I really like this time of year. It is not just that the year still feels shiny and new, but also that it is the time of year when we start preparing for Mobile World Congress. The congress will discuss the biggest issues of the day, including artificial intelligence (AI) and the Internet of Things (IoT), and not just in theoretical terms. We are all starting to grapple with the big issues around the ethics of AI use, and it is good to see a whole session on that. There is also plenty on transition to 5G, an issue high on the agenda for telcos just now.

Getting the fundamentals right

New toys are always fun and exciting, and they don’t come much newer than IoT or AI. But just as parents have to make sure that they have plenty of batteries available on Christmas Day so their children can play with the new toys, so telcos need to look to the fundamentals to ensure that they will be able to generate full value from AI and the IoT. There is much change coming, and preparation is vital.

There are some very real challenges for telcos, particularly in capacity planning and management. If customer satisfaction is crucial — and all the evidence suggests this is true — then getting the infrastructure right to ensure that the network has enough capacity is essential. So too is having a secure network architecture that will support the digital IoT revolution, and also prevent fraud effectively.

So while I am excited to be preparing for Mobile World Congress, I am also glad to see that attendees at the Capacity Middle East conference later in March will be talking about the basics as well as the baubles.

Analytics for capacity optimisation

Perhaps one of the most fundamental issue is capacity planning. As an area, this has some very specific problems, but can gain a lot from effective use of analytics. For example, it tends to suffer from both under- and over-investment. One European telco found that 66% of its network upgrades were planned for nodes with a low saturation risk, and low value: in other words, that neither needed, nor returned the investment on, improved capacity. On the other hand, no upgrades were planned for 80% of nodes with a high saturation risk. Around 41% of the capital expenditure planned needed to be reallocated to manage the under-investment issue, but 25% could be deferred to limit over-investment.

By looking at upgrade costs against congestion at the cellular level, network analytics can be used to identify which cell upgrades generate the best return on investment. This type of granular forecasting not only improves the return, but also helps to keep the upgrade programme within budget by showing which areas do not need investment. Here’s a quick example: By applying network analytics (which consists of a combination of advanced forecasting and data mining techniques), a European telco operator managed to reduce their capital expenditure in different areas of network infrastructure by 15% to 25% while catering for the same service quality level demanded by their customers.

Traffic forecasting also helps to show where and when congestion is likely. Telcos can then take effective action to improve network quality predictions by more effective reconfiguration of settings and improvements in average throughput, increasing overall capacity. Network capacity planning, deployment and management, however, is only one area where telcos can make good use of analytics to improve performance.

Analytics in other areas

Analytics can also provide useful support for network service assurance, and particularly to improve customer experience. In-depth understanding of the influence of the configuration of network components can improve throughput and therefore customer satisfaction. Network failure prediction can also benefit from an analytical approach, allowing and supporting predictive maintenance of network components and better field operations. Finally, analytics can also be used directly to help telcos understand how network events can influence customer experience. This, in turn, can improve dialogue and conversations with customers.

These uses of analytics are not on big, shiny new issues. They are on fundamentals: capacity, network management, maintenance and so on. But which issues matter most to customers, artificial intelligence and the IoT, or a network that has enough capacity to make it usable? Getting the fundamentals right, and making sure that the company can provide a basic, but high quality, service is probably the most important determinant of customer satisfaction. The rest is a bonus. Telcos forget this at their peril.