IoT, 5G, edge computing and AI showed they were maturing as technologies at MWC 2018, but it's harder to see yet how the first three make much money for telcos. AI has a different problem: the field is developing so fast that best practice is changing all the time. In this report we outline what we found, and a pragmatic approach for telcos to successfully harness these technologies and create value beyond connectivity.
What we found at MWC 2018
STL Partners were delighted to again be official partners of the GSMA for the Mobile World Congress (MWC 2018). Our team decamped to Barcelona for the week, and the following is an extract of our analysis of what we saw, heard and thought – and what telcos should do.
No big surprises, but plenty of nuance
As we expected (see MWC2018: 5 things to watch out for), IoT, 5G, NFV/SDN, edge computing and AI were the main topics of discussion, alongside telcos’ continuing struggles with the challenges of managing digital transformation and creating new revenues.
With consumers, and especially in more mature markets, telcos are fighting to retain and regain relevance with their customers, and struggling to justify why they should pay any more for data, lower latency or the insights IoT data can bring.
Hence the telco business cases for IoT, 5G, and edge computing are all eyeing the opportunity to digitalise and connect enterprise applications, sometimes hyped as ‘The Fourth Industrial Revolution’. But the enterprise market is notoriously difficult to serve due to its complex demands, so success requires a good understanding of industry-specific issues, and the business cases are hard to make.
So IoT, 5G, and edge computing are (somewhat inter-related) technologies that will initially deliver the same business benefits, i.e. helping enterprises adopt more efficient, agile and data-centric processes. AI is rather different and will have a more widespread impact in telecoms as we outline below. But each faces its own challenges.
5G: patchy to start
5G should deliver more than enterprise applications in the fullness of time. But most telcos are reluctant to step straight back into a major network capital investment cycle, and still have money to spend to meet the rising demand for 4G.
So, we think the market development of 5G will look different to the other ‘G’s, and the initial uptake of 5G will be patchy:
- The US market is looking set to talk itself into a 5G war, with each operator having its own motivation to be seen as the leader in 5G. AT&T because it wants to do everything, Verizon because it has an instinct to retain its premium network tag, T-Mobile because it wants to punch all the other operators, and Sprint because it needs to find a way to differentiate. The US market is also big enough to tempt handset vendors into production – or at least into experimentation in 5G. From a PR perspective at least, the US operators’ fingers are twitching on their triggers, but it’s likely that the CFOs and shareholders will need a little more persuasion that it will work economically. This means there will probably be an initial period of ‘phoney war’ as they work out how to play it in various tests and trials, while making increasingly aggressive claims to win the opening propaganda contest.
- South Korea and Japan also seem determined to head off on the 5G path, and are probably sufficiently advanced markets that are suited to taking the technology upgrade early.
- Elsewhere though, the enthusiasm for 5G is a little more muted, but the enterprise applications could well make sense for reasons we discuss later in this brief report. Having said that, much of what we saw at MWC 2018 would best be described as “technology seeking a business case” – some interesting technical developments, but little coherent economic rationale that even vaguely approaches a credible business case. “If we build it they will come” may turn out to be the most pragmatic argument, so it is again initially likely to be something of a “test and pivot” approach, as operators dip a toe in the water to see what they can make work, and where, before betting more widely.
- Wider new applications such as autonomous cars and AR/VR won’t move the needle until later in most markets. Autonomous cars because they won’t be able to rely on a network until it’s widespread and highly proven (see Autonomous cars: Where’s the money for telcos?), and AR/VR because it will take significant time to develop as a widespread and highly used technology (see AR/VR: Won’t move the 5G needle).
Many network vendors, notably Huawei, while hoping to bet big on 5G are sensibly taking an ‘incremental’ approach to 5G, and designing it as an add-on or upgrade to 4G or 4G+ solutions.
NB. We will soon be publishing a detailed report on our analysis of 5G’s progress.
Edge computing: the need to move beyond technology
MWC 2018 has shown that the industry is still mainly focussing on technology-focused PoCs in the area of edge computing. Key use cases that were demonstrated typically revolved around optimised video delivery, IoT edge gateways, and control of autonomous vehicles or drones.
However, it seems that the industry hasn’t got much closer to articulating viable business and monetisation models for these – technically impressive – use cases. Telcos still need to find out how to cope with three fundamental challenges and uncertainties which represent the key road blocks for success at the edge:
- Commercialisation: It is – and will be for the foreseeable future – unclear which edge use cases will deliver significant commercial value to telcos (and to their customers for that matter). In addition, telcos lack clarity on which business models need to be employed to monetise individual use cases. We have addressed this issue in Edge computing: Five viable telco business models.
- Operationalisation: Edge computing capabilities might be relevant for very different parts of the telco organisation – for both internal and external use cases, as well as wider efforts which are related to NFV and 5G. This calls for a certain degree of coordination within the organisation. Equally important is the need for an edge platform which ensures flexibility and speed in developing and onboarding applications.
- Ecosystem orchestration: Telcos need to work out what their role in the wider distributed (multi-)cloud ecosystem should be, as edge computing is not solely a telco concept (see e.g. AWS Greengrass). This boils down to the question of who telcos should partner with and who they should compete with in the edge computing and edge cloud space. In addition, telcos are starting to acknowledge that there needs to be significant technical and commercial interoperability between operators providing edge computing capabilities to third parties. Otherwise, telcos will stumble over the usual problem of market fragmentation, which would make it unattractive for application providers and developers to offer their services through the telco edge cloud.
STL Partners is currently undertaking primary research on this topic to identify the key short-term and long-term strategic principles for telcos to overcome the above barriers and ensure commercial success at the edge. The findings will be discussed in an upcoming report.
IoT: the struggle to add value
What we saw at MWC 2018 lined up with what we said in Monetising IoT: Four steps for success. “IoT is not a quick win for telcos. The value of IoT connectivity is only a small portion of the total estimated value of the IoT ecosystem, and therefore telcos seeking to grow greater value in this area are actively moving into other layers, such as platforms and vertical end solutions.”
Telcos therefore face a conundrum in IoT. At one extreme, they can focus on the connectivity, and can do reasonably well by providing SIMs plus functionality in reporting and management. Above and beyond that, which is where the bulk of the economic value lies, industries need sophisticated and evolving solutions that integrate connectivity, applications, machines and data.
Such industry solutions need to be tailored to niche demands and limitations, such as specific regulations or legacy infrastructure. As an example, hospitals want to digitalise their operations, but they can’t afford to replace expensive medical equipment like 15-year-old MRI machines, which might not be compatible with the latest technologies and application programmes. On top of that, connecting sensitive medical data comes with sensitive data protection and security requirements. To make all this work well and securely is no small matter, and the province of some highly sophisticated specialist players, well beyond the appetite of most telcos. (NB. We discuss other possible healthcare approaches on page 8 of this report.)
To be successful at a wider level in Industrial IoT will take serious knowledge at both technical domain and sector level, and this is not easy to put together for even one industry, let alone several. It takes vision, commitment, time and investment.
NFV/SDN: grinding forward
NFV/SDN at MWC 2018 was again very much in line with our previous findings and expectations. Operators are making progress, but it is slow and piecemeal.
On the ‘progress’ side of the equation, two new open source initiatives to develop standards for RAN virtualisation were debuted at MWC: the Open RAN (ORAN) project (in which AT&T, China Mobile and Deutsche Telekom are pooling their efforts) and the Cisco-led Open vRAN initiative (vRAN and C-RAN will feature in our forward research). On the ‘piecemeal’ side of the equation, these developments only add to the sense of fragmentation in industry efforts to arrive at NFV standards and interoperability frameworks, which will be critical for realising the potential of NFV to support 5G use cases in areas such as IoT, edge computing and network slicing (as discussed above).
In our forthcoming reports on emerging NFV / SDN technology and use cases, we will build on the analysis in our recent report NFV/SDN deployment pathways: Three telco futures and attempt to show how strategic clarity about the type of telco they wish to be, and the social and economic functions they see their services as performing, is vital to inform operators’ engagement in NFV and SDN, and their selection of NFV / SDN use-cases and associated vertical markets (see discussion of expansion into new verticals below). The first of these reports will take in the MWC ‘hot topics’ of 5G, edge computing and network slicing, with subsequent analyses looking at SD-WAN / on-demand enterprise networking, and industrial and telco automation.
And of course, we are continuing to build our database of information on live, commercial deployments of NFV / SDN, the NFV Deployment Tracker, with further updates adding data on deployments in the Asia-Pacific region, the Middle East, Africa and Latin America, in addition to those recently completed for Europe and North America.
AI: Strategies taking form
AI is an exception to the other technologies because it has both
- Internal applications for telcos, helping to streamline repetitive and data-intensive tasks across their businesses,
- And the potential to complement external enterprise solutions.
While 5G, IoT and edge computing seem to be all about finding the use-cases, every sector and every part of telcos’ businesses, from network planning and operations, to back office functions, customer experience and product development, can be streamlined with more advanced data analytics and automation. So, it’s not surprising that all telcos are thinking about how to implement AI.
Tier 1 players such as AT&T, SK Telecom, Orange, Deutsche Telekom and Vodafone already have clear plans on what they are prioritising in the short-term, and what they want to do internally versus partner with vendors like IBM. Improving customer experience is by far the priority area right now, reflecting telcos’ desire to regain credibility with consumers, as well as the relative maturity and awareness of natural language processing within the wider field of AI/machine learning (ML).
Despite having clear ideas of what they want to do, most telcos are still in the early days of implementation. Any live telco chatbots still have limited capabilities and are only operational in one or two markets, so 2018 will be a year of scaling into new markets and channels.
The crucial initial promise of AI is to save costs across the business, and behind the highly hyped chatbot programmes there is evidence that telcos are taking steps towards using machine learning for predictive care (see our report AI in customer services: It’s not all about chatbots), back office resource planning, and network maintenance and operations.
The major question for telcos is how they should organise their AI initiatives and skills for the highest impact, and where and with whom they should partner. Right now, most telcos are taking a decentralised approach to deploying AI, so as not to stifle or hold back progress across various business units or opcos, with an aim to shift towards becoming more centralised.
Whether this is the right approach is still up for debate. We are addressing this question as part of an interview programme with telcos on their progress and strategies around AI and will publish our findings in a forthcoming report outlining the telco AI roadmap.
To read on about what telcos should do, please login and download the report, or contact us to subscribe.
- What we found at MWC 2018
- No big surprises, but plenty of nuance
- 5G: patchy to start
- Edge computing: the need to move beyond technology
- IoT: the struggle to add value
- NFV/SDN: grinding forward
- AI: strategies taking form
- So, what should telcos do?
- Next steps
- Figure 1: TELUS Health Exchange