Over the last few years, many streaming video publishers have begun migrating their video ad technology from client side ad insertion to server side ad insertion (SSAI). With SSAI, ad creatives are inserted into the video stream as media segments with the same playback quality as that of the video segments. The primary objective in using SSAI is to improve a viewer’s experience, bringing the experience of ad breaks in video streaming closer to the seamless ad breaks viewers are familiar with from watching traditional television. Secondary objectives include simplifying video player development, as there is no need to instrument ad SDKs, and minimizing ad losses due to ad blockers.
But does SSAI technology actually improve the viewer experience? A comparison of quality of experience data for server side versus client side ads helps validate whether this expectation holds true.
Conviva analyzed 3.3 billion on-demand video streams across desktop, Roku, and Android Fire TV players during the months of June and July.
First, we compared the video completion rates and video startup times for videos with client side versus server side ads. We found that content completion rates are higher by 5% – 10% for videos with server side ads, even though video start up times are longer by 1- 2 seconds. This suggests that SSAI systems introduce an additional latency for videos to start which may cause viewers to drop off while waiting for the video to start. However, once the viewer has started playing a video, then they are more likely to watch the video to completion.
Second, we wanted to validate whether dynamic ad insertion in SSAI systems increases video startup times. By comparing video startup times on SSAI video streams with no dynamically inserted ads against those with ads we found that video streams with no ads had a one second shorter video startup time and also resulted in 2% – 5% higher video completion rates than those with ads.
We also compared ad creative buffering and average bit rates between client and SSAI streams and found these metrics were comparable with differences mainly attributable to the differences in video player and platform rather than the use of SSAI.
We concluded that, while SSAI ads certainly help with better viewer experience once a viewer starts watching, it could potentially cause more viewers to drop off before even starting to watch the video. Our recommendations for best practices as a publisher moves to SSAI are:
- Work with your SSAI vendor to improve the video startup times when inserting ads
- Improve the video app experience by combining client side and server side ads – for example, you could play a pre-roll client side ad while the SSAI system prepares a personalized video stream with dynamically inserted ads; you can then switch to the SSAI stream with mid-roll ads
- Balance video streams with no ads versus those with ads – for example, you might run video streams with no ads or few ads for your most valuable viewer segments/subscribers
Utilizing Conviva to monitor your streams across content and ads allows you to identify areas for improvement as well as to implement and monitor strategies to balance the various factors that must be considerd when moving to SSAI.