White Paper Advanced Predictive Network Analytics: Optimize Your Network Investments & Transform Customer Experience Prepared by Ari Banerjee Senior Analyst, Heavy Reading www.heavyreading.com on behalf of www.sas.com February 2014
Using Network Analytics for Communications In the communications industry, network performance is directly connected to customer experience. It is therefore critical for communications service providers that are focusing on providing superior customer experience in order to differentiate themselves from their competitors to be able to seamlessly connect the dots between network performance and customer experience. The other aspect that communications service providers struggle with is their poor capacity utilization and network resource management. Therefore, these service providers need solutions that will help them manage their existing assets and enable them to do more with less by providing accurate network capacity and usage information that will avoid unnecessary network build and infrastructure upgrade activity. Communications service providers have traditionally operated with complex, disparate sets of data silos, with useful information residing in multiple systems, including: customer relationship management (CRM), billing, inventory, provisioning and fulfillment, service management systems, network elements, element and network management systems, probes, deep packet inspection (DPI) devices, and application-specific databases. They also have multiple systems in different generations of network architecture, each holding different types of data, in various formats. However, what service providers lack is the analysis capability that would help them to collect network performance data and correlate that with the end-user service experience. Heavy Reading's research clearly shows that network data is the most underutilized albeit valuable resource in most service provider organizations. Providing optimum customer experience, accurate capacity planning, customer-centric service assurance, focused marketing and campaign management, etc., requires the strategic utilization of network data. The only way service providers can maximize their revenue potential from network traffic is to have real-time visibility into the network traffic information and the necessary analytics-driven software infrastructure to deliver insights on how to deliver superior customer experience and make informed and accurate network investment decisions based on projected revenue. What is needed today is a fundamental paradigm shift when it comes to analytics infrastructure, which we define as predictive network analytics. This approach to decision-making is significantly different from the traditional enterprise data warehouse (EDW) approach, the main aim of which is to achieve a single shared version of the truth that everyone needs to align around. This offline data reporting has been typically used for planning purposes. However, the major gap that always prevailed is operationalizing that analysis into real-time actions that impact a service provider's business processes. Figure 1 illustrates the critical areas where network intelligence will play a key role in the communications service provider environment. As the figure shows, a service provider network analytics strategy needs to rely on four fundamental pillars: capacity planning, service assurance, control and optimization, and customer experience and marketing. In the following sections, we will look at these four dimensions in detail. HEAVY READING FEBRUARY 2014 WHITE PAPER ADVANCED PREDICTIVE NETWORK ANALYTICS 2
Figure 1: Role of Network Intelligence in Service Provider Environment Source: Heavy Reading, 2014 HEAVY READING FEBRUARY 2014 WHITE PAPER ADVANCED PREDICTIVE NETWORK ANALYTICS 3
Capacity Planning The only way for service providers to compete in today's economic environment is by improving operational excellence. However, operational expenditure (opex) remains stubbornly high for most service providers and typically consumes 30 to 40 percent of revenue. Network operations account for about 45 percent of this spending. The expansion of network footprints due to organic and inorganic growth has resulted in poor capacity utilization. This leads to higher opex and capital expenditure (capex), and has adversely impacted customer service. Inconsistent information about network usage, poor use of existing assets and the propensity to overprovision network resources to mitigate future service disruption all directly relate to carriers' spiraling opex spending. In this context, accurate network planning, trending and optimization is becoming an area of critical importance. The dynamism and accuracy needed to optimize network infrastructure will require dedicated solutions that will provide a sufficient arsenal to business users as well as IT users to plan, predict and optimize use of their existing resources. Combining network data with true analytics, service providers can optimize network investments. With advanced predictive analytics, operators can forecast using data down to the cell level, predict usage, technology (2G/3G/4G) requirements and trending with incredible accuracy. They cannot rely on doing this assessment based on simple data and spreadsheets, as these resulted in inaccurate analysis that impacted operators' capex investments. Hence it becomes critical to bridge the gap that exists today between operational systems with planning and capacity management systems. Other critical aspects that remain causes for major concern for most service providers revolve around freeing up existing resources when services are deactivated or turned down. Many service providers do not have a robust mechanism to free up existing resources such as ports when services are turned down. As a result, stranded and unused assets often become common place in an operator's environment. As operators find themselves cash-strapped by challenging economic times and facing an insatiable consumer appetite for bandwidth-heavy applications (video in particular), one central challenge is how to balance network resources with an eye toward customer experience and profitability. Operators are equipped with a finite asset (network capacity), and the cost of increasing that asset through greater equipment spend on base elements such as edge and core routers can be prohibitive. On today's dynamic IP networks, some network elements will fall under increased strain at unpredictable times and for unpredictable reasons, making the challenge of effective capacity planning even more daunting. Analytics-driven network planning and capacity management solutions can play a pivotal role in combating this problem. Through a proper understanding of overall network usage and particular usage per application, operators can engage in asset allocation in a more informed fashion. In an environment where operators are often forced to overprovision the network to the degree that it is running at as low as 30 percent of peak capacity, this more intelligent capacity planning can be a critical tool in providing an optimal subscriber experience while maintaining (or reducing) capex. It is imperative to keep the customer experience at top of mind when using network-based data for either future network resource allocation or ongoing maximization of existing provisioned capacity. Operators believe that advanced predictive analytics-based planning solutions will play a pivotal role in helping them meet their business objectives. In a recent HEAVY READING FEBRUARY 2014 WHITE PAPER ADVANCED PREDICTIVE NETWORK ANALYTICS 4
survey of 70 global operators conducted by Heavy Reading, operational planning, real-time service assurance and product optimization are key areas where operators believe advanced predictive analytics can play an integral role in meeting their business objectives. Figure 2 illustrates key findings from that survey. Figure 2: How Do You Plan to Use Advanced Analytics to Deliver on Defined Business Objectives? Source: Heavy Reading, 2013 Intense competitive pressures, technology changes and service convergence have resulted in network growth. However, opex and capex continues to remain high, and the challenge to control and manage information for all interested parties (i.e., network operations, finance, planning) keeps increasing with time. Current problems associated with accurate operational planning include: Network either over- or under-built, with negative impacts for capex/opex as well as business flexibility to support service/customer growth effectively Lack of input from marketing on sales forecasts, service take-up trends Ongoing requests for financial data and financial modeling requires reliance on technical staff Planning is often very high-level, not utilizing detailed cell-level network data; spreadsheets need to be replaced with comprehensive analytics Hence, building out excess, unneeded capacity and the inability to forecast when network build-out is essential can put immense pressure on operator capex and result in inefficient use of expensive assets. The only way an operator can remain competitive in today's complex, hypercompetitive communications environment is by optimizing its network assets usage. Operators need advanced predictive analytics to federate and correlate data from multiple network data repositories, as well as sales forecasting systems such as CRM. This will provide operators with: The ability to plan, predict and optimize their investment in network build and rollout, identify potential stress points Optimized network investment plan based on service forecast demands HEAVY READING FEBRUARY 2014 WHITE PAPER ADVANCED PREDICTIVE NETWORK ANALYTICS 5
The ability to anticipate and implement necessary network changes just ahead of the demand curve An analogy can be drawn with manufacturing and retail companies such as Toyota and Wal-Mart. These companies maintain a very lean inventory and hence sustain their competitive advantage based on accurate resource utilization and their cost drivers. Why can't the communications industry adopt the same principles and optimize their resource allocation capability? Why do service providers still believe in creating network capacity based on a "just in case" (JIC) model, rather than adopting "just in time" (JIT) concepts like the manufacturing industry? The reasons lie in the service provider's traditional stovepiped approach to network planning, execution and capacity management. To be effective, all operational functions (marketing, finance, network operations and planning) need on-demand access to network resource data. To get meaningful information, this data requires ongoing input from all stakeholders as well. What is needed is a holistic approach to operational planning, which needs to take into account all dependencies that go into the planning process. Figure 3 illustrates different scenarios and parameters that need to be handled by analytics solutions to streamline network capacity management and resource utilization processes. Figure 3: Scenarios & Parameters Handled by Analytics Solutions Source: Heavy Reading, 2014 HEAVY READING FEBRUARY 2014 WHITE PAPER ADVANCED PREDICTIVE NETWORK ANALYTICS 6
Advanced predictive analytics solutions need to interface with network inventory solutions, service activation solutions, network discovery information and via service modeling and correlation of utilized resources need to help in accurate operational planning by predicting network resource exhaustion in a timely manner. It needs to drive capacity optimization and provide network planners with the ability to create "what if" scenarios based on past utilization trends, sales forecasts and service consumption trends. The challenge of the next-generation network is to accurately map network changes as they happen more rapidly because of myriad complex services. Hence successful planning needs to have visibility and control over the end-to-end processes to resolve exceptions and have the capability to accurately plan across multi-layer and multi technology domain. Advanced predictive analytics should be able to: Provide accurate "just in time" (JIT) network information will accelerate the provisioning success rate. Provide operators with the capability to predict and optimize network investment requirements and provide network engineers with the tools to optimally locate point-to-point routing demands from the traffic forecast. Based on past capacity utilization, marketing demand forecasts as well as service consumption trends can provide network operations the tools to efficiently plan, process and predict network growth. For service providers to go beyond just simple marketing data based merely on customer's home address when it comes to accurate capacity planning. For better accuracy, need to go down to the cell level, busy hour (or 15 minute) and also look at handset technology (2G, 3G, 4G). Automatically determine the forecasting models that are most suitable for the historical data. Generate an appropriate model for each item being forecast based on user-defined criteria, and model parameters are automatically optimized. Handle any number of business drivers and supplied events which should be automatically considered for inclusion in the models. Enable network planners to test "what if" scenarios, such as how changes to pricing or promotions, and determine their likely effect on future network capacity demand, which is very useful in designing sales and marketing programs to help proactively drive customer demand into more favorable patterns. We believe today's network planning activities need to support revenue generation and customer experience initiatives, going beyond their comfort zone of resource optimization. Combining subscriber, location and billing information will help mobile operators build a subscriber-centric view that can deliver value to their customers. By understanding network resources and transaction data with subscriber-specific information such as device, location, etc., these solutions can help generate different pricing schemes and targeted promotions for each individual customer, which will help optimize pricing by time, geography and customer profile. The notion of subscriber and plan revenue as it relates to network capacity can be an extremely valuable solution for operators that can also provide valuable input to capital allocation in tactical capacity planning. HEAVY READING FEBRUARY 2014 WHITE PAPER ADVANCED PREDICTIVE NETWORK ANALYTICS 7
Preemptive Service Assurance It is a known fact that service providers have no shortage of network data but very limited insight into this data. The transactional systems most have in place work well at pinpointing critical outages when they occur; rarely they were designed to allow network performance to be analyzed over a long period of time to understand how and where service issues are trending, how they are impacting one's most profitable customers, and how they might ultimately impact customer retention. With fixed or dwindling capex budgets, network analytics will play a greater role in decision-making as service providers are under extreme pressure to boost network performance and improve customer retention while reducing costs. As we have mentioned before, network lifecycle events have a large impact on customer satisfaction, as there are many moving parts that will impact customer experience. The residual impact of suboptimal network performance can result in longer call center calls, higher customer support cost and unsatisfactory customer experiences. Optimized and integrated data can enhance customer experience. Preemptive Service Quality Management In Heavy Reading's opinion, preemptive service quality management (SQM) means responding to network issues based on SLAs; measuring and adapting delivered services through real-time analysis of streaming data direct from network elements and consumer devices; and proactively enhancing handset and software performance by analyzing performance based on device type and software load. Advanced predictive analytics in many ways precedes that of the monitoring and maintenance element of preemptive SQM. It encompasses a range of functions and capabilities that would allow service providers to map service and customer commitments to capacity and service delivery. That would in effect create optimization between available network resources and services. It would extract and synthesize service components to create unique service bundles that are correlated to quality parameters. Advanced predictive analytics coupled with preemptive SQM has the ability to improve operational efficiency by focusing the operations staff on problems with a large business impact. For example, when a service provider has a brownout or blackout, the question becomes: "Which outages are more important?" In the new environment, the answer is: "Which outage carries the traffic of my best customers?" Therefore, another perceived benefit of this approach is its ability to focus on the pieces of network infrastructure that are delivering the highest value to service providers' premium customers. Figure 4 illustrates how network and customer information helps to provide real-time insight and help in preempting service quality issues. Data from network correlated that with location, profitability and service status and behavior of the customer can provide service providers with unparalleled decision-making capability. Operators can easily monitor individual subscribers and corporate customers and their transactions on corporate access point nodes (APNs). Operators can get immediate information on provisioning and configuration issues of corporate subscribers trying to access APNs. Operators can also identify a potential MSC failure and reroute traffic of their most valuable customers to a different MSC to avoid service degradation while notifying their less profitable subscribers of potential problem and providing them with estimated timeframe when the connection and service will be restored. HEAVY READING FEBRUARY 2014 WHITE PAPER ADVANCED PREDICTIVE NETWORK ANALYTICS 8
Figure 4: Combining Network & Customer Information to Provide Preemptive Service Assurance Source: Heavy Reading, 2014 By combining and comparing dropped calls, service metrics, latency for video based services, etc., with best-practice KPIs and connecting them with subscribers' dynamic and static information, operators can identify cell towers, MSCs or HLRs that are performing poorly and impacting the service experience of their VIP customers. This insight will help service providers to take preventive actions such as capacity increase, network upgrade, use over-the-air to update device patch, etc., before high-value customers' experience is negatively impacted by potential service degradation or service failure. This approach can enable operators to analyze, provide better insight and visibility over time, evaluate network performance and quality of service from a customercentric perspective, enable operators to take preemptive measures and help in answering questions such as: HEAVY READING FEBRUARY 2014 WHITE PAPER ADVANCED PREDICTIVE NETWORK ANALYTICS 9
Which regions in my network had the most dropped calls in the past hour, day, week, month, year, and which of my customers were most affected? Are these customers profitable and what is their likelihood to churn? Is my network performance breaching SLAs agreed with certain customer segment? How can I prioritize their traffic in order to avoid SLA breaches that will result in penalties? Which of my customer outages were due to handset problems, wireless coverage problems or switch problems? How can I prioritize where I should invest new capacity in my network based on real customer revenue and profitability impact? In which zone are my customers' calls being routed to another network operator most often, costing me fees? Can you deliver services that adhere to customer experience-based SLAs? Preemptive Approach to Issue & Service Management Advanced predictive analytics should be able to correlate information about the customer that is available from the surrounding systems, networks, social networks, etc., and be able to trigger certain actions to prevent problems before they occur and help maintain consistent user experience. Unlike traditional business intelligence tools, which stop at producing anomaly reports for postmortem purposes, the main advantage of advanced predictive analytics is that it is action-oriented and can help drive and streamline actionable decision-making for service providers. Advanced predictive analytics can help in preemptive service assurance so that latency-sensitive services (such as those based on video content) can be fixed before any fall in QoS impacts subscriber experience. Advanced predictive analytics should enable service assurance solutions to perform complex root cause analysis. This will help operators take into account a range of factors that may not be visible to traditional network management systems. It may help unravel complex problems such as: Correlating poor performance for subscribers all served by a single server Sufficient bandwidth is being delivered, but service degradation is a result of original encoding Turning up quality of service in one area has unintentionally impacted on quality of another type of service Advanced predictive analytics can help diagnose suitable fix strategies based on experiences that are extrapolated from other networks. This predictive approach will help to prevent problems before they arise or at least before subscribers notice them. HEAVY READING FEBRUARY 2014 WHITE PAPER ADVANCED PREDICTIVE NETWORK ANALYTICS 10
Control & Optimization It has been always been a challenge for mobile operators to isolate, identify and fix service problems with pinpoint accuracy. In a wireless world where subscribers are always on the move and consuming services indoors or outdoors, networkcentric information cannot capture all of the nuances that can affect service quality. Mobile operators need solutions that provide them with an analysis capability that captures all related radio and subscriber-level information into a single enterprise geolocation platform that can help remove the guesswork involved in fault isolation and reduce mean time to repair. Let us take a look at some of the control and optimization scenarios where advanced predictive analytics can play a critical role. Intelligent Wireless Offload Advanced predictive analytics solutions that are capable of combining data from remote cell site monitoring solutions (across various generations of network), DPI systems, customer usage systems, backhaul network management systems and subscriber data repositories can be used in real time to push different types of traffic belonging to different types of customers to different cells, depending on their subscription levels, the applications they are using and according to the traffic loading on different cells of different types. This might go beyond cell load balancing, but also take into account the variety of different backhaul technologies available to different cells (e.g., microwave from a cell site, versus DSL from a home femtocell), the cost of getting traffic to or from its end point (e.g., by getting Internet traffic off the wireless network as fast as possible) or a combination of the QoS and the cost (by using backhaul routes closer to locally cached popular Internet video content). In the context of 4G, where Wi-Fi offload is a common phenomenon, contextual intelligence can, in the case of congestion, correlate customer information with their lifetime value, spending pattern, type of services they are running, servicelevel agreements (SLAs) attached to that particular customer, etc., and intelligently decide which subscribers should be offloaded on Wi-Fi. In most cases today, operators do this blindly and risk breaching SLAs and dissatisfying their high-value corporate customers. Accurate Small Cell Placement Utilizing small cells effectively can provide operators with effective ways of delivering the increased data capacity required to satisfy the traffic demands of modern smartphones. However, the introduction of small cells increases network complexity and requires a different type of management from traditional macrocells. Controlling the heterogeneous network is a challenge for most operators, and the market today demands analytics-driven solutions that can provide guidance about where the cells should be positioned in the network. These solutions should be able to provide detailed information about how small cells are performing, exactly where they have been deployed and where else they are needed. Solutions should be able to incorporate call trace data, with subscriber, device and service data combined with location data to identify where small cells should be deployed, highlight network demand both indoors and outdoors, and identify the devices and subscribers responsible. HEAVY READING FEBRUARY 2014 WHITE PAPER ADVANCED PREDICTIVE NETWORK ANALYTICS 11
RAN Congestion Management RAN congestion has emerged as a major problem for mobile operators. Solutions that incorporate subscriber information with their services and location information can provide visibility at individual sub-cell level and provide priority to certain subscribers based on their tiers, etc., when they are moving across certain cells that are suffering from congestion issues. Since congestion events are often fleeting, making use of historical information about congestion from OSS systems to preempt similar problems and deploying RAN congestion only in those areas where congestion is anticipated are areas where operators are demanding mature solutions. Capacity management with the aid of advanced predictive analytics has the potential to play pivotal role in this context by being able to receive real-time feeds from network and back office systems to adjust network resources, load balance traffics and also help in adjusting customer's services, rate plans, tiers, etc., so as to help service providers maximize revenue from network assets. Intelligent Load Balancing When groups of users move from under one LTE cell to the next, the distribution of network load can change dramatically and some cells may overload, leading to service degradation or even dropped calls. Given the fact that adoption of SON is still in its infancy stage the load balancing mechanisms which operators have today are inefficient, manual and fraught with inaccuracies. An advanced predictive analytics-based network planning solution can not only help in understanding the load on cells by correlating traffic information and subscriber information to understand the current situation, but also perform "what if"-type sensitive analysis to forecast the growth of traffic in the near future. Based on such calculations, it can extend the range of unloaded cells to take traffic load away from a neighboring overloaded cell based on profitability of subscribers, traffic pattern, location, etc. With expected growth in data traffic this load balancing will help service providers to optimize their network resources, make sure services can be delivered with expected quality of experience with a view on subscriber profitability. HEAVY READING FEBRUARY 2014 WHITE PAPER ADVANCED PREDICTIVE NETWORK ANALYTICS 12
Customer Experience & Marketing Operators also have to make judicious capacity planning decisions in terms of the impact of over- or under-provisioning particular assets on the customer experience and how that might impact customer profitability and customer churn. Moreover, as large-scale operators entertain wholesale offerings for virtual network operators and other alternative service providers, an accurate view of real network capacity, both current and forward-looking, can illuminate the viability of this approach. Rather than guessing at the demand picture for a wholesale service of a certain virtual network operator, the use of granular analytics enables the wholesaler to understand exactly what resources are available. Operators have to couch capacity planning in terms of the impact of over- or under-provisioning particular assets on the customer experience and take into account subscriber-level information such as what services they are using, device info, location info, etc., while charting their planning process. It is imperative to keep the customer experience at top of mind when using network-based data for either future network resource allocation or ongoing maximization of existing provisioned capacity. When groups of users move from under one LTE cell to the next, the distribution of network load can change dramatically and some cells may overload, leading to service degradation or even dropped calls. The concept of customer experience management can be broken down into four dimensions: anticipating, controlling, responding to, and optimizing the customer experience. Since communications services operate in real-time, the impact on customer experience is more immediate than in many other industries. Thus, there is the need to anticipate poor customer experience in real-time or near-real-time, so that it can be dealt with proactively before it becomes a problem, to control the customer experience in real-time and to respond and even optimize the customer experience in very short timescales often seconds, minutes or hours compared to other industries, in which timescales may be days or weeks. Figure 5 is based on a recent Heavy Reading survey of 70 unique global operators which highlights the key activities that service providers believe is critical for superior customer experience. Service providers acknowledge that they need to understand their customers' individual experiences in a timely manner (82 percent) and on a continuous basis as they interact with their organizations both on- and off-net. This puts service providers in the best position to manage all the moving parts that can affect the customer experience, including the network, services, subscriber devices and contact center interactions. The ability to capture and understand the events that can affect the customer experience in time to influence and optimize that experience is a major source of differentiation for service providers. From a service providers marketing teams perspective, it is critical to have accurate and up-to-date information that will enable them to confidently launch, extend or modify campaigns. The ability to get up-to-date information on the existing network footprint's ability to adequately support a service launch can be critical for service profitability. The ability to accurately assess costs for launching any service, deciding whether a particular service is profitable or not, and making sure that any sudden surge of service demand catalyzed by aggressive promotion does not negatively impact profitability are also critical aspects of successful and profitable service delivery. HEAVY READING FEBRUARY 2014 WHITE PAPER ADVANCED PREDICTIVE NETWORK ANALYTICS 13
Figure 5: Activities Service Providers Believe Are Crucial to Deliver Superior Customer Experience Source: Heavy Reading, 2013 HEAVY READING FEBRUARY 2014 WHITE PAPER ADVANCED PREDICTIVE NETWORK ANALYTICS 14
Vendor Analysis: SAS's Network Analytics Vision One vendor whose real-time advanced predictive analytics solution can really meet the critical needs of service providers is SAS. SAS understands that service providers have a wealth of knowledge hiding in their network data which service providers continually capture and pay to store. Figure 6 illustrates SAS's vision of network analytics. Figure 6: SAS Data-Driven Network Decisions Reduce Churn/Improve NPS Source: SAS Institute, 2014 Unlocking network data's full potential using SAS for network analytics will provide service providers with new insights: Ensure continually good network service. SAS applies sophisticated analytics to traditional network data, so that performance can be visualized in a new way. Service providers with detailed know-how of their best- and worst-performing cells, nodes and switches will recognize that most issues are related to configuration, not hardware failures. They can get a view of all their network elements and uncover potential problems long before the alarms go off. Visibility of network from the customer's perspective is critical, so that service providers can work toward reducing drop rates and increase data speed. Plan capacity effectively. The days of Erlangs, busy hour and simple networks are gone forever. Today's networks have multiplied in size and complexity, and data doubles about every nine months. For accurate capacity planning, service providers need advanced statistical forecasts based on actual network and handset data so they will know exactly what they need and when, and can use just-in-time growth to maximize their revenue. SAS can provide service providers with accurate know-how such as: HEAVY READING FEBRUARY 2014 WHITE PAPER ADVANCED PREDICTIVE NETWORK ANALYTICS 15
o o o Why LTE handsets are only getting 3G service on your LTE network? When and where your network is slow? What equipment is not performing and what is being underutilized? Improve customer experience. Use detailed, up-to-date network data to empower service providers call center agents by showing them each customer's true quality of service. They can see the traffic patterns for each customer to understand speed and drop issues. They can also visualize how one customer's experience compares with others who have the same handset or live in the same region, quickly pinpoint issues at the cell or node level, and predict network failures before they happen. Then optimize the network from the customer's perspective, and spend less time troubleshooting to resolve problems. Target your marketing precisely. By examining network usage and customer equipment data, operators can make decisions that help reduce churn and increase Net Promoter Scores. Data can be shared with dayto-day decision-makers guiding them in how to sell services and features that fit customer usage. Accurate information can help service providers to sell only where you have capacity. It will also help service providers to offer more relevant promotions whether it's a faster speed for customers who like to watch movies, or an international plan for a world traveler. A comprehensive big network data solution from SAS combines data quality, advanced predictive analytics and visual analytics with expertise obtained through many years of working with service providers. SAS's solution will help service providers to capture the full value of their network data and utilize analytics to make smarter decisions to establish a sustainable operational model that can help service providers increase returns from all their network investments. HEAVY READING FEBRUARY 2014 WHITE PAPER ADVANCED PREDICTIVE NETWORK ANALYTICS 16
Conclusions Converting the deluge of information into actionable real-time information is an arduous task that service providers need to tackle if they want to meet their business objectives, which center on accurate network planning, providing preemptive service assurance and delivering superior customer experience. Advanced predictive analytics will play a pivotal role in the success of operators, as they will not only provide operators with real-time intelligence, but also help them to maximize their revenue potential from a short window of opportunity. We are in age of network consolidation and convergence. A lot of companies have driven growth by acquisition. It is important to have visibility into how the networks are modeled and connected, and be able to make traffic migration plans to move traffic from full topologies, individual sites or nodes with minimum impact to customer services and product campaigns. For carriers and service providers, operational excellence is the key mantra that will enable them to become more competitive. This implies that service providers need to optimize their usage of network assets and adopt a more accurate and timely approach to capacity management. Advanced predictive analytics will play a central role in accurate, realistic and proactive operational planning capabilities, which will not only enable correct sizing of the future network, but also help service providers to reduce capacity shortfalls, minimize order fallout and increase efficiency by identifying underutilized network resources. Next-generation network planning tools also need to evolve and support adjacent areas that involve revenue generation and assessment of operators' network profitability. With global economic conditions worsening and service providers struggling to reduce opex and accurately understand the profitability of the network, there is no doubt that advanced predictive analytics is well positioned to play a central role in service provider infrastructure. HEAVY READING FEBRUARY 2014 WHITE PAPER ADVANCED PREDICTIVE NETWORK ANALYTICS 17