VAST Wrapper vs Inline
A VAST wrapper is a VAST XML response that contains a URI to a VAST ad hosted at a different location. There are two types of VAST responses: wrapper and inline. A VAST wrapper contains a link to another VAST response which could be another VAST wrapper creating a chain of VAST wrappers. An inline VAST response contains the actual video creative, it does not link to any other VAST response.
Benefits of VAST Wrappers
A benefit of using a VAST wrapper is that it allows a third party, which does not host the video asset, to insert their own tracking pixels and receive notification when certain events occur (ie video starts, video is clicked, video completes, etc.). The VAST wrapper allows a third party to gain valuable information even though they do not have anything to do with the delivery and display of the video.
A video player will receive the initial VAST response and if it is a wrapper it will continue to follow the chain of VAST wrappers until it gets to an inline VAST response and then it will display the video ad. As the video player unwraps each VAST wrapper it stores all of the tracking pixels contained within each VAST wrapper response. When the video plays it will fire pixels for all events (impression, video start, quartile, etc.) at the appropriate time from all of the VAST wrappers and inline VAST response.
Three potential problems for mobile video with VAST
The first potential problem is latency. Each VAST wrapper requires the video player to take time to unwrap the VAST response and then send a request to the third party VAST URI contained within it. As the number of VAST wrapper levels rises, the amount of time it takes to unwrap and send a request increases and a user must wait longer for the video to display. On a mobile device this wait time can become significant and can cause the user to leave before the video has a chance to display. Previous studies have shown that for every 1 second it takes to load a page on mobile 7% of users leave, so as the number of VAST wrappers goes up the amount of users viewing the video decreases.
The second potential problem is tracking pixels. Some third parties insert lots of tracking pixels into their VAST wrapper, sometimes adding as many as 20 different tracking pixels in the impression element alone. Although, the IAB VAST spec does not set a limit on the amount of tracking pixels that can be set for a single event, a large number of them can can cause potential problems.
The first issue can be that the video player needs to store and track all of these tracking pixels to fire later, and while the storage of them may not be a problem for many video players a more plausible downside is that some video players limit how many tracking pixels they will fire for a specific event. Some video players send the tracking pixels to be fired from their server so they can track them, and in some instances the video server limits the amount of tracking pixels fired to 5 per event. Any pixels after the fifth one are cast aside and never fired. This can cause major issues with tracking and reporting.
The other potential downside to VAST wrappers containing up to 20 tracking pixels per event, is the increase in the size of the response, which slows down the delivery of it. Although the size increase may appear to be minimal, if a VAST response has 3 wrappers and all of them take longer to deliver it can affect the overall latency of the video delivery and it has been proven that even a small delay in delivery and can have a large impact on whether a video is viewed or not.
The last potential problem is the number of VAST wrapper levels. Some video players have a limit on the amount of VAST wrappers they will unwrap. Once they reach that limit and have not found the inline VAST response, they will stop unwrapping the responses and move on to the next ad in the waterfall. Too many wrappers can prevent an ad provider from having their ad shown.
To “Wrap” this up…
VAST wrappers can be very useful, but can also cause problems for video players resulting in reporting discrepancies, poor user performance and lack of video display. Publishers normally do not have much control over whether a VAST response is wrapped or not, and the majority of VAST responses are wrapped so it is something that cannot be avoided. Most VAST wrappers do not contain more than 2 or 3 levels, which is acceptable, but Publishers need to monitor the performance of their ad partners, and if they see a rise in latency in the delivery of their ads and/or see a drop in their fill rate they should review their VAST wrapper responses to see if that could be a potential cause.