• by Steve Smith , Staff Writer @popeyesm, June 1, 2011

How many users does a publisher lose every time the viewer has to endure a delayed start of video content or multiple buffering pauses? And who does that user blame? Hulu or Netflix? CNN? Or Comcast and Cablevision? Surely Akamai and Level3, generally unknown to most onliners, will not take the heat. It is hard to say who pays what sort of price for poor video performance on the Web, but Alcatel-Lucent introduces this morning a video analytics product for top level service providers that is designed to ferret out where and when quality of experience problems exist in the system and help determine if it leads to frustrated customers. "The goal is to reduce churn," Alcatel-Lucent Ventures' Mark "Buck" Peterson says. "A lot of these online subscribers when they buy Internet service have a big loyalty to over-the-top content." Netflix or Hulu are now the streaming media that people are following across tablets and TVs in their homes, and consistent performance is becoming their priority, he contends. "If they get bad experiences with that they tend to move to other service providers. They blame the service provider."

Well, we probably don't know that for sure yet, but Alcatel-Lucent Ventures, a strategic investment group within Bell Labs, is betting on the growing paranoia among MSOs and other service providers who now have competitors breathing down their necks. The AppGlide Video Analytics managed service tries to look at the video experience from the end user's point of view. It uses player plug-ins to measure Quality of Experience (QoE) across the network and determine when things like delayed starts and buffering cause people to bail in different regions. It also offers tools for measuring content delivery network devices as well as analyze content usage.

Peterson says that while other products measure various pieces of the QoE equation, the AppGlide product stands out for its cross-correlation of data. He says customers have been asking for analytics that show whether their churn is attributable to bad product or to bad delivery. When the service providers try to secure new and better content from publishers, the media brands want evidence that the network will deliver good experiences. Whether the service providers will agree is anyone's guess. Peterson says the service is cloud-based and priced on a per-subscriber basis at pennies per user.

One can imagine how a broad view of QoE can inform a number of business decisions and even help imagine new product offerings. For instance, do we know whether it is worthwhile for publishers and service providers to focus energy on hi-res iterations of their video online? I for one gave up on ticking the 1080p option on YouTube long ago, even though I have enhanced Comcast broadband. The delivery is halting, whether it is made so by Comcast, YouTube or just Google Chrome's persistently weird interaction with Flash players. Who do I blame? I don't know for sure. Does it keep me from watching videos? No. Does the Hi-Def option add or retain users? Who knows? Peterson says that they do know that users are more irritated by changes in video quality in adaptive streams. Apparently it is better to start and stick with a modest bit rate than to kick resolution up and down in the middle of an experience. 

On the other hand one can imagine cross-correlated data on QoE also being used to segment high-usage audiences like me and target them for higher tiers of service. Some time soon I may be getting a note from Comcast pointing out that my drop-offs during that botched iPad viewing of Parks & Recreation could have been prevented...with a 50MB/s super-fat pipe upgrade.

Data like this could improve the end user experience. It could also be used to kick up my already exorbitant cable bill.