From network planning to monetisation by cell: the critical role of Smart CAPEX

By Wei Shi and InfoVista

What would make a mobile operator jitter the most when it decides to roll out 5G services? In all likelihood it would be the cost. When answering a recent Telecoms.com industry survey on 5G, 60% of all the respondents cited gear and deployment cost as the biggest impediment to 5G rollout.

Something worth noting is that cost does not only occur at the time when the build-out contract is signed, and it does not only occur with the first deployment. Throughout the network lifecycle, be it a greenfield deployment or densification or expansion, different spending plays different roles. This means optimising spending should also cover the whole network lifecycle, from plan and design through deploy and test, operate and support all the way to report and monetise. And it is not one-off but happens in a continuous loop. It is also worth noting that optimising cost does not equate reducing cost. Instead, it means delivering the best return on every dollar spent.

New and Complex Networks Need Smart Tools

A main driver for the need to continuously optimise is the complexity of 5G networks and the nature of constant change of its services. To start with, 5G’s network architecture is complex, which includes conventional physical networks, virtualised network functions, and cloud infrastructure. Demands on 5G network capabilities are also complex. To meet the demands of advanced use cases 5G needs to deliver defined throughput rate, capacity, and latency. These come on top of coverage requirements regulators impose when 5G licences are awarded. The business cases 5G sets out to serve are also complex, many of which keep changing and many of which we do not know yet, especially those expected to come from the enterprise market.

Some operators and their partners choose to do case-by-case planning projects to prepare for network rollout and serve their business needs, using a fusion approach combining automatic and manual planning. Such an approach, by definition, is not replicable from one case to the next. The lack of scalability will become a serious handicap when the deployment or densification needs to happen at a large scale while under time pressure. Probably a bigger challenge will face the case-by-case approach when user requirements and service offerings change and become more diverse and complex, which is fully expected when operators start offering full-fledged 5G services.

Optimising cost throughout the network lifecycle in such demanding environment therefore needs robust, agile, and highly automated planning solutions. They should be able to flexibly integrate, reconcile, and process data from network and non-network sources, then produce advanced AI/ML-based predictions for technology and business outcomes, automated analytics of different technology options, and automated rollout scheduling.

Smart CAPEX is one such analytics solution, though it does more than just providing smart CAPEX projections. With network and non-network data as input, equipped with advanced AI and ML analytical capabilities, and with technology and business projections as output, Smart CAPEX can deliver precise investment recommendations for every stage of the network lifecycle. It is therefore serving the end-to-end service lifecycle process and playing a critical role in the overall Network Lifecycle Automation (NLA) process.

What Smart CAPEX Is About

It all starts with data input. In order to deliver precise predictions including those of network coverage and quality, subscriber’s QoE, potential churn, as well as revenue potential by location, the analytics solution takes in both network data (e.g., radio propagation parameters, base station form factors, RAN specifications) and non-network data (e.g., terrain, population density, customer behaviours), breaking the silos between them, health-checking and reconciling the data, and processing them with advanced AI and ML capabilities.

While incumbent operators have already troves of data to feed the analytics engine, greenfield operators may find it hard to run service journey simulations without existing customer data. This is where the analytics solution should play a more proactive consultative role. The AI algorithms are developed with the best understanding of the technology and market, so the operators can still build the sites in the best possible location with the best possible configuration to achieve optimal TCO and ROI targets.

The advantage of AI and ML powered automatic systems like Smart CAPEX is that the algorithms can be quickly tuned according to each operator’s unique local and regional market context when the accuracy of the predictions are tested in live deployments. Insights gained through the fine-tuning of algorithms can be used for next round expansion or densification, therefore developing a positive feedback loop.

With output like revenue projection to cell site level (or potential revenue opportunity, for that matter) and ROI heat maps, Smart CAPEX enables operators to accurately predict both technology and business KPI results, for example QoE, TCO, TTM, ROI. This in turn helps operators more accurately plan their spending on every stage of their network lifecycle, and accurately adjust their spending level to improve on KPI results when needed.

In more advanced implementation cases and to get the best out of the AI and ML capabilities of Smart CAPEX, fully digitised and extremely accurate data input like Digital Twins should be incorporated in the loop. In such cases operators can minimise engineers’ site visits, not only to optimise the use of their time, but also, in times of pandemic for example, to minimise their exposure to risks.

In a nutshell, Smart CAPEX, as its name suggests, supports operators to smartly plan spending on their networks, be they greenfield deployment or densification or upgrading, or launching and updating new services, by projecting results of both technology and operation KPI results ranging from whole network scale to specific local level. The accuracy of the projections keeps improving as a feedback loop is developed for the AI and ML engine to learn from its own records and the tuning of the algorithms, making Smart CAPEX a critical component of the whole NLA approach.

https://www.infovista.com/NLA

https://www.infovista.com/blog/maximizing-network-roi-smart-capex-network-lifecycle-automation