Every day, I speak with retailers who are planning to launch and/or scale their product video program and they have questions about best practices. They want to know how to set up their studio, which products to include in their first flight of videos and what style seems to work the best. The most common question I hear pertains to video length: everybody wants to know if there is a perfect duration for a product video.

Thankfully, Liveclicker has the most expansive warehouse of video commerce data in the world. We are equipped to answer product video length questions better than anyone because we track more than a billion video impressions every year. As an organization that was founded by guys who previously led a web analytics company, Liveclicker is passionate about helping retailers prove the effectiveness of their video programs. We were the first company to offer in-depth video engagement tracking, content analysis, search engine optimization reporting, side-by-side video comparisons and video-to-conversion analysis.

Today, Liveclicker proudly released a series of next-generation video commerce analytics called "Custom Insight" which allow ecommerce professionals to perform in-depth ad hoc drill-downs. Let's say our customer OnlineShoes.com wants to compare the conversion rate of 45-60 second videos hosted by female presenters with 45-60 second videos hosted by male presenters. They will be able to do that within seconds. If our customer Petco wants to compare the dollars-per-video-view of their dog videos with their cat videos, that will be easy to do. When our customer Advance Auto Parts wants to analyze the average order size generated by their educational videos on Facebook as opposed to the ones shown on advanceautoparts.com, they will be able to include data from any historical period of time. Liveclicker customers will have access to unparalleled ad hoc reporting not offered by any other video solution on the market. Additionally, Liveclicker has always partnered with leading web analytics companies like Omniture, Coremetrics and Google Analytics to allow its customers to integrate our data with their measurement of other online marketing initiatives.

With data like this, it is also possible for Liveclicker to draw macromarketing conclusions by aggregating and benchmarking multiple retailers together.  As an example, I did some initial analysis which incorporated the data from more than 50,000,000 views of 6,694 product videos for 25 well-known retailers during the past 90 days.  I then drilled down on four segments within the mountain of data:

• Videos where more than 75% of the viewers watched until completion
• Videos where less than 5% of the viewers watched until completion
• Videos where a purchase was made after they were viewed
• Videos where a purchase was not made after they were viewed

There were some obvious confirmations that arose. For example, we would all suspect that the low-engagement videos would be longer than the high-engagement videos. That goes without saying. However, there are dozens of more specific questions that can be answered with this type of analysis:

• Exactly how much longer were the low engagement videos? 85 seconds - 1:54 as opposed to 0:29.
• Were people less likely to even click the play button on longer videos than on shorter ones? Yes, people were more than four times as likely to play a video which is 30 seconds or less than one that is more than 90 seconds long.
• Did viewers abandon these videos right away, once they noticed how long they were? No, actually more of them watched at least 10 seconds of the video than their high completion counterparts.
• Did the abandoned videos drive lower conversion rates than the full completion ones? Yes, the completed videos were 22% more likely to lead to a conversion than the abandoned ones, but the conversion rate for the abandoned videos was still surprisingly high. In many cases, a viewer stops watching a video because he/she is ready to make a purchase.

I also broke down the videos into time ranges and analyzed the conversion rates and dollars-per-play of each group. While the 30-second videos were excellent for engagement (often retaining viewers for full completion), it was the 60-90 second videos which yielded the highest conversion rates and revenue numbers. Amongst the videos that did not lead to a conversion, most of them were less than 60 seconds long. Thus, engagement and conversion rate are not always related, which proves that retailers need to have both types of reporting at their fingertips.

Obviously, the type, cost and complexity of a product are all determining factors in the ideal length of its explanatory video. Also, educational videos can be much longer than product demonstration videos, and branding videos can differ from those found on product pages. I also firmly believe that viewers' attention spans during 2008-2011 will be longer and more patient than in 2012-2015, when their tolerance for uninformative content will decrease. Simply put, that high performing 1:24 video might need to be trimmed to 0:40 a year or two from now.

In the coming weeks, I will reveal some more detailed and actionable findings from Liveclicker's Custom Insight reports. As the hosts of the annual Video Commerce Summit and the leaders in video commerce, we have always been excited about sharing our subject matter expertise with retailers who appreciate the value of video.