The Devil's in the Details
Streaming-Only Versus Multiple Types of Content
So do you really need a tool that measures streaming content, or will a standard web analytics tool work just as well?
A typical web analytics tool, such as Google Analytics, is based on a simple Java tracker script, which resides on each webpage and tracks visits from Java-enabled browsers. This covers a large number of desktop browsers and a more limited number of mobile browsers. A web analytics tool provides details such as unique users, requests, pages visited and revisited—the latter being visits by unique viewers that occur within a particular time frame (e.g., 30 minutes, 1 hour)—and particular search keywords used to find a page.
There’s another level of web analytics tools that may already be available in your organization, which includes those linked to a content management system (CMS) or enterprise resource planning (ERP) solution. One example of this type of solution is Documentum, which is owned by storage giant EMC. Documentum features what EMC calls "Audio Video Transformation Services", including a direct integration with products such as Telestream’s FlipFactory.
Yet these tools are often geared towards details on file storage and delivery, where the logs may note a file as delivered or not delivered, but they don’t tell the story of whether the viewer had a quality experience or even watched the entire video file.
By contrast, a streaming analytics tool offers those summary details—often integrating with a free tool such as Google Analytics or a paid web analytics solution such as Omniture—plus a significant set of details needed to gauge streaming delivery. The additional detailed reports may include file duration, percentage viewed by unique viewer, ad insertion log files, and error messages or status codes generated as part of the streaming experience. The latter two reports are especially valuable to the streaming engineer, who can use this information to modify delivery, while all of the reports are beneficial to the marketing or communications team, which needs to assess content effectiveness as well as provide information to a growing customer base that may need help preparing browser or plug-in settings for optimal streaming playback.
The way a streaming reporting solution works is frequently done by segmenting the content into different time intervals. Some of these solutions need to be pre-encoded, while others are deployed by content delivery networks (CDNs) who then offer the reporting as a part of the CDN delivery service alongside log files for Flash, QuickTime, RealMedia, Windows Media, and Microsoft Silverlight.
Skytide’s Insight for Flash Video Players is a video-centric analytics tool that attempts to not only analyse "viewer engagement" with content but also measures quality of service in terms of delivery. This service was chosen as the winner of the 2009 Streaming Media Readers’ Choice Award in the reporting and analytics category, and it uses a server-side code that is sent to the client Flash player to analyse quality and user interaction.
For streaming analytics, Skytide and others are looking to answer basic questions that can’t be answered by a traditional web analytics tool, such as the following:
• Which types of content deliver the highest engagement levels?
• How long do viewers stay with a particular video?
• Where is the optimal position within a video to place an ad?
• Which CDN delivers the best quality of service?