As I said last time, the juxtaposition of ‘5G’ and ‘Enterprise’ at Nokia’s recent Global Analyst Forum got lunchtime (and beer time) conversations to conclude that the 5G business case depended on the enterprise market – despite the American carriers’ race to launch consumer-market 5G as fast as they possibly can.
Not only had we been bombarded with enterprise use cases for the two days of the event, but many of the unique characteristics of 5G seem to only be justified by IIoT, public safety networks, industrial drones, untethered robots, etc. etc.
True, consumers will want the ‘higher/faster/stronger’ 5G bandwidth to, paraphrasing Bell Labs’ Marcus Weldon’s words, get their cat videos faster and be able to do their online shopping more effectively. Since the beginning of the broadband era, the industry has demonstrated time and again that there is never enough bandwidth. Consumers are never happy with their broadband experience and the entertainment and gaming industries quickly escalate their demands for faster processing, denser graphics and quicker response times.
But do consumers really need 1 millisecond latency from their smartphones when the human brain takes one or sometimes two orders of magnitude more time to respond to stimuli and make decisions?
No. One millisecond latency is needed for machines to respond to stimuli and make decisions, not human beings.
Do consumers need 10-year battery life from low-power devices?
No. While smartphone and tablet replacement cycles are lengthening, they are nowhere near 10 years (and no one in the standards bodies have promised us 10-year battery life for smartphones). We are likely to bore of our smartwatches even implanted health devices or want the latest and greatest physical tests, the better to satisfy our paranoia and desire to live forever. That is, we are likely to turn over our personal devices long before 10 years and so what is the point of a battery that lasts longer than our ownership of the device?
Ten-year battery life is designed for install-and-forget industrial sensors (which may be in a consumer’s house or car but used by their suppliers not by them).
Do consumers need network slicing?
No. Their carriers may use network slicing to manage consumer or Fixed-Wireless Broadband differently from other 5G applications or to manage B2B2C applications, but consumers will not be buying slices themselves. Businesses will buy these just as they buy VPNs today.
Industrial plants or smart cities may need millions of IoT devices per square kilometer but consumers will not. The densest city on the planet is Manila with 43,079 persons per square kilometer. Will each of them have hundreds or perhaps thousands of sensors on their bodies or in their homes so we would need millions of devices per square kilometer? Even in a Phillip K. Dick-inspired dystopian future? And would that happen within the service-life of 5G or would that be a 6G or 7G requirement? (And since large, dense cities like Manilla tend to be poor ones, who would pay for it?)
Consumers will benefit from Ultra Reliable Low Latency (URLL) in public services and security but my belief is that they will not pay for them or at least not directly. They will pay their taxes or pay their house insurance or pay for a home security subscription, but they will not pay their mobile operator separately for the communications component. That will be imbedded in some other charge.
Of the three axes of 5G – Enhanced Mobile Broadband, Massive M2M and URLL – only eMBB is consumer-oriented.
I know what some of you are thinking: autonomous driving! In my opinion, it is, (at best) B2B2C with the added-value B2C profits going to someone else than the carrier who built the network. (At worst, it may never happen or at least not within the service life of 5G. More on this next week.)
You could say I’m playing with semantics but I think the distinction is important. The previous G’s have been justified and paid for by consumers opening their wallets and paying for something they could not get before. In a B2B2C situation, the channel is different and, more importantly, as Telia’s Gabriela Styf Sjöman told us at the Nokia Forum, the distribution of value is different. Most of the value will stick with the supplier who has the relationship with the consumer, for example, Uber or Ford in an autonomous car case, because that is what the customer is buying: transportation. The consumer is not buying communications, however sophisticated and expensive that needs to be.
So my thesis is that, apart from eMBB, all of the sexy new features for 5G are for the enterprise market – B2B2C maybe, yes, but without an enterprise paying for the service in the ‘B2B’ half, there will be no demand. And without an enterprise that correctly identifies a profitable B2C experience, there will be no demand. So much consumer-oriented 5G needs two business models to work instead of one: the B2C model and the B2B model: more complexity and more risk.
And just to be clear, I am not saying that because the consumer 5G case is only based on eMBB that 5G somehow does not work. I believe it does. The eMBB case is compelling (although advanced 4G will satisfy much of the demand in the short and perhaps medium term, especially in emerging markets). The enterprise use cases are exciting and seem to hold great promise for companies to do all the things that companies should want to do: save money and enhance value for their end-clients.
I am just saying the way we have justified technology rollouts in the past – and the way that the American operators are apparently justifying their deployments – is on the consumer side and, at the very least, we have to think more creatively when doing business cases.
Even the US consumer deployment can be explained as a response to scarce spectrum: the operators needed more spectrum to continue to grow their MBB business, the spectrum that was available was earmarked for 5G and they decided they might as well invest in 5G as in 4G.
But this argument begs the question of whether there is enough such revenue to justify the deployment of 5G networks. Or whether eMBB is sufficient to deploy 5G and all that fancy IoT and URLL stuff just comes along for free. The calculations say that 5G is cheaper per MB or cheaper per MHz so even if there is no revenue uplift, maybe the business case ‘pays for itself’. Or, I suppose some combination of the three. The eMBB pays for part (either with new revenue or lower cost) and the B2B and B2B2C picks up the rest.
That is a much more complex issue and one for another day.
No Comments »