The next telco battleground: Network experience and competitiveness

Declining ROIC, evolving customer expectations, and growing competition from hyperscalers will push operators to rethink the importance of the network.

In recent years, the source of competitive advantage for telcos has primarily been in network coverage and speed. Yet now, as many much-anticipated evolutions of telecommunications network technologies finally start to bear fruit, that focus is shifting to user-level quality of experience (QoE), deployment speed, and innovation in network services.

These technological breakthroughs range from advances in analytics, virtualization, cloudification, and open-network architectures to network management, service orchestration, API exposure, and higher achievable degrees of automation. They have the potential to help operators uncover new levels of efficiency and sources of revenue after years of declining ROI, which in certain cases is below the weighted average cost of capital (WACC).

They also, however, open the door to hyperscalers and digital attackers to compete more directly with operators, particularly in the B2B space. In this environment of increasing customer options, the QoE driven by network performance, deployment speed, and innovation pace only grows in importance. As a result, operators will have to seriously consider upgrading the network to become a competitive differentiator for both consumer and enterprise customers and to avoid becoming a cost-intensive commodity. In the process, it will massively drive operators’ value creation, potentially yielding 2 to 3 percent improvements in ROIC and EBITDA, reallocating up to 30 percent of capital expenses and reducing time to market by up to 30 percent.

Turning the network into operators’ next competitive battleground requires refocusing, though, as operators face the many strategic technology choices related to the network that will help determine their degree of success. In addition to the investment choices that operators face, they will also likely need to decide on the right transition path from their current architecture.

Drivers of change are accelerating

Finding the right path forward under these conditions requires a clear understanding of three major trends that will shape telco networks over the next three to five years: data-driven decision making based on user-level QoE; open, virtualized and cloud-based networks (softwarization and cloudification); and the move to increased active network sharing. All three are accelerating, as they enable one another.

The imperative of customer-level QoE

Individual customer experience is the primary differentiator for operators. Network experience accounts for 40 percent of new users’ key buying factors and churn; price and customer service are second and third drivers (Exhibit 1). Even if price still drives a large share, competing on price is not sustainable for the profitability of the industry.

Network experience is one of the leading factors in telco customers’ decision making process.
We strive to provide individuals with disabilities equal access to our website. If you would like information about this content we will be happy to work with you. Please email us at: McKinsey_Website_Accessibility@mckinsey.com

User-level QoE is critical to optimizing network capital investments and increasing ROI, but it requires effective measurement of customer perceptions for the services they use and value, and the traditional approaches for doing so fall short. Many operators are still unable to go beyond generating aggregate network metrics, both for wireless and wireline networks. This limits operators to allocating capital based on metrics that are disconnected from variations in user-level network experience, such as average speed, marketing claims, or average revenue per user (ARPU).

Without user-level QoE data, an operator can’t know whether they’re allocating capital in the right place. For example, an operator might allocate capital to a micromarket where the QoE is already high despite experiencing lower than average speed than competitors, and overlook a higher ROI micromarket that has a larger proportion of users experiencing low QoE, despite the average network performance in that area being above target thresholds. Such a misstep would lead to lower gross adds (additions of new subscribers) and higher churn. Having user-level QoE data in hand, by contrast, has resulted in up to 30 percent increase in ROI for operators through lower churn, increased gross adds, and overall capital expenditure reductions.

Furthermore, user-level QoE scores can be used to improve the care experience and sales. For example, the scores can enable care queue prioritization, targeted communications aimed at managing poor QoE, and personalized offers, such as offering new services to only those customers who already have a high QoE. One operator learned this the hard way by experiencing unusually high churn after marketing a streaming package to users with poor QoE—such an offer for additional services can backfire if the network can’t deliver the new offering effectively, increasing customers’ frustration with the operator.

Softwarization and cloudification of the network

Network softwarization is enabled by three key technological shifts: network virtualization (including network function virtualization, or NFV, and virtualized radio access network, or vRAN), which decouples network functionalities from the underlying hardware; software-defined networks (SDN), which separate the control plane of a network from the user plane and centralize it; and standardization and openness of network interfaces and API.

These advancements are forcing changes that will reshape entire aspects of the industry in ways that were unthinkable just a few years ago, opening the door to faster innovation and greater agility that is likely to result in a reshaping of the vendor ecosystem and increased innovation that looks more like the software industry and other digital-native sectors. Softwarization also comes with additional operational benefits, including higher potential for automation, flexibility in resource utilization and scalability, and advanced redundancy and management features already standard in the cloud and data center world.

Reaping the benefits of softwarization, however, is not as straightforward as it might seem. For example, many operators have seen costs increase with NFV, while vRAN setups are facing energy-efficiency challenges, and the business case for an open radio access network (oRAN) is not a one-size-fits-all solution. Real-life implementations of such modern architecture concepts have yet to prove their financial advantage when deployed at scale. Operators would be wise to take an approach rooted in value creation for customers and procurement excellence to offset considerable swap/migration and systems integration costs.

Moving telco network functions to the cloud, in conjunction with softwarization, gives operators a further opportunity to potentially improve financial performance. Certain areas of core network and RAN baseband processing could be run on cloud infrastructure, allowing operators to reduce upfront hardware costs and instead opt for a “pay as you grow” model that scales with business growth. To take advantage of computational load variability, capacity could be dynamically scaled up and down, thereby avoiding the need for site-level permanent provisioning of capacity that is only required for certain peak occasions. Infrastructure management, operation costs, and complexity and could be streamlined, helping to increase focus on value-creation drivers (Exhibit 2).

Operators have cost saving expectations for oRAN, vRAN, and cRAN across multiple areas of network deployment and operations.
We strive to provide individuals with disabilities equal access to our website. If you would like information about this content we will be happy to work with you. Please email us at: McKinsey_Website_Accessibility@mckinsey.com

Cloudification could also unlock new asset ownership and deployment models. For instance, a third-party infrastructure provider could offer operators generic and highly efficient turnkey hardware solutions (for example, power, HVAC, backhaul, and radio hardware; baseband hardware; and core hardware) on which operators could run their specific software solutions to serve their customers. We already see parts of this being used in more-traditional infrastructure-sharing business models.

Two operators are already working to realize the potential of softwarization and cloudification. Rakuten is pursuing capital expense efficiency and building its own private cloud using a low-complexity hardware design. With extensive use of vRAN, it is opting for heavy standardization of its individual network components, which allows a higher degree of network automation. Dish is focusing on operational cost efficiencies, and has partnered with Amazon Web Services (AWS) to build a new architecture based on hyperscaler infrastructure and an automation framework around a softwarized stack. Rakuten’s approach of building its own cloud is still quite capital intensive; Dish’s approach is lower intensity, but requires crafting partnership models that are new to the industry.

The outcomes of these two approaches are still unknown, as it is not yet clear whether public or private cloud will turn out to be the most efficient and under which conditions. As part of the public cloud approach, for instance, hyperscalers will need to adapt their pricing and operating model to the high fluctuations involved with real-time mobile traffic. On the other hand, the private cloud approach requires operators to develop extensive new skills and evolve their operating model. Both approaches will have to prove they can lead to a documented increase in average revenue per user (ARPU) and lower total cost of ownership.

Our analysis shows that operators could gain up to 20 percent in total-cost-of-ownership savings over legacy deployment by leveraging softwarization (including oRAN) with cloudification over legacy deployment cloudification, in the right conditions and context. This promised efficiency is the main reason why, in our market modeling, we believe network virtualization will reach a share of about 45 percent in core and around 35 percent in RAN over the next five years.

Network active sharing

Operators will continue turning to sharing as a key lever to significantly reduce cost and improve deployment time or network coverage. Fifth-generation (5G) deployments, which can lead to significantly increased capital expenditures in scenarios that require densification, are accelerating this trend. The valuation benefits that TowerCo models can bring are having the same effect. Network sharing will likely be a standard tool for most operators to optimize their rural network spend and multiple business models will coexist to satisfy the demand for more-active network sharing and edge computing (vRAN baseband).

Mobile network sharing is a common model for operators in Europe. Fiber sharing is increasingly being adopted across the globe, as investment returns are much less attractive because of fiber overbuild. In some cases, even governments are pushing the industry to network sharing.

Operators can achieve 30 to 35 percent savings from doing active network sharing, including spectrum, on mobile. The speed of new technology rollout can be also doubled when two operators share the network. In the past, 4G rollout has driven many network-sharing deals.

Defining the path forward

To be ready for this new era, operators could determine what skills they will need and make appropriate build-versus-buy decisions for the future. In the meantime, they will still have to navigate the current ecosystem and try to achieve capital efficiency improvements, including a better way to allocate investment where it matters most to customers, and a faster way of building out their networks, all while continuing to lower unit costs.

Shake up the way capital is deployed

Operators should consider rethinking their road map for both wireless and wireline in terms of actual customer value. Rather than pegging capital to support marketing claims centered on speed and newer technology coverage, operators could allocate capital to those parts of the network that actually affect customers’ quality of experience, such as using user-level QoE to target the 20 to 30 percent of users with low QoE, and launching new products that customers actually value. For example, one European operator optimized both the timing of its migration to 5G stand-alone architecture and its edge infrastructure deployment by accounting simultaneously for customer value creation and technology readiness, which resulted in a more than 15 percent net present value improvement.

Capital expenditure goes hand in hand with operating expenditure. For example, Rakuten’s nearly vendor-agnostic model is almost completely virtualized, with a low-complexity hardware design to reduce site complexity and to accelerate deployment. Such a setup not only promises capital expenditure savings but also operational budget efficiency when being scaled. Whether the expected efficiencies in terms of operating and capital budget total cost of ownership will materialize for other operators, and whether such an architecture will bring benefits in a brownfield setting depends on the specifics of the operator pursuing it.

Leverage partners efficiently

Operators have long operated largely on their own, only drawing on suppliers to define their road map and to build out traditional networks. But a softwarized and cloudified network requires operators to partner closely with hyperscalers, system integrators (SI), and new software-as-a-service (SaaS) vendors. This requires a complete new governance model to ensure that partner costs and delivery delay do not cancel out expected benefits. For example, operators have managed these partnerships efficiently through agile cross-functional sprints with a single end-to-end business owner, transparent business-driven deadlines with clear value implications, and process automation for rapid testing/feedback.

Augment current talent and operating model

Whether the future of the network lies in the public or private cloud, large chunks of network function will soon live in the cloud, and operators’ workforces will need to be capable of managing them. Operators will greatly benefit from operating models that fit their technology choices, and talent is a key piece of determining what to build and what to buy. Until recently, operators’ network engineering talent needed to specialize in radio capabilities—planning and optimizing the RAN, with many other aspects being outsourced. But as the network becomes virtualized, differentiation will depend on a wider range of tech skills and capabilities, including expertise in automation, platforms, and network asset management.

Operators will also want to rethink which activities they want to keep in-house to focus on the value-creating sources of differentiation and to avoid becoming overly reliant on system integrators, which could end up eating the savings generated by virtualization.

The sheer volume of talent that operators will need to build and scale cloud infrastructure is unlikely to be available in any one local market, so telcos may want to set up talent hubs in a variety of regions and consider acqui-hiring. Building the right cloud architecture takes specific processes and skills—for example, scaling up cloud infrastructure requires modern DevOps practices and organizational methods, for which telcos need to acquire cloud engineers and associated specialists across regions. The demand for such talent will be intense. Many operators are looking into building up competence centers for cloud network engineering with sizable head counts, and doing so in Eastern Europe due to higher availability of relevant talent.

The way work gets done is also critical. Just as IT organizations had to embrace agile and other digital methods of working over the past decade, telcos should strongly consider enabling an end-to-end digital operating model, including API and data governances to drive the agility and automation required to rapidly innovate and market new products and services. Operators slow to build these capabilities risk losing chunks of their market to hyperscalers and digital attackers.

Simplify deployment process and tools

Current processes will not be enough to keep up with operators’ fiber and 5G ambitions. Digital and analytics can be leveraged to streamline deployment journeys, enable productivity gains for operators and for suppliers, rationalize tools and architecture, and create a unified view. This can result in 30 percent faster deployment time, 8 percent lower costs, and improved user experiences both internally and for suppliers).

Rethink supplier strategy and model

Despite labor and material shortages, operators will need to rely on unit cost reduction as they accelerate 5G and fiber deployment. To both secure supply and achieve unit costs, they will also need to holistically rethink their supplier strategy to generate efficiency gains and to reduce their business risks. For example, network deployment labor, which represents the largest network costs bucket, can be tackled in a number of new ways to create efficiencies and reduce business risks. These ways include uniformizing margins across services, geographical exclusivity, improved short- and long-term forecasting, new terms and conditions where both parties make commitments to each other, and leveraging advanced analytics award optimization to maximize efficiencies. Operators leveraging this playbook have been able to achieve historical unit cost reductions while securing labor despite market momentum projecting double-digit increases.


With network differentiation shifting to user-level quality of experience, softwarization, and cloudification, operators will need to fundamentally change their operating model. They can no longer rely on their OEM for their technology road map—there is no standardized blueprint for outsourcing vendors for key technology skills such as those possessed by developers, cloud engineers, and data scientists. Operators will also probably need to build new in-house competencies just as strategic as radio frequency engineering to define their network of the future and transition plan, evolve their planning process to leverage analytic-powered user-level QoE, and manage multiple partners to capture value. There is no time to waste, though, as undergoing the transformation required to achieve these gains is usually a journey of 12 to 18 months.

Explore a career with us

Related Articles