‘IoT is not about LoRa vs Sigfox anymore; it is about Design Process X vs Design Process Y’
The low-power wide-area (LPWA) end of the IoT sector is characterised by the capabilities and limitations of its main connectivity technologies. But this is wrong. These technologies – the likes of LoRaWAN, Sigfox, and NB-IoT – should not be so easily reduced to a checklist about throughput, payload, and roaming, and so on. Such a nerdy matrix of performance parameters is too limited to capture practical issues around deploying and scaling complex IoT solutions.
There is a whole other dimension to it, which stretches out in each case around the functionality of the stack, flexibility of the business model, and dynamism of the ecosystem. So says Wienke Giezeman, co-founder and chief executive at The Things Network (TTN), the Amsterdam-based group that has helped to precipitate and unite about 30,000 gateways, hosting about 800,000 devices, on public LoRaWAN networks around the world.
These other aspects correspond more directly with the likely success of each technology, as the sector matures, he says. They describe the ‘design process’ that goes into making IoT work, and dictate the shorter-term viability, longer-term benefits, and final-term logic of ‘massive IoT’. And for all of that, read: cost. And not just the upfront calculation, but the ‘total cost of ownership’ (TCO).
Because business doesn’t care about technology; it only cares about fixing problems and making money. But it is Giezeman’s point, and he should tell it. He explains: “These technologies are presented as commodities – as Commodity X versus Commodity Y. But it isn’t that. Because you’re not choosing LoRaWAN or Sigfox, or whatever; you’re choosing Design Process X versus Design Process Y.”
He adds: “Most of the time the design process defines the TCO, and the TCO determines whether the business case works.”
If success in the LPWA market, like every market, is dictated by longer-term costs, the TTN-end of the LoRaWAN community seems to have fixed the price about right. The latest numbers from the Dutch outfit, issued last month at The Things Conference, the group’s annual shindig on the Amsterdam waterfront (forced online again because of lockdown), say TTN is routing 600 messages per second and loading-on new traffic at a rate of 60-70 percent per year.
Giezeman says: “That is business data, flowing. And because the bandwidth is low, the value density is high. So it is the only metric that really matters. Chips-sold, devices-sold – those are not really relevant measures of success. What is relevant is that someone has coughed-up the TCO to install the sensors to keep the data flowing. And 600 messages per second, growing at that rate, says everything.”
But, as always, it is not so simple. Low-end IoT has been sold to the world on the promise of ‘massive’ scale, and failed to deliver. Of course it has; it is easy to multiply a small number by a big number, and come up with a massive one – and paint a future market around it. But easy maths does not translate to easy sales, especially when the margins are tight, the technology is fragmented, and the use cases are just so kaleidoscopic. Is it any wonder?
Speaking later, Giezeman talks about the sheer bloody-mindedness required to make it in IoT. “We’ve made a massive amount of screwups, like everyone else in IoT” he reflects. “But we’ve learned, and kept ploughing on. And at some point you succeed – and at some point you succeed more. That’s how we got here. You have to be able to embrace failure. If you can’t handle the heat, you have got to get out of the IoT kitchen.”
It takes true grit, as Mattie Ross would put it. Giezeman comments: “Because IoT is not easy; it is really hard. You have this massive cold start. But once you break out – and one in 20, maybe one in 10 break out – and once you’ve solved the business case and the challenge of scalability, it goes very fast. That is where we see this exponential growth.” Indeed, the traffic on its networks looks good; but what about that “break-out” figure, at five-to-10 percent?
There’s the bottleneck in the mission to scale. Speaking in a keynote at the same TTN event, the LoRa Alliance talked up the growing maturity of LoRaWAN – about it going from a period of horizontal-spread, across industry ‘verticals’, to a new phase of upwards growth as early adopters become early majorities in each. Giezemen says the same, talking about its “distribution across geographies and industries”.
But the sector is young, still, and that five-to-10 percent conversion rate for LoRaWAN sales, while anecdotal, says new maturity only describes the shift from bawling infancy to spotty adolescence. A multi-billion connected-everything, or whatever, is out of sight; the market is nowhere near stepping into the big-sized shoes it was fitted for. The LoRaWAN brigade, the story goes, are just following fastest behind, tracing these giant steps in the sand.
Giezeman comments: “People want to know about the killer app. But there is no killer app, just hundreds of them. Which is massively strong, and the reason it won’t go away, and is already delivering a lot of value for businesses.” But the problem remains: how to close the gap, cleaved by fragmentation and complexity, between the promise and the reality. “The market is sold on the hype – this idea that it costs $10 if you order a million units,” he explains.
He says something about “an ecosystem that [consequently] downplays its partners to sell products”. The audio is snaffled on Zoom; but the sense is the IoT market has to go beyond penny-pinching on hardware, and under-valuing its componentry, to make the business case stack up. The sales narrative has to bounce from the BOM to the TCO – from the one-time bill of materials (BOM), signed-off in cap-ex, to the total cost of ownership (TCO), teased-out in op-ex.
This article is continued here, under the header, Once upon a time in the (New) West – how LoRa is looking to strike gold on the IoT trail.