Playing to Win in the Video Game

It is perhaps no surprise that the way in which users watch mobile video varies greatly. What may come as a surprise, however, is that “viewer abandonment” studies have found that about 50 per cent of viewers will browse away from video clips after watching less than half the duration of a clip.

Knowing how users watch video and what they watch can make a dramatic difference to mobile network operators, most of whom now have video optimisation on their agenda. Mobile video is a growing cause of traffic on networks: today, video accounts for 60 per cent of all HTTP traffic, and is projected to reach 64 per cent of all mobile traffic by 2013. Mobile operators are looking to optimise video, without excessive investment in resources, while still ensuring that subscribers continue to receive a high quality viewing experience. They need to find the right balance between user experience, network availability and profitability.

So why do viewers abandon video? Imagine a scenario if you will. A group of friends gets together. One of them is sent a link to a video and he starts to watch it, laughing. His friends want to watch too, so he starts the clip again. But when it becomes clear they can’t all watch around the same phone, someone else gets their phone out. After watching the first clip they start browsing through the video suggestions. The first is boring, the second one is slow to load, the third is good to start with… but then it gets dull, so they switch it off. The fourth video is great, but then the phone rings.

Best results
As this typical example illustrates, the group watched a number of videos, but how many in total, and how much of each one? Most importantly, how should the operator optimise those videos to get the best results for the consumers and the network?

Video optimisation can reduce video traffic on a network by as much as 40 per cent. But to keep the network resources required for optimisation in check, mobile operators need to use user-aware video optimisation techniques.
User-aware video optimisation solutions continuously analyse user viewing patterns and make real-time decisions about where best to utilise the optimisation resources. Operators are able to optimise only the most-viewed videos, rather than every clip that crosses the network. In this way, they can get the desired bandwidth savings, and also reduce their optimisation resource requirements.

One user-aware video optimisation technique is called Statistical Caching Optimisation, which caches and optimises only frequently-viewed videos. Mobile operators can also choose to optimise only the most-viewed sections of each clip, rather than the whole of the clip. Real-time analysis of user behaviour highlights the most-viewed parts of clips, and operators can choose to optimise those segments only, reducing resource usage, while still providing bandwidth savings.

Just-in-time delivery
Another user-aware optimisation technique is called Pacing. Rather than delivering an entire video as quickly as possible, Pacing - also known as ‘Just-in-time’ delivery - serves each video stream to match the pace at which a person is viewing it. This further opens up bandwidth to enable more concurrent sessions. Although there are various ways of implementing Pacing solutions, the most efficient way is to actually analyse the content of each video and to optimise each part as necessary: something that Deep Packet Inspection (DPI) can’t do.

Regardless of whether users watch a video to the end or not, how long they actually watch before they browse away is relevant for individual and statistical purposes. Other significant user behaviours include how they pay and where they are when they stream video - at home or roaming, for example.

Using the mobile network operator’s policy and charging rules function (PCRF), optimisation can either be applied or not for each user according to the operator’s rules database. This is something that the Mobixell Seamless Access policy and charging enforcement function (PCEF) can enforce. For example, an operator might want to set how optimisation is applied for video streamed to prepaid users who are charged by packet, or roaming users might, by default, receive all their video optimised, and have controls to choose to turn off optimisation.

Video optimisation techniques such as these reduce the volume of data required to deliver video traffic, enabling mobile operators to defer expensive network upgrades, while maintaining excellent quality of experience for subscribers. What’s more, user-aware optimisation helps operators to significantly reduce the amount of network resources required for implementing optimisation.
When combined with device- and network-aware techniques, user-aware optimisation provides the best viewing experience possible for our group of friends and every other user, whether they choose to watch clips all the way through or whether they use their imagination to fill in the ending.

Noam Green is VP marketing at Mobixell