There are clearly a lot of things going on in the telecom space. For well over a decade, operators have been experiencing erosion in revenue per bit. Some, particularly in areas like the EU where competition is fierce, have sought subsidies from big-tech firms to compensate for the traffic they add, traffic users themselves are unwilling to pay for. Neutrality rules change in the political winds. Businesses, under pressure to do more with less, are looking for ways to replace the costly business-specific services like Ethernet access and MPLS VPNs. Cloud providers seem to be targeting at least some aspects of business services. What’s a telco these days supposed to do? Let’s look first at these disruptions, then at the sum of options to address them.
Revenue per bit is not a particularly good way to measure telco profit potential, or at least not a comprehensive way. Over time, because of technology improvements, it’s common to see the marginal cost of bandwidth decline somewhat. The big problem is a lot more complicated. Over time, the broadband total addressable market (TAM) tends to plateau simply because the number of unconnected users declines in both consumer and business sectors. At the same time, the utility of broadband and the number of things users want to do with it stabilizes, which means that average revenue per user has little upward drive. Competition, then, actually lowers ARPU. However, whatever the source of the pressure, there is no question that telcos are under pressure to sustain and even raise profits, and natural forces are pushing in the other direction.
Neutrality poses a problem for two reasons. At the high level, the fact that the Internet model has always been one of bill-and-keep, and because rules often prohibit priority handling for a fee, it’s difficult for access providers serving the consumer market to share in the benefits created by large traffic generators like streaming video. While this rule isn’t always in play, given that telecom regulation is essentially political, the fact that it might be imposed again when there’s a change in government means operators can’t rely on a policy over time, which means they aren’t likely to deploy a premium service model.
The business services side is also critical in the sense of TAM/ARPU growth. Obviously, the number of business sites grows primarily as the retail presence for walk-in grows, and if anything online services in many forms has been pushing more and more in the other direction. The number of satellite sites of multi-site businesses has been shrinking gradually for about six years, and it’s this kind of site that generates new adds to an MPLS VPN. In addition, enterprises increasingly rely on the Internet and cloud to reach prospects, customers, and partners, simply because no other economical option exists. Given that, there’s a steady pressure to reach small office sites the same way, and since “business” forms of what’s essentially consumer broadband Internet access are steadily improving in QoS and availability (Comcast, for example will offer SLAs on theirs), the reasons to pay for MPLS/Ethernet access to sites are weakening.
All these changes are creating competitive options for the “interior” part of network services, and leaving traditional operators little more than the access piece. This happens to be a third of the total capex and almost 60% of opex, and thus the least profitable part of a network. It also includes mobile infrastructure, which has a history (with 5G) of driving a significant wave of investment that generates minimal return.
So, the net of all of this is that telecom as a business is being marginalized, and nothing is going to save it short of new revenues or new subsidies. The latter is obviously something that would likely be considered as a last resort, and that is surely a political issue. Thus, we have to look at the former.
Operators have tried to enter OTT service competition, and have generally not been successful for many reasons not worth going through again. Suffice it to say that the humor definition of insanity is to continue to do the same thing while expecting different outcomes. The only realistic form of new revenue would therefore have to come from either an extension of a current operator access service paradigm, or competing in a completely new area. As it happens, the two are likely related.
You can’t invent consumers or businesses; when the TAM has plateaued, you’re done in those spaces. However, you can serve devices independent of both, meaning some form of IoT. This truth was accepted by telcos, but with was (let’s face it) characteristic myopia, they saw it as meaning that you sold cellular plans to IoT devices. Obviously that’s not practical in most cases. Further, operators saw this as a simple connection mission, without considering the fact that devices are not inherently sociable and thus don’t talk spontaneously among themselves (please don’t tell me that AI will create just that sort of device!). So, if you want to sell access to gadgets, you need to have an application framework that creates both your opportunity and a viable value proposition for whoever/whatever is going to pay on their behalf.
There are probably over a billion IoT devices out there, the percentage of which are served by a connection some operator is getting paid for falls below statistical significance. Consumer IoT is almost completely associated with various elements of the smart home, including security, lighting, and heating/cooling, and WiFi and the Internet make any connections not made by wires or a special control protocol like Z-wave or Zigbee. Business IoT devices are largely installed in fixed facilities like plants, warehouses, etc. In some verticals, like healthcare and utilities, there may be an opportunity for mobile services, but enterprises say that where WiFi would not serve, they prefer private cellular services to a public operator because the devices are operating in a small geography.
So what happens, what could be a positive? The most obvious thing would be some application that relied on a lot of mobile IoT elements, a lot of new elements placed where wire connections could be expensive, or both. The problem, as it often is with the stuff that’s supposedly obvious, is that even identifying such an application is a challenge, and showing how it could make a business case for the buyer/user and all the stakeholders is even more so.
There is genuine momentum in industrial IoT applications for 5G/6G adoption, but far less so for public IoT services. Part of that is due to the fact that the vendor stakeholders in the applications are reluctant to make a push in the absence of any tangible service offerings with specific prices and features, and telcos have been slow to take the initiative in providing tangible/specific stuff absent specific use cases. Classic chicken-and-egg.
The cloud providers may end up being the players who break the deadlock here. They’re a middle player in the story; something that applications run on and networks connect to. While that may help get things moving, the driving player in any new initiative seeks to maximize their own value, and to make the initiative appealing to buyers, minimize the take of any other player. The network operators need to think about that point going forward.