What you need to know about using video online


We received a lot of questions about testing and using video at our recent Web clinic on testimonials. Some examples:

  • How do I use video to help my landing page increase sales or subscriptions?
  • Which kind of videos work best? Short or long? Autoplay or Userplay?
  • Should I edit my videos or just leave them raw?
  • Are there specific bandwidth or file format issues that might hurt my conversion?

video.jpgBefore we answer, it’s worth noting where we are today with video as compared to just a year or two ago. Much like the ramping-up period for RSS feeds and corporate blogging, the examples have been around for a while but bottom-line numbers remained elusive.

For the past few years of testing here at MarketingExperiments, the results we’d seen from video had usually been underwhelming, mainly due to bandwidth. So our response to questions about video was often that “they generally hurt conversion.”

More recent test results are hinting that online video has grown up. Fast.

For example, SiteSell.com recently shared that adding video to their landing page increased conversion to around 20%. What’s interesting here, besides the increase, is that the video was used as core content stating the company’s value proposition, not just tucked away in a sidebar.

Another example is a recent Sherpa case study where a UK Entertainment Brand site embedded a 1MB video into an email and achieved a 50% increase in conversion. They also split tested the subject lines: the email that mentioned the video had a 14.6% higher open rate than the one that did not.

And in our own research, a redesigned page that featured a video testimonial yielded three times the click rate of the control (discussed in the testimonials clinic).

So while we’ve seen video reshaping the Web in terms of content for a while, we’re now seeing more numbers on the marketing side. And now that test results are justifying further exploration with video, and the audience is much larger than early adopters, our analysts are working to answer many of the same questions as you — including those four queries above.

However, like other aspects of marketing, there are few if any easy answers. The real answer is you have to test and retest.

Much like long copy vs. short copy, call-to-action button styles and colors, or form length, the answers will change with the context in which these elements are used. To find out if a longer video will outperform a short clip, you’ve got to test it and break down the variables involved: the content of the clips, the audience and its expectations from the page, the goals for the page and video, and so on. The same applies to bandwidth, editing, autoplay, etc.

Before you can answer the broader types of questions, consider the context of your videos and look at friction and usability; for instance:

  • What are you asking/expecting visitors to do with your video?
  • Do they need to watch all of it to get through your conversion process, or is it an add-on to complement a registration process?
  • Is it a testimonial, a how-to or a product demo?
  • What need is the video trying to fill, and is it in the proper place in the conversion process?
  • Does the video create friction with longer page load times, or by interrupting the site flow or eyepath?

The most reliable answers are specific to the usage and take the context into account. You’ll get those answers from tests that are set up and measured with the right research questions (see our recent clinic on testing).

Let us know if you’d like us to cover this topic in more detail in an upcoming clinic. And if you’re getting results from your own video tests, feel free to post about those here, too.

Austin McCraw, audio/video producer for MECLABS, contributed mightily to this blog post.

You might also like
  1. Rob says

    The use of any iconic device to increase conversion may be affected by geographical location.
    Not only is it the case that even now, some parts of the world have low bandwidths, but also certain populations may be more receptive than others.

    Are we finding that, for instance, the UK population provides better conversion than say the USA?

    Is it the case that you might be wating your money if you used (say) video in the UK rather than the USA?

  2. Hunter Boyle says

    Thanks for posting, Rob. I haven’t seen research that directly answers those questions, but if you’ve seen any, please feel free to post some links. It would be interesting to correlate some broader conversion trends with geographical access data and see if any parallels exist, even if that wouldn’t be considered hard science. Then again, I’m no scientist and I’d still like to know. Great questions to consider.

  3. Robert Lamb says

    I read your article and have thought about the long vs. short approach. With no research data to back up my premise, I’ve decided on a “short but many” approach. So my “many short” equal one long. I’m speaking of product demo and how to when I’ve envisioned this thought. I’m not sure if this is right but I’m thinking of taking a stab at this approach initially.

  4. Hunter Boyle says

    Thanks for the comment, Robert. That sounds like a very reasonable approach to start with, especially for demos and how-to. If those types of videos are being viewed before a purchase, the shorter format would likely be enough. You might also complement the short video clips with a product-specific FAQ or shortlist of specs and benefits. That can help reinforce those key points for people even if they watch the clips, and quickly inform those who decide not to click the play button. It’s also a good idea to ensure prospects know the clip length is only 30-45 seconds: make that clear before they click, and reinforce it with the video player timestamp. Let us know how your video testing works out.

  5. j. noronha says

    I think shorter videos work best to sell an idea or product, nobody wants to watch half an hour of something they, many times, don’t even know what is about.

    Longer videos may be useful in HowTos and product demonstrations though.

  6. Steven Sonsino says

    Hi Hunter

    Just some stats from me to add to the debate. I started an email newsletter in 2005, added audio podcasts in 2006 and video podcasts in 2007.

    I tracked open rates through 1shoppingcart and clickthroughs with feedburner. I found over 24 months (56,000 downloads) that:

    a) folks are 4x more likely to open an email promising an audio podcast than an article

    b) folks are 8x more likely to open a video podcast than an article.

    Tracking whether folks actually listened/watched is tougher, but:

    Audio: anything up to 12 minutes long – everyone who opened, listened to the whole podcast on site. No one bailed out. All were ALSO downloaded. (Tracked with SonicMemo.)

    I only ever posted one audio of 47 minutes and that was downloaded, but rarely reviewed on site.

    In 2007 I had 37 videos of 3-15 minutes on iTunes and saw 36,000 downloads in 6 months in 2007. (The podcast became the most popular leadership podcast on iTunes inside 3 months.)

    In the last 2 months I have had 20 videos (from 1-30 minutes) viewed 3,282 times on a WordPress blog by 818 unique visitors. No active traffic generation strategies used… yet.

    Average time viewers spend on my current video site is 5.27 minutes (Google Analytics), while the time readers spend on my article site is 2.12 minutes. Viewers stay on my corporate site (some resources but mainly about our company) less than a minute.

    I’m sure you can draw your own conclusions.

  7. I E says

    I think 30sec – 1 minute videos work the best. Quick and to the point.

  8. Hunter Boyle says

    Hi Steven,

    Thanks for sharing those interesting results. There’s a lot to think about there and I’ll be sure to (re)visit your site to see the numbers come to life, as well as mention the data to our team of analysts. By the way, I like the video on your homepage — and I couldn’t agree with you more.

  9. Stephanie says

    I can’t believe how fluid videos are online these days–both in streaming video or in sidebar advertisements. Although the more does not equal the merrier – I’ve been to pages where three video ads were playing simultaneously. Yikes!

Leave A Reply

Your email address will not be published.