Budgeting for product videos for the following year had not been completed, and the request came from the production manager for data to understand the ROI that could be expected from new video production.
Since the products included in the test had videos produced for them - which were already live on the site - the test was skewed more heavily toward including them in the control group.
As a visual merchandiser, I want to know if videos succeed in making a better first impression, because our resources are finite and should be spent on projects that have the highest ROI
Over a 2 week period, qualifying product views, video plays, AOV and conversions were tracked. After enough data was collected to reach statistical significance, we found:
Because no additional development was required to set up the test, it launched quickly with no blockers. The only deficiency was a lack of product-specific conversion tracking. This means that while only those visitors who saw product with a video-available were included in the test, the final conversions numbers would be comprised of transactions with any products, not only those with video available. Separate analytics could have been gathered to build product-specific conversion numbers, but the resources were not immediately available to gather them. In any case, the results were deemed positive enough to move forward with budgeting video production for the most popular products for the coming year.
With the increase in conversion and AOV for visitors who watch the product videos, the question quickly becomes how to increase the percentage of visitors who watch:
