In order to give you a better service Deutsche Messe uses cookies. If you continue we assume that you consent to receive cookies on all Deutsche Messe websites. Legal Notes


Relayr CEO Josef Brunner: "We Germans are the Chief Worry Warts"

The Berlin-based startup Relayr wants to connect old industrial plants to the web. But CEO Josef Brunner prefers not to talk about the Internet of Things. He explains why in an interview with

05 Oct. 2017 Source: t3n Lisa Hegemann
Source: Relayr
Everything began with this toolkit (Illustration: Relayr)

Relayr is the flagship German startup when it comes to the Internet of Things. This Berlin-based startup has set itself the ambitious goal of finally digitizing German industry. It wants to network existing facilities with retrofit kits and turn conventional mechanical engineering firms into service providers. The company has raised almost USD 37 million in venture capital for this idea from coveted investor Kleiner Perkins – and won over large groups such as Munich Re as customers.

Even so, CEO Josef Brunner doesn't like to mention the Internet of Things or IoT when he's talking with customers. He explains to why these terms are intimidating, why Relayr is now offering credit and why Germans worry too much. Josef, with Relayr you're taking on a Sisyphean task: you want to bring Germany's old industrial machines into the web. How do you explain to a small business owner what the Internet of Things is?

Josef Brunner: I don't explain either the Internet of Things or the technology to them. That's one of the key lessons we've learned in the past four years. Can you explain?

When I sit down with the CEO of a chemical company, he or she may have never even heard about the cloud. So the last thing I need to do is start talking about the Internet of Things. The just wants to know how much can be earned with a networked machine, how it can be sold shareholders, and how earnings per share will shift. So I get that across in such a way that the CEO can easily grasp it. How do you accomplish that?

I explain what he or she can achieve with our technology, for example in terms of saving money, boosting productivity or helping to get a new digital business model off the ground. That is what is relevant to a CEO – not some label we happen to be using for the technology involved. What kinds of solutions are we talking about?

It depends on what the customer needs. Small manufacturing companies are seeing huge changes in their markets, and noticing that their customers suddenly also want services. This creates a major challenge for them. We offer models that allow them to offer not just products, but also services. The second big segment is what we call the industry whales. So the big groups?

Right. They're looking more for savings potential and can see that many new digital business models are taking shape that make their old production pathways more efficient. They want us to network their industrial plants and provide comprehensive support. We do this with retrofit kits. And the third group is the GEs and Ciscos. They mostly use parts of our technology to complete their own solution. Usually device management is the topic here: How do you network a machine, how do you keep it secure, how does the software updating work?

Everything began with this toolkit: The Wunderbar by Relayr allowed developers to build their own Internet of Things applications. (Illustration: Relayr) So you build an individual solution for everyone?

Not necessarily. For most companies we generate the same data points. For example, we can learn a lot from acoustics. If the doors of a lift squeak, we can learn from the sound that they don't close correctly. Or we can see from the laser that the lift isn't stopping in just the right place. Technologically speaking it's not too hard, and can be transferred to other areas. The challenge is the mathematics behind it all, and making predictions. Can you give us an example?

Source: Michael Hübner
(Photo: M. Hübner)

One of our customers is a company that cleans water for large industrial parks and automotive companies. The water has to meet specific standards to be returned into production or back to nature. It's important for the company to know if the water isn't clean enough to be used again.

To achieve this we network the sluices and water tanks. Sensors let the company see where the process isn't working optimally. Or there's an alert if a tank needs to be serviced because the aggregate data show that algae has formed there. We've seen in recent months that networked machines can come under attack. How do you protect your customers?

Developers tinker with an application in Relayr's old offices. (Photo: Michael Hübner)

It's important to say first of all that we're one of the few companies with no dependency on the cloud. Our customers don't need to use our cloud, they can store their data on their own proprietary computers. That's very expensive, so few customers do it. But for us it's a major distinguishing factor. And we can manage hybrid models. So some data is in the cloud, and some isn't.

Correct. A company can decide that its production data never leave the factory, for example. In that case we create a metaset of new data to send on, and the machines don't need to be connected to the Internet at all. But you also offer cloud solutions. What happens if they're attacked?

You have to start by defining the scenarios: In one it's the infrastructure that's attacked and someone uses it against the company and their customers. That would be the classic blackmail scenario: „If you don't transfer me x amount of money, you won't regain access to your data.” I would say this is a comparatively more harmless situation. Harmless?

Yes. I know that no one has gotten into my system and I can close the gap. In a different case the hackers create brand new business models from the data, such as when a hacker invades a system and no one figures it out. Or they sabotage the plant. That is much worse. Why?

Take a pipeline that's been digitized with pressure sensors. It's very easy to take over control of the sensors and transmit false pressure data. That can lead to pressure rising and the pipeline exploding – a horror scenario for companies like GE and Cisco. That's why they use our solution, because we have a strong security focus.

"Without digitization you can be secure, but also definitely broke in five years." So exactly how do you prevent that in practice?

We have three blocks: With device management we network devices. With middleware all the data streams come together. And at the top is an application with dashboards and analytics. The middleware is the core, with a security net over it. We can use algorithms there to see whether the data make sense or if there are things that stand out. We bought the startup Neokami from Munich for this at the start of this year. It analyzes data streams using artificial intelligence and checks whether the data that a company is gathering is valid. This is called anomaly detection. What does that mean?

Let's stick with the example of the pressure sensors. If someone hacks a sensor and changes the pressure, our security net recognizes that. Then it checks whether the pressure is changing at other sensors. If not, that would be the first anomaly. The second notion is additional sensors. Now we have something physical: If pressure is higher there might be a different air displacement in the pipeline. Our system can also measure the density. If the data don't correlate, this confirms that one of the sensors is providing false, invalid data and an alert can be triggered. Even with this system, hackers will always manage to get into a system. Aren't your customers worried about this?

Absolutely. And these worries are justified. But we Germans are amazingly good at worrying. I like to describe us as Chief Worry Warts. You also have to realize that Germany is very strong in the industrial sector because German companies build good machines. Now the economy is shifting towards software and service models. Other countries are stronger there. If companies would rather stay offline, fine. Their intellectual property might be safe, but they can also be sure that they’ll no longer be around five years down the road. A bold statement.

Of course I’m exaggerating a bit. But I firmly believe that companies have to take the risk, because otherwise they won't be able to offer the same service as their competitors. Now Relayr is not a decades-old company that everyone knows, but rather a startup. Companies have to trust that you'll still be around in 10 years, that you'll still be delivering software updates 10 years from now. How do you convince them of this?

First of all with our technology. And the fact that we boast a big financial cushion is certainly another reason. Another key point is our partnership with Munich Re as an ultraconservative insurance company that instills confidence. The Munich Re reinsurance company joined us in 2016. With this partnership we're guaranteeing the production savings that we had promised. This lowers risk for corporate customers. But that also means the risk lies with Munich Re. What does the insurance company think of this?

We pay a perfectly ordinary insurance premium and we assume we'll be able to keep our promises. How did this collaboration come into being?

It wasn't our idea. Munich Re was looking for partners in the financing sector. And they came to us.

That's why Relayr does credit financing now. What does financing mean in this context?

When a company networks its systems or wants to begin offering services, it's very expensive. It needs to have savings available for this. And not all companies do. That's why we buy the sensors and hardware, take care of the networking and prefinance the equipment. The customer pays us a monthly or annual fee for its use. That sounds like a kind of lending operation.

It's really mezzanine capital. But we try to not focus on that in our discussions. Because it gets confusing when a company builds hardware itself, is a service provider at the same time and then also does something with financing. Our core business remains our technology. How will the prefinancing model amortize for you?

We're betting that customers will opt to work with us long-term. Because it's unlikely that a company will go looking for a different technology provider after two years. Our solution will be so strongly integrated that it would have to find an adequate replacement first. But you're not yet making any profit.

No. But we've collected USD 37 million in capital and have a positive cashflow. That's because some of our customers pay us one to two years in advance. But these revenues are only prorated in the books. So in spite of the high cashflow, we're still registering losses. But profits aren't an issue for the next few months. Maybe we can talk about that point again in 2019. Josef, thank you for the interview.