Communications service providers (CSPs) have worked doggedly over the last two decades to prepare for a new, data-led mode of operation. The plumbing is now done but extracting actionable insights from data lakes and then trusting that for fundamental operational processes is the next giant step operators need to take, writes George Malim.
CSPs have seen how other organisations utilise data to improve efficiency within their businesses. Financial institutions have, for example, been using early forms of data analytics to enable credit scoring since the 1950s. CSPs themselves have relied on data to assess credit risk for device subsidy, propensity to churn and for segmentation.
These are nice-to-haves and hit CSPs’ priorities which are, in common with all businesses, to make money, save money and achieve compliance. This trinity of aims enables investment to be approved and long-term strategy to be set.
However, CSPs are in a complex industry that is transforming digitally on multiple levels. The market is saturated and that means the industry’s old approach of simply entering a new market to gain subscribers no longer enables them to access new revenues. Instead, to win a customer they have to demonstrate they offer a better experience than a rival. Price can’t be a differentiator because CSPs have raced to the bottom and communications services have commoditised.
It’s not all doom and gloom, though. CSPs have fabulous network infrastructure and with the arrival of 5G and ultra-fast fibre many exciting new services and experiences are coming to market. CSPs are keen to harness these to move beyond simply providing the basic connection and see opportunities in mobile edge computing, IoT, high-speed gaming and a raft of other applications and services that require specific network and adjacent capabilities.
The common thread with these is far greater segmentation of the customer base is required to target users accurately and effectively. The market of one might seem a distant prospect and it will require enormous automation to enable across a subscriber base of millions. However, if operators can analyse their data and make timely, relevant and attractive offerings automatically and then also dimension the network, prioritise high value traffic and charge for it in automated ways, they will both move up the value chain to higher average revenue per user and achieve operational efficiency. The cherry on top is that a data-led mode of operation can also ensure compliance with telecoms regulation and data regulations.
The carrier barriers
Although CSPs are highly technical organisations with strong familiarity with data management, they have so far not reached the scenario outlined above. There are many obstacles in their paths, ranging from legacy to culture to competence.
CSPs’ legacy IT often extends back to the mainframe era when it was deployed to manage a basic communications service. These environments don’t have the flexibility to react rapidly to changing market conditions nor the ability to integrate easily with newer systems. Fortunately, many of these systems have been retired and the upgrade path is set to move to more flexible and cloud-based IT. Compounding this challenge is that most CSPs’ IT is composed of a web of systems that don’t necessarily interconnect. This is because of the industry’s heritage of mergers and acquisitions that has seen CSPs bought and sold and become part of large carrier groups. Although outwardly integrations appear complete, this is often not the case and it’s common to see mobile operations running on separate systems to cable or fibre offerings.
To provide the holistic, responsive experience that customer want, CSPs have to overcome these traditional demarcations and consolidate data from across all these systems.
Another significant challenge is the culture of the CSP. The traditional view is that network engineering is the heart of the business. The network is the prized asset and telco-grade engineering is the key attribute of the CSP. This means that CSP employees are composed of skilled workers that operate to high-level network engineering practices. They see the network via engineering metrics such as availability and throughput rather than consumer metrics such as video buffering or jitter during gaming.
Suggesting to a network engineer that a software-controlled network will reconfigure itself to support content uploads at a sports event or the peak loads of ultra-low latency gaming is almost an affront to their engineering heritage. The industry is moving in this direction but there is a culture shock that CSP organisations will have to overcome.
Finally, there is also a skills gap. Network engineers, marketers and customer service representatives are experts for a previous connectivity generation. The new environment is software-controlled and fueled by data insights in order to feed network automation. In mobile, network slices will be spun up and down to support specific apps, consumption and services and this will happen so frequently and at such enormous volume that automation is the only option.
CSPs are starting to embrace this. They’ve invested in the data infrastructure and analytics so the next step is to move to data-led operations. At the moment there’s a disconnect. The data is gathered, stored and analysed but the insight created is then acted upon with human involvement. This slows the process and increases cost.
Ultimately, it’s a question of trust and CSPs aren’t quite there yet. They simply aren’t ready to let their data feed automated systems that control their networks, serve their customers and feed their business support systems. Surely these businesscritical activities are too sensitive to entrust to a machine?
Therefore there is a transitional stage for CSPs to go through in which they familiarise themselves with the power of their data and understand more about what they can do with it. At the same time they’ll learn more about new techniques such as networking virtualisation and network slicing and know much more about what they can do with their software-controlled network.
This period, which has already started at pioneering CSPs, will see them learn the most important thing of all: to trust their data. Once trust in data is established, it will be used to drive CSPs’ decision making and from there it’s only a further, yet still significant, step on to data-enabled automated operations.