In his latest book The End of Telecoms History, William Webb presents an interesting perspective on conventional thinking about future developments in telecoms. Webb argues that those who are well connected, with good home broadband and good mobile coverage – even if it is only 4G – have all the connectivity that they need. Chapter 3 charts the mobile network traffic volume growth curve at global and regional aggregate levels and concludes that data volume growth will cease in 2027. Webb presents a plausible analysis on why mobile data speeds beyond 10 Mbits/s are not needed and the only remaining problem is ubiquity.

The book addresses both fixed and mobile, but I focus my comments on mobile connectivity. While I agree with some of the points made, such as misplaced focus on headline speeds, lack of ubiquity, and the poor financial health of the mobile telecoms industry, I disagree with William’s conclusion that user requirements are nearly met and that we have all we need.

Looking at the growth of mobile data volume at a regional or global level makes little sense when data volumes in countries with similar per capita GDP differ by as much as a factor of 10. For example, Finland reports that between January to June 2024, monthly mobile data consumption per capita was 74 Gbytes, with roughly half of that from smartphones and half from data-only devices1. In Germany, the average monthly data usage per mobile customer (rather than per capita) amounted to 7.4 Gbytes and this may have risen to around 8 Gbytes in 1H 2024. In 2024, mobile data consumption in Finland was around 10 times higher compared to Germany. Using s-curves to predict future growth is useful and I myself make use of this methodology. However, it does not make sense to take an average of very different datapoints to generate a growth curve and predict a trend. Given that Finland and Germany are economically similar, what might cause this difference?

In Finland mobile operators have implemented 5G-SA and sell a user experience in terms of speed (Mbits/s) as opposed to data volume (Gbytes). As of October 2024, Elisa Finland offered a speed of 300 Mbits/s with unlimited data volume for €34.99. By contrast, in Germany, Telekom’s offer for 20 Gbytes is priced at €39.95 per month. For unlimited data usage Telekom charges €84.95, which is 2.5 times more costly than Elisa’s unlimited offer. It is unreasonable to assume that there is no price elasticity of demand. Surely, if prices in Germany were like those in Finland, monthly mobile data usage per customer would be much higher.

Webb predicts mobile data usage to plateau at around 15-20 Gbytes/user/month2 by around 20273. This ‘prediction’ does not make sense when in many markets mobile data usage already exceeds 20 Gbytes/user/month with significant growth rates. In Finland mobile data usage in 1H 2024 amounted to 74 Gbytes per capita of which 33 Gbytes from smartphones with a 12% growth from 1H 2023. Mobile networks in the USA, Austria, Latvia, Iceland, Sweden, Lithuania, Ireland, India, Malaysia among others had mobile data usage in excess of 20 Gbytes/user/month. For example in India (population 1.45 bn, the world’s most populous country), with a mobile data usage of 24.1 Gbytes/user/month in 20234, fixed broadband availability is scarce and as was the case with voice, mobile networks will be the primary carrier of data traffic. Unless we expect India to remain a low connectivity country we could expect mobile data usage to grow to well above 100 Gbytes/customer/month.

Webb’s assertion that we have all we need is another example of being misled by averages. Analysing monthly data usage in a mobile network by 10 percentiles, the top 10% of users typically account for 35-45% of data usage with consumption of five times the average or more, often in excess of 100 Gbytes per month. This shows that it is possible, even with today’s applications, that at least some smartphone users generate 100 Gbytes of data or more. This is a far cry from the “plateau at around 15-20 Gbytes/user/month” predicted by Webb. The most relevant variable in explaining these differences in monthly mobile data consumption is the age of the user. For example, a study by the Finnish mobile operator DNA found that on average, adult under the age of 30 use 74 Gbytes per month on their smartphones (excluding FWA) compared to 44 Gbytes in the 30-39 age group, 38 Gbytes in the 40-49 age group and less in each older 10 year cohort5. If one uses S-curves to forecast growth, then at the very least the upper asymptote should slope upwards to reflect the age shift over time.

However, the focus on data volume is misplaced – 5G is about delivering a user experience. I agree with Webb that as yet 5G has not delivered, but this is primarily because most mobile operators have not delivered the requirements for 5G (IMT-2020) as set out by the ITU6, namely to provide a 100 Mbit/s user experienced data rate in the downlink and 50 Mbit/s in the uplink, anytime, anywhere and while ‘on the move’. History teaches us that data usage is to a large extent supply driven. I fully agree with Webb that there is a lack of ubiquity in the delivering the user experience in terms of speed.

The high data usage in Finland is not only linked to price but also to the quality of the network: 5G actually works well. Contrast that with the experience of mobile users in London. Upon exiting an underground station, the wait for Google maps to load can be 30 seconds or longer if the satellite map type is selected. Attempting to load a discount voucher to the supermarket loyalty app at the checkout is extremely slow and often tends not to work at all due to poor mobile connectivity, and discussing an intended purchase on a WhatsApp video call rarely works. Approaching a bus stop in Berlin, you may find it very slow to activate a ticket on the app because people waiting at the bus stop are busy watching things on their smartphones causing area traffic demand (Mbits/s/m2) to exceed area traffic capacity. Commuters on the London underground have no coverage in tunnels and slow connection speeds in stations. Yes, they can download a newspaper before their morning commute, but they cannot watch the morning TV news during their 20-minute commute. These simple everyday examples illustrate that Webb’s assertion that “we have all we need” is divorced from the real experience of mobile users in most European cities.

There is nothing esoteric here – these are mundane applications which do work well. If there was ubiquitous reliable ‘speed coverage’, people could make much more use of their smartphones. Webb rightly bemoans the lack of mobile indoor coverage and advocates Wi-Fi as the solution, which would involve persuading all building owners to support a system where all mobile users from any network could seamlessly connect to Wi-Fi. Whether this will ever come about, how it would work for devices other than smartphones, and how it could deliver quality of service for applications that rely on QoS – all of this is doubtful. However, in advocating for a Wi-Fi solution, Webb in effect concedes that current 5G usage is constrained by availability, i.e. demand is constrained by supply, and that if the constraint is removed usage would increase.

The book asserts …that there is no need for densification. This ignores the fact that small cell densification is mostly driven by the need to improve the user experience in places where macro sites do not provide sufficient speed coverage. Enlightened municipalities have recognised the benefit of excellent speed coverage and offer municipal infrastructure such as lampposts for small cell installation for a nominal site rental. Indoor deployments are part of the densification story, with neutral host installations gathering momentum. These indoor small cell installations such as Ericsson dot are much cheaper than legacy DAS, and the total cost of ownership may be less than 5% compared to a rooftop macro site.

5G-SA is of course different from Wi-Fi. Webb focuses on data speeds, however 5G is not simply a new radio interface but a new core. With the generational development from 4G to 5G came new features such as network slicing, and it is linked to standardising new frequency bands for mobile, widening the channel bandwidth from 20 MHz to 100 MHz, and the deployment of high orders of MIMO. New spectrum and MIMO are key in reducing mobile operator’s cost per bit and have allowed operators to offer vastly better value for money to mobile users, with a decline in the effective retail cost per Gbyte in the order of 80% over 10 years. It is hard to think of any other industry which has delivered such an improvement in value for money.

A key aspect of the ITU vision for 5G is that is addresses business and institutional uses as well as consumer use. New features, such as network slicing, enable new use cases. For example, the Kaohsiung City Police Department in Taiwan operates a ‘smart patrol car’ using a network slice on a 5G-SA network to support licence plate recognition. With this application, patrol cars equipped with high-resolution cameras use an AI image analysis solution to identify vehicles that have been reported stolen.

Regarding new features, notably network slicing, ubiquity, including indoor coverage, is a requirement for many use cases. This ranges from asset tracking (which does not require a lot of bandwidth) to body cameras worn by first responders and other field operatives. Live streaming by body-worn cameras is not some futuristic application, it is already being implemented. Live streaming body cameras can give a rapid indication to control rooms about the risk level of a situation. The live streams will be analysed by AI triggering alerts for control room intervention when needed. While currently only a small percentage of workers are equipped with this technology, this is only one application that would rely on the ubiquitous mobile speed coverage, outdoors and indoors.

The book states that data rates beyond 10 Mbits/s of mobile … will not make any meaningful difference to end users. Looking at the video setting on my slightly dated iPhone 12 Pro, I can see that my setting of 4K with 60 frames per second would use 440 Mbytes per minute. That is a data rate of 20 to 80 Mbits/s depending on compression. While this is only one data point it shows that even today there are applications in our hands which demand a data rate of roughly 2 to 8 times more than 10 Mbits/s stated in the book. The human ability to perceive quality watching a video on a small screen is one criterion. However, users may want to zoom in on a football tackle without pixelation and look at it in slow motion, both of which require higher resolution or higher frame rate and hence a higher bandwidth. Security applications can be far more demanding. AI analysing a security camera video stream may identify a suspect person. However, the alert to the control room is only triggered when AI zooms in on the person’s hand to see whether the individual holds a folding umbrella or a gun. For streaming at higher resolutions, higher bandwidths are required.

There are an estimated nine million CCTV cameras in the UK. Cameras could be hooked up by wire or Wi-Fi, but it is much more flexible and convenient to use 5G connectivity. Convenience is of course a major driver in mobile and Wi-Fi connectivity. The largest consumer of data in a home is the connected TV. The TV does not move around but it is connected using Wi-Fi because it is more convenient and cheaper than running cables. Convenience is also why many 5G mobile users – provided that there is good speed coverage – no longer bother with Wi-Fi even at locations where is it free. However, often simple tasks such as using a store card voucher in a retailer app frequently cannot be accomplished without logging onto the retailer’s Wi-Fi network, which is cumbersome and inconvenient for users who have become accustomed to not having to think about connectivity. Why would a consumer log onto multiple Wi-Fi networks while shopping on a high street instead of benefiting from seamless 5G coverage indoors and outdoors?

This brings us to the lack of ubiquity or what I prefer to call lack of ‘speed coverage’. This where I am in full agreement with the author. European policymakers focus on arbitrary headline speed goals and delude themselves that 5G has 81% population coverage in Europe, when where most people spend most of the time in cities, i.e. indoors, there is little or no 5G connectivity. In London there is even a lack of 5G coverage outdoors in central areas such as Holland Park. The misguided focus on rural coverage targets which, for example requires German mobile operators to deploy 3500 MHz sites in the middle of nowhere, is a waste of money. A Euro invested in indoor coverage in cities would benefit more people more of the time and would generate more socio-economic benefit than incremental coverage to ever more remote areas. Regulatory intervention such as the equivalent of wayleaves for indoor mobile sites may help and will foster awareness among consumers and building owners.

If applications such as streaming, video telephony, or simply looking up a map with satellite image detail do not work properly, then this amounts to having no coverage at all. In other words, today’s 5G networks in cities, where most people live and work, do not even come close to ITU-R’s requirements for 5G (IMT-2020). It is highly likely that demand – i.e. what people and businesses want to do – is constrained by the lack of area network capacity (speed coverage). If people could actually use their smartphones reliably anywhere, anytime, usage is likely to increase substantially.

For businesses, municipalities, and public services the lack of ubiquitous speed coverage is worse. How can a business or municipality build a process or service which relies on 5G connectivity when in urban areas there is only patchy 5G coverage. That is not what 5G is meant to be.

In this context, I find great merit in Webb’s suggestion, in chapter 6 of his book, to enable consumers and regulators to make informed choices based on quality. Today 4G and 5G non-SA mobile networks deliver an undifferentiated best effort service where price is driving consumer choice and effectively creating a race to the bottom. I for one would happily pay double to benefit from a true 5G service.

A further issue is that consumers find that when streaming video over 5G, the battery drains rather quickly and hence many people ration use. There are two aspects to this, network design and availability of charging points.

With regard to network design, when I compare the battery life in London vs. Toronto, London falls short. In London (using a UK mobile operator) I barely get through the day whereas in Toronto on the Rogers network my battery is still not empty by the end of day. Poor network quality leads to shorter battery life and consequently many people ration their use of 5G.

Looking at the proliferation of smartphone charging points, new buses in Paris, Berlin, and London now feature USB charging points at seats. With consistent 5G speed coverage, a commuter could watch TV news without fear of running out of battery.

Webb rightly points to the poor financial health of the telecoms industry. If governments stopped extracting billions in spectrum licence fees, perhaps more cash would flow into network investment, notably indoor and outdoor speed coverage, and we would see a virtuous circle of increasing area traffic demand (Mbit/s/m2) and increasing area network capacity. Above, we contrast the excellent mobile services on offer in Finland, where the annualised cost of spectrum accounts for less than 1% of service revenue, with Germany where it is around 10%. Low spectrum licence fees are a significant factor in Finland’s 5G success story.

Lastly, RAN equipment has a useful life of around eight years before it needs replacing. Of course, operators would install the kit that delivers the best performance at the lowest price (lowest cost per bit/s) with the lowest relative energy costs. There is therefore a natural continuous upgrade process through which innovation and lower costs are delivered to end-users. Like Webb, I am sceptical with regard to 6G, but looking at ITU requirements for 6G, there are useful features such as better positional accuracy of sub-1 metre outdoors. Such accuracy in combination with smart city developments can be very useful. Today, when parking on a street in London, people need to use an app to enter the location, estimate the time they will need, and then pay. Would it not be much more convenient if the smart city simply recognises where people park and for how long and then charges the appropriate fees? Of course, this only a trivial low-bandwidth example, but there may be many more useful improvements to our daily lives. Hence, I do not think we have reached the end of telecoms history, but rather that we can look forward to incremental improvements which may be quite useful.

1 TRAFICOM, 1 Oct 2024
2 The End of Telecoms History, W. Webb, p.37
3 The End of Telecoms History, W. Webb, p.100
4 India Mobile Broadband Index 2024, Nokia
5 DNA press release, 12 December 2023, figures reported as daily average converted to monthly averages (30 days/month)

About Stefan Zehle, CEO, Coleago Consulting Ltd

30 years’ experience in telecoms, Director of mobile operating company, specialises in strategy & business planning and spectrum issues. He gained his experience working in 65 countries on all continents. As the 2nd person on the ground and Director of Marketing, Strategy & Regulatory he played a pivotal role in launching Nedjma, the 3rd Algerian GSM operator. He launched new propositions transforming the market and introduced best practice interconnect agreements. Stefan worked on over 90 spectrum related projects. Speaker at the Mobile World Congress in Barcelona and 3 to 4 telecoms conferences per year. Recent publications include the “The need for sub-1 GHz spectrum to deliver the vision of 5G”, report for the GSMA (2022) and “Demand for IMT spectrum in the 2025-30 timeframe”, model and report for the GSMA for WRC-23 (2021). Co-author of the Economist Guide to Business Planning. MBA with distinction from University of Westminster, London.

About Coleago Consulting Ltd

Founded in 2001, Coleago is a telecommunications consulting and training firm. We offer an experience-based consulting approach, with project teams entirely made up of partner-level consultants, each with a minimum of 20 years’ experience in the telecoms sector. Our advice is therefore based on practical experience and proven processes and methodologies developed over many years. Since 2001 we have carried out over 140 spectrum projects in developed and emerging markets on all continents. Coleago consultants regularly speak at international telecoms conferences, on topics such as spectrum, infrastructure sharing, digital content in Europe, Asia, the Americas, and Africa.

Get in toucharrow_forward