such a sharp Copy Hackers community. 🙂

      I really appreciate your idea to test the size of the video as well. It’s another straightforward test that can give us a quick learning… which in turn can be applied to other areas of the site.


  • That’s a very interesting variable to test, Lance.
    Thank you for sharing your results!

    By the way, were you able to measure whether video was played?

    That would be a good thing, because now I’m coming up with a load of questions…


    – Maybe the variation was less identified as a video, hence visitors were squeezed into clicking the buttons?
    – Maybe a video view increases CTR? Or maybe it decreases it…
    – Maybe video is never played and the frame then acts as an image, so maybe the variation performed better as an image?

    Any answer will help!

    • Billy Gene

      Alexandre that’s a great question!

    • Lance Jones

      I agree with Billy — great question, Alexandre!

      I am sorry to disappoint, but for this test we did not have Vimeo integrated with Optimizely nor did we have a Google event tied into Optimizely. Our goal was to explore how a change to the headline/sub-head/display frame impacts clickthroughs to landing pages that ultimately result in opt-ins.

      Your questions are awesome for a follow-up test — and we’ll plan to do just that!

      Thank you for jumping in to comment!

Copyhackers Tutorial Tuesdays training calendar