Driven by computing, storage and artificial intelligence (AI) advances, there’s no doubt that telecom analytics is growing fast. According to research firm MRFR, by 2023, the global telecom analytics market will soar to more than $11 billion with a 33 percent Compound Annual Growth Rate (CAGR).
Thanks to data infrastructure improvements, data-rich carriers can now yield meaningful intelligence to transform their businesses. Carriers can now merge and analyze customer data related to spending and plans, as well as behavior data like Internet usage or duration of networked activities. Carriers can also now get insights from location data, such as roaming or most visited locations, and device data, such as brands and screen sizes.
All these analytics are poised to transform how carriers operate networks and businesses. Here’s a look into what we can expect in 2019.
#1: Network Intelligence Will Optimize Networks
It’s well-known that mobility and new 5G investments challenge carrier networks. With the introduction of 5G, carriers will have to constantly reallocate resources to manage changes in network traffic. The only solution to handle this will be for carriers to widely deploy virtualization along with machine learning-based traffic analytics.
Virtualization using Software-Defined Networking (SDN) and Network Function Virtualization (NFV) inherently offer flexibility and agility. Together they enable agile networks. Of course, an agile network will quickly find new performance limits if it’s only based on traditional, descriptive traffic analytics.
Enter machine learning-based traffic analytics. Machine learning depends on a constant flow of high-quality data (which network operators have). Over time, network management machine learning models learn, and become better and faster. Thus, in virtually any scenario of seemingly random traffic chaos, carriers will be able to very accurately predict capacity problems as well as an individual customer’s behavioral response. They will also be able to use machine learning to suggest the next best step they should take to solve problems, and continuously provide high quality and reliable connectivity.
#2: Subscriber and Customer Experience Insights Will Lead to Happier Customers and Lower Churn
On both mobile and broadband networks, subscriber usage data helps carriers understand behaviors and optimize customer experiences. Carriers have real-time visibility of subscriber behavior - be it making a call, texting or surfing the web - and visibility into the success and failure of that activity at every network location, including both edge and core.
In the months ahead, carriers will start to merge that behavior data with information collected from modern customer support channels like call centers, primary research surveys, emails, bots or even Tweets.
By applying machine learning to understand sentiments, carriers can identify at-risk subscribers or causes of current and future churn. They can then develop marketing or proactive customer care programs and offerings to prevent churn. They can also identify subscribers who are ripe for data plan upgrades, likely to respond to certain ads or likely to buy new services.
#3: Ad Hoc Intelligence Will Fuel Better Marketing and Planning
Rohde & Schwarz
The entire population of subscribers’ behavior data rolls up to a market view. This means even small, but fast-growing, services may be meaningful indicators of future demand. On the other hand, small and fast-growing service attrition might indicate a new competitor or another marketing problem.
Analytics from traffic data also allow carriers to segment subscribers by behaviors and 3rd party data sources can lend richer understanding. For example, carriers can create subscriber groups or segments based on subscriber usage of leading-edge services today. From there, they can merge behavior data with other 3rd party customer intelligence platforms for a 360-degree view of the customer. Finally, carriers can determine characteristics of the most loyal or highest-value customers and find more like them.
In 2019, all of these analytics will help carriers understand tomorrow’s subscribers and uncover market changes faster and earlier. Carriers will become more agile, change or offer new services and improve customer support.
#4: Artificial Intelligence, Machine Learning and Data Will Drive New Revenue
The combination of AI, machine learning and mobility unleash new applications and content for enterprises, the government and consumers. This includes location-based advertising, premium personalized content or streaming video to mobile and other latency-sensitive Internet of Things (IoT) apps.
Of course, this means it’s increasingly necessary for applications and security to run closer to users than they do now. After all, if an app runs closer to users at the network edge, app data travels shorter distances using fewer network resources at reduced latency. With the changing role of the network edge, location data will become increasingly important. By knowing location, apps will be able to use AI to process data and reduce the quantity sent to centralized data centers.
Carriers have real opportunity to monetize the network edge. They can equip it with compute, storage and network resources, as well as the tools that enable rapid innovation. Location and network analytics on the carrier network edge will also assure dynamic allocation of network resources when and where they’re most needed. The result will be better performance at reduced cost.
#5: Security Detection Intelligence for Safer Networks
In addition to more data and more traffic, more devices and endpoints mean heightened risks to carrier networks. At the same time, it’s impossible to deploy perimeter defense everywhere. To make sure there is robust security, carriers will increasingly rely on intelligence from security detection solutions that use deep packet inspection.
Deep packet inspection (DPI) is critical here. DPI continuously monitors all data packets, looks for anomalies and makes decisions based on pre-defined rules. It examines network metrics, such as traffic volume, bandwidth consumption and protocol usage according to “normal” usage. Any departure from “normal” is an anomaly that could be a threat, thereby triggering an alarm.
Even more important is early anomaly detection. This is challenging for a few reasons. First, it’s hard to arrive at a well-defined notion of “normal usage” in a highly dynamic environment where systems, devices and behaviors constantly change. Second, it requires the ability to detect anomalies without relying on known patterns that can only emerge over long periods of time.
In other words, carriers will increasingly need anomaly detection to uncover very subtle changes. Early anomaly detection is tailor-made for machine learning and it’s only made possible by ever-improving big data infrastructure, machine learning advances and abundant, high-quality network data generated over long periods of time.
In the coming years, carriers will increasingly compete on telecom analytics. No doubt, greater use of DPI increases the likelihood of detecting network crushing and business-killing security anomalies. However, it won’t be enough for carriers to innovate in data infrastructure and develop the most efficient machine learning algorithms alone. To dominate, carriers must excel at both generating - and applying - intelligence and insights across the network and business operations.