Jon Lindén, CEO, Ekkono

What are you waiting for? The Internet of Things (IoT) has been on everyone’s mind since the beginning of the century. After connecting everyone, we started connecting everything. Things with sensors become accessible over the Internet from a far, which should enable the fourth industrial revolution. And so it is, right? But why is it taking you so long to connect your products?

I don’t actually know. We are beyond the point of determining if things will get connected. Every product company I talk to confirms that it is a question of when. While the business models for IoT have not always been proven – you see, someone must show the way – there is no doubt that customers will pay for better SLAs, less downtime, less emissions, lower power consumption, performance indicators, and other valuable insights and expertise that the OEMs are best suited to provide. If anything, the pandemic has demonstrated that we cannot take for granted to go onsite to service products all over the world. The technology is there, and the price is right. In short, IoT as a concept is here to stay.

Already in 2020, Juniper Research predicted in a study that the total number of IoT connections will reach 83 billion by 2024, rising from 35 billion that same year. This is not an unreasonable expectation as everything, from large to small, from trucks and elevators down to smart sensors and running shoes, get connected all around us. And while consumer products are fascinating, the industrial sector (IIoT), including manufacturing, retail and agriculture, will account for over 70% of all IoT connections according to Juniper’s study. Which makes sense since uptime and performance are critical and drop straight down to the bottom line.

Edge Machine Learning and IoT

My fascination for IoT comes from the potential I see in product OEMs extending the relationship with their products to after they leave the factory – out into the field. This requires automation, or Smart IoT. And ever since I first met my co-founder, our CTO, Rikard König in 2015, when he was completing his university research that is the foundation for Ekkono’s products, I’ve been totally absorbed by how Edge Machine Learning, or Edge AI, empowers Smart IoT. But to release this ginormous potential, things need to be connected – first you connect things, then you make them smart. IoT is one of the most intuitive places for the use of AI (artificial intelligence). This is the reason why I’m impatiently monitoring this development.

So why is it that machine learning works so well with IoT sensor data? Most products already have traditional sensors like temperature, pressure, accelerometer, and vibrations. These are cheap today and not necessarily your traditional big data generators. But they provide comparable data. And when combined, and decorated with lags, sliding averages, etc., they become a strong dataset that offers a comprehensive view of how, where and how much the product is being used. Edge Machine Learning is actually just another tool to make use of this data. The real value of Edge ML is that you can do custom learning per individual device, that you can learn more the more the product is being used, and you can predict what will happen – or at least what is supposed to happen. This opens a box of business opportunities.

Those who award the future have seen and understood this. In the last quarter alone, Ekkono has been awarded Red Herring Top 100, Siemens’ Future of Energy Awards, and Dena’s SET Demo Day. I think you as a product OEM have understood this as well, but you might struggle to get this off the ground at scale. It tends to become a catch 22 where you try to develop business cases to justify the investment in IoT, but the business cases need data that can only be retrieved by connecting the products. So, unless you are hesitant about whether you will connect your products or not, this is the time to stick your neck out and just do it. This will set the ball rolling, and you have a chance to come out of the gate ahead of the competition. Don’t boil the ocean – set a roadmap with incremental steps (just like we software companies do), but most importantly, do something (as opposed to doing nothing)!

Industrial Internet of Things ­– IIoT

When you connect your products to the Internet, it enables you to track the product, measure its use, check how it feels and performs, update it with new functionality, and support it remotely. All of which are costly manual features, but perfect for automation; Automation that requires some smarts. But you should always approach this from solving a real, relevant, and urgent problem. Otherwise, it becomes a science project. Nine out of ten cases start with maintenance, and the ambition to migrate from reactive maintenance based on error codes, to proactive maintenance based on predictions.

Smart IoT enables this predictive maintenance (PdM). The PdM market reached $5.6 B in 2020 and is projected to grow to $28.2 B by 2026 at a CAGR of 31% from 2021-2026 (State of IoT 2021). To get there, I think you as a product OEMs have to step up your game. And you should since 83% of the PdM end-users reported a positive ROI, and 44.5% reported amortization in less than one year. With the ongoing pandemic and a remote workforce, service and maintenance staff need tools to do condition-based and remote maintenance.

Smart IoT needs IoT

The point I’m making is that I don’t think the IoT industry has taken off the way it should and could have done. There are many reasons for that – we always overestimate the short-term development, there is and has been a pandemic, and we probably underestimated what it takes to connect things, and in particular to connect the installed base. You see, your customers won’t replace their current product just because there is a new IoT-version of that same product. But I also think that it has been slowed down by an eternal decision loop where many end up doing nothing. For now. That’s why I’m asking “what are you waiting for?”. Now is a better time than ever to get going, and in particular if you are already convinced that you will connect things, that you want to become more data-driven and service-centric, that you want to act and support your products all over the world, that you want to offer added value, and that you don’t want to be left behind with the old-fashioned product in the market. So go for it! And let me wrap up by giving some advice on how to get started:

  • Start with maintenance – uptime and remaining useful life of crucial wear parts are concrete and quantifiable cases.
  • Make it a strategic initiative to offer better service on your products by transitioning from reactive to proactive maintenance.
  • Manage every product individually, just like you do with your customers. Manage them based on where, how, how much and for what they are used.
  • Start with solving real, relevant and valuable problems – the IoT project should relieve you by solving an issue you are working on rather than creating a new project on top of all the others.
  • Become your own first customer – it takes a smaller effort, you will learn how to use the technology, and gather insights and evidence that you really can deliver what the end-customer needs. After that it is a small feat to build a business case.
  • Think roadmap – do small incremental steps, but ensure that you enable yourself to make upgrades and changes in the future; Become a software-defined business
  • And call/email us if you want to know more. Don’t forget that “doing nothing” is also a decision, and often the worst of decisions.

If you want to read more predictions about Edge Machine Learning and IoT, I would recommend these two industry analysis:

Top 5 Edge AI Trends to Watch in 2022 – NVIDIA

10 IoT technology trends to watch in 2022 – IoT Analytics

Good Luck!

Leave a Reply

Your email address will not be published.