Telecoms data analytics – Where’s the real value?

Executive Briefing Service

Purchase report

This report is available to purchase.

Buy Now

Login to access

Want to subscribe?

This article is part of: Executive Briefing Service

To find out more about how to join or access this report please contact us

Although nearly all operators aspire to deploy autonomous networks and personalised customer services, few have actually implemented advanced analytics at scale across their organisations. Almost universally, telcos are hampered by incomplete and siloed data sets and cultural resistance. What have the industry's leaders done to overcome these challenges?

Why telecoms data analtyics matters

Telecoms data analytics matter because telcos currently face big challenges. As connectivity services are increasingly commoditised, telcos are seeing a steady decline in core revenues. They are at risk of becoming seen as providers of basic utilities, rather than offering innovative services to their customers.

Improved analytics fuels benefits in multiple layers. Initially in the (human) management of operational performance, and then increasingly through using analytics and managed AI to control automation. This can apply to existing and new services.

Future-proofing: why data telecoms analytics are essential to today’s business, 5G and beyond

There is a lot of noise from the telecoms industry around fifth generation (5G) mobile networks and how 5G may provide a renewed source of revenue growth. There is no doubt that 5G will unlock new vertical opportunities for telcos, however, if telcos do not invest in developing additional services, revenues will primarily still come through connectivity. While some may gain a first-mover advantage, over time, 5G will experience the same diminishing returns per user that we have seen with previous generations (see Figure 1). 5G, through connectivity alone, is therefore likely to make only a short-term impact on telco revenue streams.

Figure 1: The effect of increasing 4G subscriber penetration on ARPUs

Source: Data from company filings, analysis by STL Partners

STL Partners has been writing about the commoditisation problem for many years and has seen that operators increasingly accept it as inevitable. Most, in one form or another, are looking beyond connectivity to improve the bottom line. Telcos are adopting two main strategies:

  • build or acquire new revenue streams outside of connectivity
  • cut costs.

The first of these is increasingly popular. Telcos worldwide have accepted the idea that they must develop new capabilities outside of their core service area and find ways to make money from them. These capabilities, and how well they link back to existing connectivity offers, vary widely. For example:

  • Some, realising telcos’ technical expertise, are developing end to end solutions based on new technologies such as multi-access edge computing and 5G. Although technologies such as 5G may not bring sustained growth through connectivity alone, they do offer the potential for telcos to access new areas of the value chain and derive new growth opportunities.
  • Some are developing new services in specific verticals. For example, TELUS in Canada and Telstra in Australia are both building service platforms in the healthcare sector, primarily through acquisitions of health-tech companies.

Unfortunately, due to heavy capex constraints and debt regulation, many telcos face challenges in investing in innovative technologies and only some have shown real success in building new offerings outside of traditional telecoms. All telcos are, however, implementing the second strategy, focussing on cutting costs and driving efficiencies throughout their organisations. Although exploring new verticals and areas of opportunity outside of connectivity is a must to drive sustained growth, in order to defend their territory against the likes of Amazon (who operate on razor thin margins), it is essential that telcos look internally and cut costs across their businesses.

While we see many variants and combinations of these two core strategies across the industry, there is one key element that ties them together. Operators are increasingly taking the view that the key to success – both in building new revenue streams and keeping costs down – is finding ways to make better use of data.

Through their networks and customer interactions, telcos collect a broad array of data. This data comes from both internal (for example data on network performance) and external (for example customer location data and usage data) sources. Telcos can extract and leverage insights from this data more accurately and more quickly through advanced analytics, informing key business decisions, creating efficiencies for internal processes, and unlocking data-enabled new service areas including the facilitation and adoption of technologies like 5G.

Building an advanced analytics capability

High ambitions: data and the AI continuum

When we talk with operators globally about data analytics, a key point of discussion is artificial intelligence (AI). “AI technology” is often cited as a powerful way to cut costs, increase ARPUs, and reduce churn – across an operator’s business. Indeed, at STL Partners we have written extensively about how this could be achieved. However, much of the discussion around AI in the industry is just that – discussion. Many AI solutions are still in their nascent phases and there is a lot more talk than live implementations that deliver measurable business value.

We raise two points to help cut through this hype and understand the real-world value for operators, both in the long and short-term.

  1. All AI is equal, but some AI is more equal than others”. It may seem out of place to paraphrase George Orwell, but the truth is that operators and vendors alike market an increasingly broad set of solutions to customers and the analyst community under the blanket term “AI” (“all AI is equal”). This is often misleading, if not erroneous. “AI” can mean different things depending on who you speak to, ranging from computers following simple instructions or rules set by humans, to more complex fully autonomous computer systems that learn and improve with limited human interaction (“but some AI is more equal than others”). These examples differ strongly – but both fit within a generic definition of “artificial intelligence”.

Agnostic of what you include in your definition of AI, there are clearly tiers of AI solution which are based on the algorithm’s complexity, its ability to implement decisions independently (in terms of rights/permissions and integration with automated processes), and the level of human interaction or guidance necessary. At STL Partners, we have written previously about how we see advanced data analytics and AI as a continuum, with stepping stones on a journey towards the fully autonomous telco (Figure 2) The detailed explanation and formulation of this continuum is more thoroughly explained in a previous instalment of our AI research series.

  1. Most live and scaled deployments fall under our definition of rules based automation. Operators speaking to us about AI tend to want focus on innovative AI use cases that fall in the right-hand side of Figure 2. Examples include automated and self-improving chatbots that can solve any customer query and translate a complaint into a sale, or self-healing networks that fix themselves with no need for engineers to intervene. It’s true that these use cases will deliver high-value for telcos and help to answer the big questions set out above. However, should telcos be prioritising these if their data systems cannot yet tell them which customers are having a poor experience, or give them a full, real-time view of network performance?

Where are operators compared to their AI aspirations

Source: STL Partners

In terms of real progress, we have seen only a handful of leading Tier 1 operators deploying telecoms data analytics solutions that truly fit under the ML/AI banner within our framework. Most operators are still much earlier on in the journey towards automation. Even those pioneer operators have deployed only in specific geographical regions and in specific parts of their business. They face problems in deploying more complex solutions at scale and deriving measurable value.

At STL Partners, we believe that too much focus on a poorly defined end-goal risks stalling necessary work that must be done up-front. Operators should strive for and research innovative uses of data, but we believe the focus in the short-term, for Tier 1 and 2/3 telcos alike, should be on laying the necessary groundwork to ensure that data is accessible and clean, with a clear governance structure, as well as building the analytics capabilities necessary to make full use of it.

Laying the groundwork: stepping stones toward data analytics

There are three key components to building even the most basic data analytics capabilities:

  1. Clean, unified data
  2. The skills and tools to process and analyse it
  3. The ambition and drive to do so – data-centricity

This may seem straightforward but telcos globally (including even the most advanced operators) have faced challenges in meeting these requirements. For example, 77% of the operators we have spoken to stated that data collection and management was a key issue for them in implementing an analytics strategy. Furthermore, over a third of the operators we spoke with mentioned a lack of both internal and external skills with regards to advanced analytics (see Figure 3).

Figure 3: Top 4 issues faced by telcos looking to make use of data

Source: STL Partners research programme, October 2018

In order to overcome the issues listed in Figure 3, and to build future-proof telecoms data analytics capabilities, telcos must develop the three components mentioned above. Without doing this in the short-term, operators will lack the underlying platform from which to springboard into developing innovative solutions that leverage AI or ML.

Contents:

  • Executive Summary
  • Future-proofing: what to do?
  • Building an advanced telecoms data analytics capability
  • High ambitions: data and the AI continuum
  • Laying the groundwork: stepping stones toward data analytics
  • In practice: Assessing real analytics use cases
  • Improve business as usual
  • Monetise user data
  • Enable next-generation services
  • Conclusions
  • Key recommendations
  • Conclusion

Figures:

  • Figure 1: The effect of increasing 4G subscriber penetration on ARPUs
  • Figure 2: The journey to AI and telco automation
  • Figure 3: Top 4 issues faced by telcos looking to make use of data
  • Figure 4: Telefónica’s data management structure across multiple opcos
  • Figure 5: What is your biggest challenge in leveraging analytics?
  • Figure 6: The opportunity areas for telcos in advanced analytics
  • Figure 7: A comparison of Iliad against the leading Italian operators
  • Figure 8: A graphical representation of KPN’s Data Services Hub
  • Figure 9: Where operators are compared to their AI aspirations