By Chris Koopmans*, VP and GM, Service Provider Platforms,
Citrix
As radio access technologies have evolved and downlink speeds have increased, we’ve seen video content taking up a larger proportion of the total data traffic on mobile networks. Specifically, as RANs have moved through 2G to EDGE, UMTS and into HSPA, the combination of faster networks, cheaper data, smarter devices and better QoS for video means that video as a percentage of traffic has risen from 30 percent to over 50 percent.
In the five years since the iPhone first launched and kick-started a smartphone revolution, video has become a much bigger portion of traffic on mobile networks. Unfortunately for mobile operators, the proportion of revenue that video brings in has not grown at the same rate as the traffic. As we move into LTE and towards LTE-Advanced we’re going to see the share of video continue to grow.
We may not need to wait long for video traffic to accelerate its growth. Facebook looks set to begin moving forward with video adverts, a step that may be setting off alarm bells at networks operators around the world – or at least wherever Facebook has a presence. In the Citrix Bytemobile Q2 2013
Mobile Analytics Report we found mobile Facebook usage has grown five-fold in a year, with the level of traffic Facebook generates growing from one to five percent of overall data traffic. That includes very little video. So, what happens when we add video adverts to Facebook?
In the same Mobile Analytics Report, we indicate that a single mobile YouTube session contributes the equivalent of ten mobile Facebook sessions -- that’s the power of video. When Facebook rolls out video adverts, there will be an unavoidable bump in data traffic associated with Facebook. When video adverts are added to all of the other online services, that new traffic could start to have a seriously detrimental impact on the quality of service subscribers enjoy.
Of course, even in the absence of video adverts, operators have had to manage the growth of data on their networks – or risk financial implosion. The tools used range from relatively blunt instruments such as throttling and lossy compression to more nuanced approaches such as adaptive traffic management, the latter employed in the service of ensuring the best possible subscriber QoE within the constraints of the network. More recently, operators have begun to consider QoE as part and parcel of a differentiated services plan in which the subscriber experience – including access to particular content – is tied directly to willingness to pay. The question at hand is: which of the above approaches – or other approaches -- are most relevant in the context of video adverts?
There are a few things to consider. First, these adverts will be pushed over the network over and over again, making the excess traffic challenge that much worse. Second, at approximately 2MB each, these adverts, if seen many times per day and month, can serve to consume a significant portion of the subscriber’s quota. Finally, the effectiveness of the advert and the likelihood of having a subscriber click through will be related to the video experience, for example the amount of time it takes for the advert to load.
Given these considerations, it’s clear that some degree of data optimization is required in order to reduce the overall transport costs and, perhaps more importantly, to minimize the hit to subscriber quotas. Adverts can be easily cached and highly compressed, so the blunt instrument approach could certainly be justified. After all, the operator isn’t making any extra money from these adverts. But they could be.
If ever there were an opportunity to pursue a two-sided business model – one in which the operator collects revenue not just from subscribers but from content providers as well – this is it. Both Facebook and its advertisers have a financial incentive to generate as many click-throughs as possible. Meanwhile, the operator -- in partnership with an optimization vendor -- has the tools at its disposal to affect this click-through rate while better serving the interests of its subscribers.
At the most basic level, an improved user experience enabled by optimization would better ensure the timely delivery of the advertisement. Taken a step further, the operator could use its understanding of real-time user experience to only send adverts when the subscriber is likely to have a good experience. Who’s benefitting here? Facebook (or equivalent). The basis for an agreement between the operator and the social networking site is thus established.
And it doesn’t need to end there for either party. The operator has the option of zero-rating the traffic associated with these video adverts, so as not to unfairly consume the subscriber’s data quota. The presence of a business arrangement between the two makes doing so even more palatable. Perhaps of most interest (and most controversy) is the idea of sharing certain elements of the subscriber profile – elements inaccessible to Facebook or the advertiser – thus enabling the more effective targeting of adverts. This, of course, is exactly what all parties want.
What looks at first like a network burden that must be ignominiously accommodated like so many others is in fact an opportunity for operators to establish partnerships with the content providers, for whom an enhanced subscriber experience translates to revenue.
________
*Chris Koopmans joined Citrix as vice president, Service Provider Platforms, with the acquisition of Bytemobile in 2012. He is responsible for product development, management and marketing for the Service Provider Platforms group, as well as the Citrix business strategy for telecommunications service providers. Koopmans was a founding engineer at Bytemobile in 2000 and rose to the position of chief operating officer, responsible for all aspects of product development, management, marketing, delivery, and support, as well as information technology (IT). He has over 14 years of industry experience in hardware and software engineering and architecture.
The trend of cloud-based analytics reporting continues (see Openwave, Movik).
Posts
