Manufacturing is currently replete with new technologies that can unlock unprecedented levels of productivity. Additive manufacturing, flexible automation and smart, connected devices can make mass customization cost-effective, free human workers from dull, dangerous tasks and transform our ability to see and analyze process performance.
However, many of these new technologies—such as 3D printing or collaborative robotics—are sometimes criticized for being solutions in need of a problem. Can these actually add concrete value to manufacturing operations happening today?
Relayr is one company focused on finding such concrete applications, specifically in the field of industrial Internet of Things (IIoT).
Guneet Bedi, Chief Revenue Officer at relayr.
What is your background with industrial IoT and how did you join Relayr?
I spent about twenty years in large technology companies, including Cisco, Oracle and Symantec. During this time, I came to understand how the technology works and how the business of the technology world works. I gained a product management, engineering and business development background.
I have known the CEO of Relayr, Josef Brunner, for almost ten years. Several years ago, he built a company called JouleX to disrupt the energy management market. I led the integration of the JouleX acquisition at Cisco in 2013. Then Josef invested in and joined Relayr. I joined Relayr in 2016.
At that time, similar to other companies in the industrial IoT space, we were at a very interesting juncture: we had figured out the basic technology on how to get data from complex machines to help industrial manufacturing companies. When I came in, the goal was to build the U.S. market.
Today, the journey has led to our being acquired by an insurance company, HSB, which is really interesting. There is a lot of technology developed over the last few years, and now it’s time to make this technology solve real business problems, which really excites me. It’s much more exciting than selling data centers, switches and routers.
What does Relayr do specifically in the manufacturing industry?
We think of the manufacturing industry in two or three different parts. Number one, there are the operators on the assembly lines and, of course, the company which operates the assets.
Number two, we think of companies who actually manufacture the equipment, the machine builders. Number three is companies who service these assets. All three of these, surprisingly, have been good markets for us.
On the operator side, it's more about getting digital insights and monitoring conditions. How do we increase efficiency? How do we predict bottlenecks at the line? We are really familiar with this space because of our work with automotive suppliers. They need to be efficient to manage to supply to the Fords and the BMWs of the world.
The other two categories are faster-growing markets for us, because, as you know, significant capital is spent buying these machines, but then a lot of money is also spent during manufacturing—maintaining and operating the assets. This includes spare parts, service, efficiency, downtime, etc. We are focused on creating the equipment-as-a-service business for them, so they don't need to sell machines anymore, instead selling per-hour of the outcome on utilization. We bring not just the technology, but we also handle the finance and the insurance behind it.
You mentioned providing a product as a service. Is this monetization of industrial devices a growing trend in the industry?
It really is. And we have found that many machine builders believe the place to be is in IoT or in providing connectivity features that will drive revenue. So, we say, why don't we start there? Why don't we help you figure out the best way to monetize these connected industrial machines? Start from there, and we’ll figure out what the right solution is.
For a machine builder, it makes sense why you'd want to move to this ‘anything-as-a-service’ (XaaS) model. But from a customer's perspective, it may seem that you're switching from owning your home to renting it, metaphorically. So, how would you say a product-as-a-service model provides value to customers?
That’s a good question. It relates to the larger issue of guaranteed risk. You’re right: If we just give you, "you can buy a machine, or you can buy a machine per hour", that doesn't help. Actually, it might be worse. There is a question of who is maintaining the machine. I’ll give you an example.
One of our customers does aluminum heat treatment in automotive. They said to BMW, "Hey, instead of you buying a million-dollar machine, how about you pay us by cost per part treated." To their customer, this is great because it reduces OPEX. But now, the supplier is maintaining the machine. If that machine goes down, my entire supply chain is affected. I don’t just care about heat treatment, I care about the whole car being ready, right? And that’s where breakdown insurance comes in.
HSB, our parent company, is the largest equipment breakdown insurer in the world, creating actual models to find when machines will fail. But all of this has been based on historic data, which is where we come in. In addition to historic data, we create intelligence with real-time, centralized data. This allows us to provide more sophisticated and more economical policies for equipment breakdown.
With this model, we can offer customers a certain OPEX, and assume responsibility for machine uptime, including maintenance, spare parts, and even for providing understanding of how to continuously improve the speed and quality of the work. Under this model, total cost of ownership of the machine should go down for the end user.
That’s interesting: You’re using predictive analytics of machine data to set insurance policies for equipment breakdowns?
Exactly. Insurance against breakdowns, but also against downtime—so, it's not just the breakdown of the machine, but if the breakdown of the machine causes a cost of a million dollars a day, we insure a million dollars a day.
What would you say is holding back IoT adoption in manufacturing?
I think, one is just the conservative nature of the industry. If it's been working for fifty, sixty, seventy years, why change it? I think getting to the right business model and producing the best ICO is important, but I think that is also the biggest hindrance.
When a company decides to implement an IoT solution, we’ll do a pilot project, which means getting the first few devices connected. This helps answer: what can really happen? How will this help in day to day business? I think that's definitely one factor. The second factor is that I think industry does a disservice to the concept of predictiveness. The way we make predictive analytics seem, it sounds really easy—but it's not. Getting to anomaly detection and unsupervised learning is pretty easy, at least that's what we specialize in, but that doesn't virtually predict the system to a full extent. And you know, eventually you can get to a predictive system, but a lot of people think they will get to predictive too soon, and then it doesn't meet the promise.
That's a very interesting point, because in writing about industrial IoT I've noticed that there's a lot out there about the opportunities, the applications, for things like predictive, remote management of assets. A lot has been written about all the potential, and all the hype, but there's a lot more interest now for information on getting things up and running, requirements and costs. The more concrete side, beyond the hype to the real nuts and bolts of how to get something up and running. So, it's an interesting point you make that hype is actually doing a disservice to the adoption.
Yes. I know Silicon Valley people love hype, I lived there for eleven years! You can make a case that when it comes to the Indiana motor shop, for example, hype would not necessarily be good, and facts and figures are probably more valuable.
For more information on how IIoT is changing manufacturing, check out the article The Connected Factory and More: 5 Examples of How IIoT is Changing Manufacturing.