How can you plan for the unknown?
Network operators today have to grapple with this question at a massive scale. 5G encapsulates a wide set of topics. It’s not only a case of delivering wide coverage at a rapid pace in the face of higher network density and virtualization; there’s also the question of how to support a new generation of advanced connectivity services – each with its own unique set of performance requirements. In many cases, the coverage requirements themselves vary for different services too – some will require street coverage, others indoor coverage, others rooftop coverage, and this is before we look at different capacity, latency and throughput requirements. We often don’t even know what those services will be.
A shift in emphasis is needed, integrating business information alongside network information as part of the planning process, optimizing planning decisions for business KPIs, such as quality of experience (QoE), customer churn, revenue and, ultimately, return on investment (RoI). Of course, this isn’t to say that network KPIs like coverage and throughput aren’t essential too. In many cases we work with leading operators that do indeed have contractual obligations to deliver against specific targets in these areas for regulatory compliance.
Instead, we need a balance, one that takes account of a holistic understanding of the interplay between network capacity roll-out, the time and cost involved, and the likely impact it will have on incremental revenue.
For this to work in practice, though, several conditions need to be met. Any planning system that you use clearly needs to be able to flexibly ingest different data sources, but there is much more involved besides. For instance:
- Advanced AI/ML-based predictions, essential for the accurate forecasting of these more complex business outcomes, like what would happen commercially if you don’t invest and how much revenue could potentially be lost – don’t forget other factors are at play, such as competitors’ presence and activities, price plans and price sensitivity;
- Automation of analytics for complex decision-making activities, which involve considering options and trade-offs around form-factor, cell location, backhaul availability and technology options, and all of this alongside consideration of the business targets mentioned above; and
- Automation of roll-out scheduling, to optimize the ‘time-to-market’ (TTM) of networks, densifications and expansions, which is a critical factor for many operators that today find themselves in highly competitive, saturated markets, where growth still comes largely from the acquisition of customers that churn from other operators.
This all sounds complicated enough. But what happens when user requirements and service offerings change and become more diverse and complex? When landscapes and buildings change? When networking technology changes?
Network CAPEX is all about long-term investment, meaning that you must plan and make predictions for the best shot of a good return over time. But how can you approach this scientifically when everything is in a state of flux? How can you make sensible, data-led adjustments?
Like so many of the challenges associated with 5G, the solution lies with network lifecycle automation (NLA).
Earlier this month, we kicked off a new series of blog posts on this theme, the first being an introductory piece entitled 5 reasons why network lifecycle automation is important for 5G, which gave a preview of five of the many commercial and technical arguments for NLA.
First on the list was how NLA empowers you to invest CAPEX of your RAN accurately for ROI. In the second part of this series we expand on this topic of ‘smart CAPEX’, showing:
- The better business outcomes you can expect;
- How it works; and
- Ultimately, the value this brings to your bottom line across the network lifecycle.
But first, a quick note on how we define smart CAPEX.
Defining smart CAPEX
Smart CAPEX concerns the end-to-end service lifecycle process. That process starts when you plan and build a network from scratch, continuing through to the operational phase with network evolution – and beyond to fully monetizing your network.
In the beginning, you have neither network data nor performance management data. You don’t know where your potential customers are. And you have no dynamic picture. But you still need to build sites in the best location from the perspectives of TCO and RoI. In other words, you must ensure that each site will deliver the best QoE while maximizing your key business objectives.
This seems obvious. But what happens when things change? Because when they do, your initial hypotheses can become invalidated, and in the era of 5G, things change at a much faster pace, in terms of service categories, customer expectations and demand profiles. The need to make decisions quickly, to be agile and responsive, is more urgent than ever, with operators needing to marry together network planning and business outcomes.
This is where smart CAPEX comes in. Smart CAPEX isn’t a tangible thing; rather, it’s an analytical method that helps you adapt from greenfield development through to the operational network evolution stage.
And the right network lifecycle automation system will facilitate it, using the data it gathers, and artificial intelligence (AI) and machine learning (ML), to make predictions that allow you to adjust the focus of your investments over time, and automate the making of network investment decisions, allowing for much faster decision making, in response to much more accurate and comprehensive predictive insights, meaning your deployment plans are always aligned with your initial objectives. Let’s see how.
Smarter planning; better business outcomes
Planning your network – be it a greenfield deployment or densification and expansion – can be fraught with numerous challenges, such as:
- Network design and TCO/TTM/RoI analysis usually being organized as separate processes, one after another, and how to ensure each metric is optimally balanced;
- ‘Language barriers’ between CMO and CTO teams, which increases your ‘time to decision’;
- The dilemma of where to prioritize deployment;
- Suboptimal RoI and net present value (NPV) after deployment;
- Difficulties with end-to-end alignment of key business objectives (KBOs) and the deployment process;
- Limited visibility on future poor QoE and churn hot spots when your traffic grows.
With increased operator competition, churn alone has become a significant problem. Vodafone in the UK saw a rate of 64.2% among its prepaying customers in the fourth quarter of its last financial year, for example.
But running through these challenges is the common theme of uncertainty. Without certainty – or at least reliable predictions – how can you plan?
With smart CAPEX allocation and automation of your RAN, as provided in forward-thinking NLA solutions, you can reduce this uncertainty and apply predicted business outcomes. At its crux is the idea that you can make the best investments early on and adjust them over time according to ever-increasing collections of data.
Tapping into advanced AI and ML-based analytics means you can better understand the ways in which customers respond to changes in the quality, speed and coverage or your network.
By creating a link between your network and business outcomes, you can derive actionable insights on how customers’ behaviours, such as consumption, spend and churn, are impacted by the performance of the network, and the investments you make in improving it. This will inform your strategy for the delivery of revenue growth. Typical related KPIs that you can predict in this way include:
- Service demand prediction;
- Subscriber QoE prediction;
- Subscriber churn prediction;
- Network coverage and quality simulation; and
- Regional revenue potential and future revenue growth.
Through outputs such as revenue and ROI heatmaps, Smart CAPEX allows you to directly link these metrics to business outcomes, thereby helping optimize your network investments over time and giving you the best chance of winning and holding on to customers.
This methodology also enables you to more easily exploit new enterprise revenue streams through better prediction analysis.
But how does it all work?
A glimpse under the hood, and further benefits this technology brings
Like NLA as a whole, smart CAPEX relies on de-siloing data – from subscribers, population density maps, network data-sources and the like – and linking it together meaningfully to help you make better and faster predictions for pro-active decisions.
Driven by advanced AI and ML, it allows operators to plan using predictions on quality and business outcomes that are tested in the real world, with ML enabling algorithms to deliver better returns over time within each operator’s unique local and regional market context. This means you can make proactive investment decisions as your network evolves.
In essence, the system becomes a classic feedback loop. This can work in a number of ways but, to give an example, a CSP we’ve worked with in the US is building a 5G greenfield network in Las Vegas. During the planning phase, their work has been informed by data-based predictions. At the deployment phase, they’ll go on to test how accurately their predictions panned out in the real world. They can then make adjustments and fine-tune their prediction algorithms accordingly, ready for further work down the line – this is the feedback loop.
‘What if’ scenarios, sophisticated simulations, better time-to-market analysis, revenue heatmaps: as part of an NLA framework, these are the smart CAPEX systems that allow you to assess the different levels of network investment needed over time, whether for building new radio infrastructure or maintaining your existing 4G infrastructure. Important during these early days of 5G, this insight means you can then apportion costs with more precision, based on your network’s needs and likely revenue.
All of this equates to a high-precision approach to investment, which is so important with 5G, with features like small cell densification, expansion and configuration, and the adoption of higher spectrum frequency bands and mmWave on the horizon.
The true value of smart CAPEX
Your objectives will vary depending on the stage in the lifecycle that your network is in, whether you’re looking at financial planning and RoI predictions, through to dynamic modelling of revenue and investment potential. But whatever those objectives are, smart CAPEX allocation in an NLA environment provides the enablers for financial success, whether it’s maximizing RoI, reducing churn or improving your network evolution strategy to capture maximum revenues.
Later blog posts in this series will expand on the other themes in the introductory article to provide more context. Look out for Part 3, which we’ll publish in the next couple of weeks.
In the meantime, take a look at our report, Network Lifecycle Automation: the telecoms industry view, along with our new brochure, Network Lifecycle Automation: Accelerate monetization while reducing costs, without compromising on quality.