Telecom operators are investing heavily to connect more people to high-speed fiber networks. In 2022 in Europe and the United States alone, their networks reached 15.2 million more households than the prior year, compared with the 9.9 million they added in 2018.1 Incumbents, newly formed fiber companies, governments, and private investors are all active in the sector, to the extent that more than one billion homes worldwide now have high-speed fiber access.2
Still, that figure means that some 40 percent of the world’s population remains without fiber access, indicating significant potential for further growth in the fiber-to-the-home (FTTH) market. The extent to which companies capture that potential, however, could hinge on their ability to improve returns on invested capital, as new markets are likely to be those with relatively high costs to serve and low revenue-generating potential. The most attractive markets already have fiber connections.
Reducing costs wherever possible is therefore crucial. But maximizing market share early on is critical too, and this depends upon speed to market. Experience suggests that the first FTTH operator to enter a market—whether overbuilding existing networks with fiber or rolling out new networks—can win significant market share as customers switch from cable and DSL providers. A second, later entrant, however, is likely to capture a significantly smaller share as the pool of customers willing to switch shrinks (Exhibit 1).
Of course, government policies and regulations will play a role in shaping market outcomes. To encourage high-speed access, governments might, for example, allow competitors to deploy fiber on incumbents’ ducts and poles, provide financial incentives to create wholesale fiber infrastructure companies, or offer subsidies to build fiber in rural and remote areas.
But companies’ own operational capabilities will also go a long way in determining how they fare. In our view, there are four main ways in which incumbents and greenfield companies can improve their internal processes to make their network rollouts both faster and leaner:
- Use artificial intelligence (AI) to target low-cost markets with high potential penetration levels.
- Establish an operating model that drives coordination and efficiency.
- Automate resource-intensive functions.
- Strike long-term partnerships with suppliers.
Our experience suggests the combination of these tactics can accelerate deployment by as much as 20 percent while lowering costs by between 15 and 25 percent, with results discernible within eight to sixteen months (Exhibit 2).
Use AI to target low-cost markets with high potential penetration levels
Companies looking to roll out new fiber have typically targeted markets likely to yield the highest revenues, which has often meant focusing on urban areas where population density and household income levels are high. But as more fiber is laid, these high revenue-generating opportunities are becoming scarce, requiring companies to put greater weight on the cost to serve a market. There’s little point in targeting an area with high population density if penetration rates are likely to be low because of low household income. The cost to serve could prove prohibitive.
Instead, companies will want to identify markets where they can find the best balance between penetration potential, average revenue per user (ARPU), and construction costs. This can be a complex and time-consuming task given the multitude of variables, beyond socio-demographic ones, which are likely to impact the balance. These include the length of network routes required, deployment methods (aerially on poles, using existing underground ducts, or trenching new ducts), the availability of high-speed backbone infrastructure, equipment and labor costs, and competitors’ deployment plans.
AI-assisted planning models are adept at handling such analysis and can improve outcomes. The experience of one operator illustrates the point. Its original fiber build strategy, formulated by an experienced team, underperformed. Costs were higher and take-up weaker than anticipated. For the next deployment phase, the company opted to use an AI-assisted planning model that considered a huge range of variables; these included the distance of every residential building from the nearest fiber route, a factor that impacts costs, and the number of wireless subscribers in each area, a factor that impacts take-up as wireless subscribers often upgrade to fiber. As a result, the model identified markets where the cost of getting fiber to homes was 5 to 7 percent lower than the markets initially targeted and potential penetration levels up to 10 percent higher, significantly enhancing forecast returns.
Building an effective AI planning model starts with getting the basics right. All financial, customer, and network data will need to be available. There should be no gaps and inconsistencies. From here, advanced AI models can forecast returns on network deployment at a granular level, looking at zip codes, city blocks, or even individual buildings, for example. And with the addition of more parameters, the model can analyze even more scenarios, such as where competitors might build fiber in the future, helping to minimize the risk of overbuilding, or the returns on alternative broadband technologies, such as fixed wireless.
Would you like to learn more about our Technology, Media & Telecommunications Practice?
Establish an operating model that drives coordination and efficiency
Fiber rollouts can only be as fast and efficient as a company’s operating model allows. By standardizing processes, setting up a harmonized IT stack, and establishing a “nerve center” to coordinate the project pipeline, companies can drive coordination and efficiencies within and across functions including network deployment, operations, IT, and procurement. These measures have the potential to cut deployment costs by a percentage point or two while simultaneously accelerating deployment.
Standardize nationwide processes
Many companies find they have different operational processes, procedures, and practices in place for different markets, as well as for different projects within the same geography, and even for different teams working on the same project. In some markets, companies outsource design, for example, but in others they conduct the work internally. In some, they award construction to a single company while elsewhere they divvy up the work. Some projects might have 20 milestones while other similar ones have 30, with comparisons made even trickier because of different nomenclature. This kind of variance means that best practices aren’t always followed, slowing the pace of deployment and cranking up construction costs as team members learn to adapt to diverse ways of working. It can also lead to mistakes, necessitating costly rework.
Standardization helps solve this. It entails the meticulous mapping of all steps in the deployment process, documenting all relevant best practices, and defining governance mechanisms so that any deviations from the standard are identified early. Regular audits, performance metrics tracking, and compliance checks all help in this respect. Change management is important too, as all employees will need to train to adopt newly established practices. Network design and architecture should also be standardized; that could include establishing the optimal distance between nodes, for example.
Harmonize the technology stack
Many operators find themselves saddled with cumbersome processes and high costs because of the legacy tools they use to manage deployment. This can be particularly problematic for incumbent operators with legacy copper or cable networks that are now laying new fiber ones, as well as for companies that have expanded through acquisitions. It isn’t uncommon to find companies with multiple different systems for managing inventory, projects, capital deployment and other resources. Consolidation of these tools is a lengthy process that requires a transition road map aimed at avoiding disruptions, but it also offers substantial rewards in terms of cost savings and deployment speed.
One North American broadband company that had different deployment tools in each of the regions where it operated found itself struggling with the quality of its data and rollout effectiveness. Inventory entries were missing or duplicated, for example, and there were manual, unreliable notes regarding the completion of milestones. A phased plan that eliminated redundancies and introduced a new tech stack with modern, out-of-the-box SaaS solutions alongside customized in-house built tools quickened the pace of subsequent fiber rollouts by almost 10 percent while also reducing deployment costs.
Create a nerve center
Problems inevitably arise when a company is deploying fiber in dozens of different sites and markets and orchestrating multiple vendors and material suppliers, threatening schedules and budgets. Yet business-as-usual processes, whereby individual project teams work without visibility across the entire deployment program, can be slow to detect such problems and their potential impact or to resolve them. For example, the fact that a particular supplier has failed to deliver on schedule to two sites might go unnoticed other than by those working on the sites. But the failure could be a sign that the supplier has run into significant problems and won’t deliver on time to the next ten sites either, seriously jeopardizing the deployment schedule. Likewise, a decentralized approach can mean work slowing to a halt in one place because of a supply shortage, even though the needed material is sitting in an intermediate warehouse.
A nerve center, by contrast, collects and analyzes data that gives a bird’s eye view of how deployment is progressing relative to milestones. This might include real-time updates on permitting applications, inventory availability, and overall construction timelines. With this information, the nerve center’s cross-functional team, which includes representatives from the various markets, IT, procurement, and data analytics, can work in an agile manner to resolve issues promptly, as they have both the capability and authority to mobilize resources and address them themselves, or to escalate them when necessary. The nerve center established by one operator detected a delay in the delivery of specialized fiber-optic cables from a key supplier. Procurement approached the supplier and ascertained the problem was unlikely to improve imminently, prompting the nerve center team to reallocate cable stocked in a warehouse in a different region—a move that saved further project delays and the associated costs.
Fiber opportunity: Four deal types for investors to consider
Automate resource-intensive functions
Companies need to manage thousands of tasks to deploy new fiber infrastructure. Add in human error, and it’s not hard to see why the process is often riddled with inefficiencies that slow progress and inflate costs. Automation can help remove some of them, particularly in the following areas.
Network design
Designing a fiber network is a complex process, entailing route planning, node placement, capacity and redundancy planning, simulation and testing, and more. The process can take many months, depending on the project’s complexity, topography, and the team’s expertise. But automation can help accelerate it in a variety of ways. For example, an automated process can generate optimal routes and node locations based on topographical data, existing infrastructure, and coverage requirements. It can forecast network capacity and simulate performance under different conditions, such as peak usage or unexpected network failures. Importantly, automation also frees up employees to focus on higher-value tasks—perhaps working on designs that might require bespoke solutions because of the unique nature of a construction site or validating other designs. Our experience suggests that automation can reduce the time it takes to design networks manually by up to 30 percent.
Site surveys
Automation can cut the time it takes operators and their contractors to survey sites for deployment. Data collection, terrain analysis, and obstacle detection, all once time-intensive tasks requiring in-person visits, can now be automated with the help of digital twin models that combine relevant public and proprietary data, including satellite imagery, geospatial databases, and elevation models. Site visits only then need to be made if data is missing or needs confirmation. Yet even here, companies can expedite the process with the help of computer vision models and lidar sensors that produce high-resolution 3D maps and models of the area. Such methods have helped speed up certain site survey timelines by between 20 and 25 percent.
Permits
Obtaining permits for fiber construction is one of the most time-consuming rollout processes, often running into weeks or even months. To a certain extent, the local authorities or utilities issuing the permits dictate timelines. Nonetheless, poor management by telco operators can cause unnecessary delays. All too often, company employees find themselves trawling through different websites trying to ascertain which data (and in which format) they need to submit for dozens of different permits for, say, digging a trench or interrupting traffic, then trying to keep track of the progress of every application. Even in companies where the process is more structured, it often remains manual. Establishing a standardized, automated permit center—one that understands regional permit requirements, collects and validates the necessary data, creates and submits the permits, and then tracks their progress in real time—streamlines the process. As a result, it can boost application approval rates by 10 to 20 percent and cut the time needed to acquire permits by as much as 15 percent, while freeing up employees to concentrate on verifying more complex permit requirements.
Construction management
Project managers commonly oversee as many as 500 distinct jobs on a project, each involving perhaps 60 or so individual tasks—these are hugely time-consuming processes that are almost certain to overlook existing or brewing problems. A real-time automated construction management system drives efficiency by establishing and monitoring exactly when all routine tasks and activities need to be completed, by whom, and in what order, to keep the project on schedule. If a milestone is at risk of being missed, the system triggers an alert. In this way, project managers can dedicate their attention to evaluating the impact of any delays and resolving them. In addition, many routine tasks can be automated—creating a bill of materials or dispatching a crew, for example—leaving the manager with more time to handle complex ones, be that optimizing the network topology for maximum efficiency or addressing unexpected challenges that arise in the field. A real-time construction management system can, in our experience, quicken the pace of deployment and reduce deployment expenses by 5 to 8 percent.
Strike long-term partnerships with suppliers
Building long-term partnerships with suppliers in lieu of transactional, one-off relationships can offer substantial rewards. Long-term partnerships help suppliers plan their production schedules and allocate resources in line with operators’ needs, giving the relationship a degree of stability and predictability that can often deliver cost savings greater than those likely to be achieved through short-term price discounts offered in a one-off, competitive bidding process. Within a partnership, suppliers may be willing to negotiate favorable terms, bulk discounts, and customized solutions that extend beyond a single project. By working in close collaboration with their suppliers, some operators have been able to secure a construction cost advantage of between 5 and 8 percent over those with a more transactional approach to supplier management, taking into account the cost of materials and labor.
Collaboration also shields operators from delays and disruptions because of a more secure supply chain—an advantage some operators have sought to maximize by changing their ordering practices. Rather than ordering materials a month in advance and hoping for timely delivery, they share supply forecasts up to a year ahead and commit to purchasing a sizable portion of that volume, perhaps up to 75 percent. The advance notice and demand certainty help suppliers meet their obligations, though penalties can still be applied if there are fulfillment delays.
Forecasts must, of course, be accurate, which will mean installing robust demand planning systems that incorporate reliable data and machine learning tools that are closely integrated with deployment and financial planning. Efficient inventory management systems are pivotal too. By offering comprehensive visibility into stock levels across all stages of the supply chain, these systems ensure the availability of the right materials at the right locations and times, increasing the accuracy of demand forecasts shared with vendors.
The speed and costs of fiber deployment are central to its viability in many of the world’s still unserved markets. The four measures described here—using AI to target low-cost, high-penetration markets; establishing an operating model that drives coordination and efficiency; automating resource-intensive functions; and striking mutually beneficial, long-term partnerships with suppliers—can help incumbents and new operators improve both measures. Within eight to sixteen months, the result could be both higher fiber market share and higher ROI, while helping to ensure that millions more people get access to highspeed connections.
