One of the core challenges of bringing the benefits of the Internet of Things (IoT) to life is getting started—how can enterprises successfully begin to adopt IoT when the solutions needed are so specific to a given business or business case, and off-the-shelf options have not yet sufficiently evolved? And how can the emerging technology evolve when getting started is so difficult?
Two sister organizations in Amsterdam seek to address both of those challenges together. One, The Things Network, is a not-for-profit organization that offers open-source tools for enterprises to build their own IoT solution on a low-power, LoRaWAN network that they own and operate. The other, The Things Industries, offers proprietary tools and support for scaling a solution once it has reached minimum viable product (MVP) status, helping companies produce, run, and market the IoT solutions they create.
Serial tech entrepreneur Wienke Giezeman is initiator of The Things Network, and CEO and co-founder of The Things Industries. He spoke recently with McKinsey partner Ferry Grijpink about the reasoning behind an ecosystem approach to serving the IoT needs of enterprises, and why it makes sense for companies to start small when adopting IoT.
McKinsey: What’s the barrier to enterprise adoption of IoT? Why aren’t we seeing as much value from it yet as we believe is there?
Barriers to enterprise IoT adoption
Wienke Giezeman: The challenge is that all of these puzzle pieces of the Internet of Things—the components that you need in your digital transformation—need to start to come together with business processes. So far, they’re not connected. Data measured by a sensor may end up in a dashboard in a little graph, but our experience is that when that happens, it’s really a sign of failure. Success is when that data point ends up in a business process, when it can save a worker 30 minutes of work every day. Connectivity and IoT sensors only start to make sense in a broader digital transformation context.
At the same time, you have this massive fragmentation of different types of IoT use cases all around the world, falling somewhere on a spectrum. On the thin edge, you might have a small data point, a temperature measurement in a fridge for instance, that can serve as a data point in a food compliance report. On the other end of the spectrum, the thick edge, there are large amounts of streaming data that all need to be analyzed in the cloud. And there are other use cases in between. All of these different problems and solutions need different types of connectivity, and it’s increasingly possible to build these networks yourself and tailor them to your own needs.
McKinsey: What about LoRa [long-range connectivity] and LoRaWAN [long-range wide-area networks] technology? What makes them a good form of connectivity for IoT?
Wienke Giezeman: LoRa has been around for a while already, as a technology in the segment of low power devices. It is good at building secure, last-mile private IoT networks, which can be deployed at low cost through simple gateways that are connected via cellular or fixed backhaul. Eventually these can serve as strong use cases for 5G networks.
The more advanced version, LoRaWAN, includes more network routing and connectivity features. This enables LoRa to address a much larger market and a longer tail, because it takes away a good amount of the overhead and complexity to build IoT solutions such as smaller, after-market sensors.
For these reasons, LoRa is now serving as sort of a beachhead for IoT—it makes adoption of IoT easier and less expensive.
McKinsey: Tell us more about its potential for specific verticals or applications.
Wienke Giezeman: The industry where we see massive growth is in food compliance, where you can put a simple sensor in an existing fridge to measure the temperature and generate food compliance reports. We are now in more than 100,000 fridges all across the world with our networks, and these are generating a lot of optimization processes in restaurants and other places where food is processed. LoRa allows you to enter the market very easily—you don’t have to change anything in the fridge. Of course, the next step will be how to massively reduce the cost of ownership, and integrate this technology in the fridge earlier, going from aftermarket to OEM and embedded Internet of Things, massively reducing the capex required. Having it embedded will create all kinds of interesting new business models too.
What we’re trying to do with our community and developer engagement is to remove the obstacles to building solutions and make sure they become the next growth diamond. For example, we are supporting the tracking of 10,000 cows in Australia. We help office rental companies optimize their real estate by tracking how many and how often meeting rooms are being used. We have companies that monitor air quality and ventilation for COVID return to work environments. We have hundreds of customers with all kinds of different use cases.
Simplicity is key
McKinsey: What should companies be thinking about to help accelerate LoRaWAN-based IoT? What type of investment needs to happen for it to scale further?
Wienke Giezeman: A few years ago IoT was a complex niche area of expertise that you could only do with an external specialist. That is no longer the case. The hardware, tools, and management software can be operated by the average enterprise developer, which enables companies to move in to this market with much less risk.
So now is really the time for enterprises to start their own IoT activities. They can start by shifting some of their existing IT resources to this segment without making additional investments.
And it’s not just in the area of talent that IoT no longer requires major new spending. One of the issues that we have in the IoT industry is that we have often ended up having to buy more equipment than we need. The beauty of what we now have with low-code, no-code, and serverless cloud, is that it lets us cherry-pick the tools. It’s the realization of the microservice dream from 15 years ago. IoT is just a small but critical microservice to feed data into your digital transformation platform, and now it’s time for companies to take matters into their own hands. It’s no longer that complicated.
Only when companies get hands-on with IoT can we begin to move it from aftermarket activity, where you put sensors in devices and things after they are produced, to embedding them in OEM devices and systems and in the digital transformation strategy.
McKinsey: What advice would you give to a CEO who wants to use IoT as a way to transform their business?
Start small and simple with IoT
Wienke Giezeman: For IoT my advice would actually be to go developer first. Build your Internet of Things ops as an in-house technical competence. That way you can control the simplicity—and simplicity is critical. We’ve seen this in the industry again and again—you cannot solve IoT problems with money. It’s so tempting to try to solve these problems with cash, but really it’s the creativity and pushing for simplicity that leads to the solution, which shouldn’t be so complicated. It takes critical thinking. And starting very small and simple. Maybe one small data point can have massive impact on your business processes. Setting up a department with a budget of $50 million might not be the solution, but hiring a small set of very smart developers may well be the first step to IoT success.
Making things happen
McKinsey: Talk a little bit about how you came up with your business model. It’s a bit unusual.
Wienke Giezeman: When we started our initiative, we saw two obstacles: the market is highly fragmented, and it is still very hard to build an end-to-end IoT solution. We thought that this market can only be solved by an ecosystem approach, so we created The Things Network to offer all of the tools you need to build the initial solution on an open-source basis. We tried to make sure that we provide all the tools and intellectual property [IP] you need to build your first proof of concept successfully. We call that a zero-to-one phase—you have an idea that you develop into a device that can be brought to production.
Then the second phase is one-to-a-million. We do that with The Things Industries, where we help companies with the tools and IP that they need to put their device into production, build out their LoRaWAN networks, and scale to a million devices. We contribute to the ecosystem by helping players scale their solutions, and our business model is we only make money when our customers make money.
All these problems that IoT solves are so different. Even just in agriculture alone, similar problems in different regions can require very different solutions. What we provide is not so much a platform but more of a workbench, where somebody can pick the right pieces they need to address their particular problem.
From an initial investment point of view, this may feel counterintuitive, because you could say we are allowing an ecosystem to leach on our IP and investments. But the alternative is building a silo, or a walled garden. And people forget that a walled garden almost always has a ceiling that limits your growth as well. What we’ve seen is that if we just open up, and make sure that we are that trusted workbench to build your solution, then companies will stick with us. It’s developer first, API first, workbench first—an ecosystem-driven dev model.
Risks and rewards in IoT’s future
McKinsey: As the IoT industry develops, how big are your concerns around cybersecurity and regulation of frequencies?
Making sure IoT is a social good
Wienke Giezeman: As we move toward that big dream of billions of connected devices we must also be cautious, and visualize how that world would look. How does this technology fit into society? We see three main concerns there.
One is privacy and security. Fortunately, what you now see is that security by design has been adopted so much over the last few years that it is starting to become an out-of-the-box technology, with end-to-end encryption and secure key enclaves, et cetera.
The second concern is e-waste. We’re sending out all these devices with batteries—how are we going to make sure this is a circular supply chain? How can we use existing circular economies, for instance, that are already in place for batteries to also be used for Internet of Things devices?
The third concern is utilization of spectrum and how that evolves. How are we going to divide that in a fair way? This is more of a long-term subject, but these are all big questions we also need to answer when we’re going into hyperscale in IoT to make sure that it finds a responsible place in society.
McKinsey: How do you see the future of Enterprise IoT?
Wienke Giezeman: When we look at the future of enterprise IoT, sometimes the present seems more crazy than the future. For example, there are operations where a significant amount of employees are doing work that is highly inefficient when they could be focusing on something much more useful for their customers. We were talking to a utility company that employs people to drive around and visit customers just to check equipment cabinets to see if the red light on top of it is blinking. Or pest control companies spend massive amounts of money to have people driving all around to check if a trap has caught an animal. Looking at that, you just see so much opportunity and untapped potential—for economic efficiency and economic growth.
McKinsey: What do you see the connectivity ecosystem and market looking like ten years from now? Does LoRaWAN and low-power WAN become integrated in an overall network? Will it be more private networks?
Wienke Giezeman: Looking at LPWAN in general we will see a movement from IoT projects with purpose-built sensors and devices to a world where the connectivity is standardized and embedded in the devices. But that will take a few steps. Right now we are in the aftermarket IoT phase, where we are putting sensors into existing fridges. You can get off-the-shelf IoT devices for LPWAN technologies like LoRaWAN, NB-IoT, LTE-M, 5G, etc. The next phase of IoT is embedded IoT, where it comes pre-integrated, and the fridge already has the sensor and the connectivity in it.
Several connectivity technologies will find their place in the market. We believe LoRaWAN will become a connected network of private networks supported by nationwide LoRaWAN public networks as a backup. This will all be secured by end-to-end application-level encryption. We already see that partner ecosystem around logistics emerging, where different private networks are shared across a transport supply chain to forward data from trackers embedded in goods.
The device ecosystem for LoRaWAN is now its largest asset. Our device catalogue already has hundreds of ready-to-order devices for many different use cases, and that will only grow further in the coming years. The tools for creating the end-to-end solutions with LoRaWAN will become more and more accessible to the enterprise developers and will allow even more companies to leverage IoT.