Every year I head out to Barcelona in late February for Mobile World Congress. This year was my 18th. And probably the best. I am a self-confessed MWC lover. The opportunity to sit down in 50 meetings with the most important companies in IoT allows me to better understand market dynamics and company offerings and strategies, and always gives a fresh perspective on things. This year the biggest theme was this: the customer is not always right, and don’t expect your customers to know what they need.
But before we get going, the overwhelming highlight of the week, and possibly of any MWC week, was our inaugural MWC IoT CEO Dinner on the Sunday night. We welcomed a select crowd of the CEOs and business unit leaders of the top IoT players to join us for an unparalleled night of networking and reconnecting, and not to mention amazing tapas. It was quite humbling to be joined by the top brass from 1NCE, AT&T, BICS, Blues, Cubic Telecom, Datora/Arqia, Deutsche Telekom, emnify, Eseye, floLIVE, KORE, KPN, MultiTech, NLT, Pelion, Quectel, Soracom, Swisscom, Telefonica, Vodafone and Wireless Logic. I had already spoken with 80% of the people I wanted to meet in Barcelona before I even set foot in the Fira. Thanks to Frank and the emnify team for sponsoring. As a departure from my key theme of the event, certainly Transforma Insights’ customers were right in thinking it was a fantastic opportunity to get together.
Now, back to my over-riding theme. Customisation is a killer. From Connectivity Management Platforms to hardware to connectivity options, it’s the special cases that ruin profitability. Speaking with one disruptive IoT connectivity provider it has an approach of pushing a very small number of leading offers, having a selection of standard alternatives in the back pocket just in case it’s not appropriate for the customer, and if the customer needs something much more customised, they’re shown the door. Not every dollar is equally green, and in the world of $1 IoT margins depend on that. In the past we have talked a lot about hyperscaling IoT connectivity. That’s not just a technology question, it’s also a question of commercial models. And there is a trend of ‘constructive client dismissal’ (if I can term it that way) where organisations are electing not to serve particular customers that they think can’t be done in a cost-effective way.
This idea that the customer is not always right is particularly applicable in the Mobile Private Networks (MPN) space, one which Transforma Insights has been keeping an eye on for many years. The potential of private 5G (and 4G) has been touted for many years, but we are yet to really see a flurry of major companies embracing it across multiple sites and truly embedding it in their business-as-usual operations. There are exceptions, of course, and many trial/testbed deployments demonstrating some amazing capabilities. But it has not skyrocketed in the way some (but not us) might have predicted. It’s expensive and it’s a complex new technology to manage.
However, there is a very bright side. The learnings that are derived from MPN deployments in the form of S-NPN (Standalone Non-Public Network, i.e. a dedicated network) will stand MNOs in great stead when it comes to deployments of PNI-NPN (Public Network Integrated Non-Public Networks, i.e. network slices on the public network). It is always strongly suggested that companies need their ‘private’ 5G to be based on dedicated infrastructure. The argument goes that the companies want to own the infrastructure for reasons of security and control. But don’t expect your customers to know what they need. When faced with the complexity and expense of deploying, managing and upgrading an S-NPN, compared to the fully managed PNI-NPN at a fraction of the cost and including features like local breakout, many will find that perhaps their demands for ‘their’ network aren’t quite so rigid. Of course there’s still a way to go before PNI-NPN is really feasible in most places. 5G Stand Alone cores need to be fired up and coverage needs to be improved, but ultimately we expect the economics will win out. In most cases.
There are also a few more subtle ways that the trend can manifest itself. Telenor Connexion has a small range of IoT Complete products which comprises a connected module or modem supporting LTE-M or LTE Cat 4 (plus fallback) as standard offerings, and including a few standard sensors and interfaces, along with the associated management platform. This represents a baseline set of products to steer customers towards in preference to starting from scratch for every deployment. This offering isn’t unique of course, but this type of approach does play to the theme of avoiding customisation and steering customers towards an off-the-shelf proposition wherever possible. Take the LTE-M/Cat 1 product. If the customer had been consulted they might have demanded Cat 4 as well, or NB-IoT, or several other technologies. Sometimes suppliers need to be bold enough to say “you don’t need that, you need this”, rather than “of course” every time.
This situation isn’t helped by the growing plethora of new technologies coming out. 5G RedCap is the latest. A technology which got quite a lot of coverage at the show, but which we remain convinced is one that is optimised for nothing in particular. but the arrival of which causes some confusion. Whereas the addition of Release 17 NTN simplifies using satellite (by using the same device as for NB-IoT) and which we applaud (and more on which later), the arrival of a new IoT-oriented standard just causes more uncertainty and delayed decision making for the customers, who are wondering if they are backing the right horse.
Another area where we can’t expect customers to know what they need is in compliance. If there is one of our 2024 IoT Transition Topics that stands out as being the biggest headache in 2024 it’s regulation. New rules on national resilience, data sovereignty, and other associated topics are either already here or pending in many countries and regions. Navigating those rules is a headache in the extreme. And it’s not one that the customer should have to worry about. Or perhaps even have to think about. Which is why features like Aeris’ new Watchtower are interesting. The headline is about security, but the big underlying value is compliance, providing auditability for the way the connectivity is provided and some level of guarantee that security requirements have been met. Compliance-as-a-service is an emerging feature that customers demand.
If I was being trite, I might suggest that the new SGP.32 standard also allows customers to not know what they want, in terms of which operator’s connectivity they want to rely on, and opt to select it in field. It gives freedom to switch between carriers without the complexity of the SGP.02 M2M standard. We have published many reports and blog posts on this topic recently, including this. As a technology goes it is perhaps “third time lucky” as Paul Bradley of Kigen suggested on the panel at the 5G IoT Summit. And the technology is a marked improvement. But that’s only half of the story. There is still some uncertainty about the commercial models. Who does what and where? On my 6.5 hour train ride from Paris down to Barcelona I set myself the task of working through all the implications; something which I didn’t quite manage to finish. However, the week in Barcelona gave me the opportunity to test a set of hypotheses.
SGP.32 gives enterprises the ability to switch between operators. But will they? They will need commercial relationships with multiple operators, they will need the SM-DP+ and eIM capabilities to manage the switching, and they will need to manage the back-end implications of the switching. It’s not straightforward. Of course, this might all be handed over to a third party to manage. In the case of the SM-DP+/eIM element it’s almost inevitable in most cases. But what kind of third party. Certainly we’re unlikely to see an emerging cadre of specific managed eIM providers. A few months ago for a client project I dug into the viability of such an operation, considering how many switches might happen and how much could feasibly be charged for such a switch. The conclusion was that the total market in the US for such a role was less than USD100,000 in 2025. Hardly something to build a business on. The opportunity therefore is to stitch it into another offering, be it Connectivity Management Platform, Remote SIM Provisioning Platform or, and this seems the most likely, MVNO. The latter already have established commercial relationships and integrations with the carriers. Whatever happens we have about 12 months to work it out before SGP.32 is available.
At this point I should also note that eSIM/RSP evolution does not end with SGP.32. There is also a process ongoing to standardise something called In-Factory Profile Provisioning (IFPP) in the form of SGP.41 (and with SGP.42 as the actual standard to follow). This allows for a manufacturer to handle the SIM profile upload while still in the factory, rather than waiting until it gets into the field. There are numerous advantages to this, not least being able to save a chunk of battery life that’s needed to handle in-field provisioning after it has left the factory. Great news for battery-powered devices. Check out our recent White Paper for Kigen for more details. And further on SIM technology, we rather like the rSIM technology that CSL was promoting on the bustling GSMA stand, as another option for supporting critical connectivity. If you want to know more, check out our recent webinar on resilient connectivity with CSL and Vodafone.
At the risk of being hypocritical, I should also note that the advice that the customer is not always right does not mean that there shouldn’t be any consideration of what the customer needs. Far from it. The reality is that the biggest unmet opportunity in IoT is to ‘hold the hand’ of the customer as they deploy. It is notable that in an era of high interest rates where over-leveraged companies focused on a highly horizontalised (and theoretically infinitely scalable) approach have struggled compared to those offering a less scalable approach involving understanding the vertical in which the company operates. Such considerations are particularly notable in the IoT MVNO market where we see success for some players based on delivering value in particular segments, such as Kajeet in education, or CSL in security, which contrasts starkly with many providers of undifferentiated connectivity, notably those that are based on a white labelling of an MVNE offering. We expect consolidation in that space and that can only be a good thing.
This also extends somewhat to some perceived exit of the likes of IBM and Microsoft from IoT. The reality is that these companies have not exited at all. They continue to perform a lot of functions that are critical to IoT. They just never found a way to make a success of a universally applicable, all-encompassing IoT ‘platform’. I’ve commented on this many times before, including a LinkedIn article about the great IoT bandwagon. The twin ideas that the IoT space has been undergoing some ‘deplatformisation’ (i.e. less obsession with needing to be a platform and focusing more on serving customers) and that it is moving back more in the direction of being served by specialist companies rather than generalists, were a couple of themes that I discussed in my IoT State of the Union at the 5G IoT Summit on Wednesday morning.
In that self-same session, I also touched on many of the key themes that we see dominating the market this year. In particular, I discussed our IoT Transition Topics for 2024, which encompasses many of the themes I discuss above. A Transition Topic that definitely did get a lot of coverage was Release 17 NTN, which continues to be an intriguing opportunity. Deutsche Telekom IoT particularly seems to be getting ahead of the game with agreements with satellite providers, most relevantly Skylo, and what we think is the first fully fleshed out commercial offering. We see NTN as really a value-added service (for redundancy/coverage) for a cellular connectivity offering and one which is definitely monetisable; there are lots of scenarios where adding in satellite fallback could be very useful. I should note that I remain a LEO-sceptic because of the economics of keeping a constellation in operation.
It's also interesting to note what didn’t really come up. While I said during my 5G IoT Summit session that regulation is set to be the defining IoT issue in 2024, hardly anyone really was talking about it, notwithstanding the Aeris product discussion above. I’m worried that many companies aren’t giving it the focus it deserves, although maybe it’s just that the work is going on furiously behind the scenes. In the conversations I had, there was also precious little about resolving some of the challenges of NB-IoT and LTE-M. Within the GSMA there are initiatives to resolve that and we’re hopeful that the initial flurry of roaming agreements, which petered out a little bit in recent years, will be reignited.
Altogether it was an amazing Mobile World Congress that confirmed some views, challenged others and created yet more. But there’s no real opportunity for a breather in the coming weeks. Next week I’m in Stockholm kicking off the first discussions with connectivity providers for the 2024 iteration of our Communications Service Provider IoT Peer Benchmarking report (press release for the 2023 iteration is here), and then it’s off to the Channel Partners event in Las Vegas, to dig a bit more deeply into another of our Transition Topics, around channels and partner programmes. After a few other client visits in the US it’s off to the Industry of Things World event in San Diego.
At Transforma Insights we have always set out to be immersed in the IoT world, rather than just watching and commentating from the sidelines. Our immensely successful IoT CEO dinner hopefully proves that to be the case, but it’s just one aspect. Beyond that we will always prioritise getting face-time with all the companies in the field.
