In a recent Heavy Reading survey of more than 90 mobile network operators, network performance was cited as a key factor for ensuring a positive customer experience, on a relatively equal footing with network coverage and pricing. By a wide margin, these three outstripped other aspects that might drive a positive customer experience, such as service bundles or digital services.
Decent coverage, of course, is the bare minimum that operators need to run a network, and there isn’t a single subscriber who is not price-sensitive. As pricing and coverage become comparable between operators, though, performance stands out as the primary tool at the operator’s disposal to win market share. It is also the only way to grow subscribers while increasing ARPU: people will pay more for a better experience.
With 5G around the corner, it is clear that consumer expectations are going to put some serious demands on network capability, whether in the form of latency, capacity, availability, or throughput. And with many ways to implement 5G — different degrees of virtualization, software-defined networking (SDN) control, and instrumentation, to name a few — network performance will differ greatly from operator to operator.
So it makes sense that network quality will be the single biggest factor affecting customer quality of experience (QoE), ahead of price competition and coverage. But there will be some breathing room as 5G begins large scale rollout. Users won’t compare 5G networks based on performance to begin with, since any 5G will be astounding compared to what they had before. Initially, early adopters will use coverage and price to select their operator. Comparing options based on performance will kick in a bit later, as pricing settles and coverage becomes ubiquitous.
So how then, to deliver a “quality” customer experience?
5G, highly virtualized networks, need to be continuously fine-tuned to reach their full potential — and to avoid sudden outages. SDN permits this degree of dynamic control.
But with many moving parts and functions — physical and virtual, centralized and distributed — a new level of visibility into network behavior and performance is a necessary first step. This “nervous system” of sorts ubiquitously sees precisely what is happening, as it happens.
Solutions delivering that level of insight are now in use by leading providers, using the latest advances in virtualized instrumentation that can easily be deployed into existing infrastructure. Operators like Telefonica, Reliance Jio, and Softbank collect trillions of measurements each day to gain a complete picture of their network.
Of course, this scale of information is beyond human interpretation, nevermind deciding how to optimize control of the network (slicing, traffic routes, prioritization, etc.) in response to events. This is where big data analytics and machine learning enter the picture. With a highly granular, precise view of the network state, each user’s quality of experience can be determined, and the network adjusted to better it.
The formula is straightforward, once known: (1) deploy a big data lake, (2) fill it with real-time, granular, precise measurements from all areas in the network, (3) use fast analytics and machine learning to determine the optimal configuration of the network to deliver the best user experience, then (4) implement this state, dynamically, using SDN.
In many failed experiments, mobile network operators (MNOs) underestimated step 2—the need for precise, granular, real time visibility. Yet, many service providers have still to take notice. HR’s report also alarmingly finds that most MNOs invest just 30 cents per subscriber each year on systems and tools to monitor network quality of service (QoS), QoE, and end-to-end performance.
If this is difficult to understand in the pre-5G world — where a Strategy Analytics’ white paper estimated that poor network performance is responsible for up to 40 percent of customer churn — it’s incomprehensible as we move towards 5G, where information is literally the power to differentiate.
The aforementioned Heavy Reading survey points out that the gap between operators widens, with 28 percent having no plans to use machine learning, while 14 percent of MNOs are already using it, and the rest still on the fence. Being left behind is a real possibility. Are we looking at another wave of operator consolidation?
A successful transition to 5G is not just new antennas that pump out more data. This detail is important: 5G represents the first major architectural shift since the move from 2G to 3G ten years ago, and the consumer experience expectation that operators have bred needs some serious network surgery to make it happen.
The survey highlights a profound schism between operators’ understanding of what will help them compete and succeed, and a willingness to embrace and adopt the technology that will enable it. With all the cards on the table, we’ll see a different competitive landscape emerge as leaders move ahead with intelligent networks.