“Now I Understand What You Mean with Contextual Intelligence in Policy Control and Charging”

Last week I attended the Broadband Traffic Management event in London, gave a speech on analytics-driven Policy Control and participated in a panel on Over the Top (OTT) and Communications Service Providers (CSPs): ‘What are the Obstacles for Two Sided Business Models?

The need for policy control is evident.  CSPs cannot keep on investing in the networks to the same extent they have been doing in the past. The costs are too heavy, and CSPs are actively looking for smarter and faster ways to monetize data.  The key topics of the conference mostly span around Wifi offload (lightly, but there), OTT (heavily, how to partner), video (heavily, how to control). Video is considered the main culprit in consuming all the bandwidth, and CSPs are very worried about the growth.

The main message of my speech was to fundamentally state that “rules-driven PCC is not going to lead to success” and “policy control needs to be augmented with real-time predictive analytics”. One of the use cases I presented was a so called “intelligent turbo boost” which brings the possibility to dynamically alter the bandwidth, price and duration based on congestion and subscribers’ propensity to pay – as opposed to  the standard “bandwidth boost of 1Mbit/s for 1 hour for 2 euros”.  The second use case example introduced “contextual video optimization” which means using subscribers’ propensity for deciding how much video is optimized for a specific session instead of using rules that are static and fixed. The difference in approach is that humans are hardly driven by static rules, and therefore building the environment on a way where fixed rules define how everything works is really counter-productive.

I received very good feedback after the presentation. Most discussions ended up in the conclusion that pools of rules are very hard to maintain, and they hardly ever meet the subscribers’ expectations. They are looking for a service provider that can offer a selection of personalized data packages which flexibly responds to their needs on-the-spot and within a particular context.  That is something we would like call contextually intelligent customer experience management.