This blog continues a multi-part series that suggests ideas by which our customers could deploy and monetize video applications. Most of these ideas (as stated) are as yet un-tested in the industry, but we believe there is enough value in the ideas to encourage our more clever readers to enable or launch a service or two which might prove successful. This time, we talk about the end user's Quality of Experience (QoE).
Ensuring the quality of content is not something often overlooked by the industry. However, when it comes to maintaining service interest and attaching advertising revenue, it is more important to establish a high Quality of Experience (QoE). QoE is a measure of how happy the customer is. A happy customer might want to buy something. A customer with a good video experience with a Taco ad might just want to go get himself a taco, especially if the Mobile Operator (MO) knows it is lunch time and there happens to be a Taco restaurant right near where the customer is located.
Advertisers want their advertisements to be attached to good quality of experience
There are two distinct aspects to quality assurance:
In order to get people to use a service, the provider needs to provide some level of quality assurance. If the service is subscription based, there is either a direct or implied Service Level Agreement (SLA) under which a customer may have to be reimbursed if the quality of the media stream falls below a certain set of parameters. Even with 'free' content, monetization requires a certain level of performance in order to attract advertisers.
Based on data we have seen, when the MO introduces a new video service, there is often a sharp uptake in the number of minutes followed by a steep decline shortly afterwards. Now, we do not know for certain what the causes were for this rate of customer drop-out, but we believe a large part of it has to do with coverage, service availability.
Over a mobile network, there are a number of things which can lead to inferior quality. Two primary categories of variables that can affect the overall viewing experience are:
1). The encoding
2). The network variations in the airlink, caused by multipath, co-channel interference, fading etc. that can lead to lost packets which are unrecoverable in a video.
There are techniques which we could assist with that could lower the bit rate such as reducing the resolution quality but sustaining the frame rate to avoid frame loss, however, employing such techniques requires that we know the condition is occurring.
Being able to monitor the quality of received video content the user is experiencing requires that we first have a baseline of the quality of the original content. By measuring the content quality at the source and embedding a watermark that identifies the measured quality and then subsequently measuring the quality of the content at various points throughout the network (i.e. handset) we can compare the quality and know that it is either degraded through the network or matched the original quality.
Now obviously this fully end to end service would require a handset application component, but an interested carrier could choose to add this capability to handsets with a certain SLA. However, this handset piece is not necessary to provide the advertisers with some sense of the quality of the original content. This solution could still be marketed to the content provider as well as the MO who wanted to insert advertisements into the stream.
Content providers would use this technique to identify high quality content with which to marry advertisers who insist their brands are only associated with good quality content. This is especially true for the UGC market where some videos can be down-right lousy. Additionally, advertisers may want a mechanism that removes their ads in the event received quality fell below a certain threshold.
Techniques requiring handset applications would require closing deals with handset providers or getting solutions incorporated as downloadable applications in an Android-type environment. However, there is still plenty of potential for the measuring of initial content quality, even if we not able to resolve the handset side of the equation.