Bernd Gross on Building Successful IoT Projects
As CTO for Software AG and former CEO of Cumulocity, a spin-off from Nokia Siemens M2M and IoT Business, Bernd Gross is no stranger to working with leaders to adopt new technologies. He’s deep into integrating the Cumulocity IoT platform with the rest of Software AG’s portfolio of products.
How do you recommend IT leaders technically begin their journey exploring IoT?
That's a very good question because we, unfortunately, had a few projects that didn't go very well. And so we identified the reason. We did some post mortems together with clients and partners, and we came up with a concept to move forward with. I call it a three-phased approach that we optimized into seven steps.
The first thing that is important is to gain ownership of all internal data. The data integration and connectivity needs to be robust because if that's not solid, you cannot change processes internally.
In the first phase, we call it application-driven. You have to have a use case and keep the application in mind. Then you focus as a stand-alone application on your new IT technology and your connectivity and ensure it works. Once the application run phase is under control, you can manage it. You have new technologies, gateways, or embedded technology to connect to and extract data from.
The second phase is what we call process-driven. You begin to think about improving existing processes with the new data you are collecting. You identify the use cases without getting bogged down in technology. Getting to the use case nowadays is quite a creative process. There are a number of emerging methods that are very innovative.
I like very much business model canvassing. This means trying to figure out what the value of the use case is that we are trying to achieve. Which customer segment are we addressing? How are we messaging this? What are the key functionalities that we need? Who are the people, key stakeholders that can help us? Do we need partners? What is the cost structure associated with it? How should it be priced? Are there new pricing considerations in play? If you have that whole picture, you can explain it to the executive level, get a commitment, and you are drawn into the next steps.
The third phase is innovation-driven. That means new business models, new revenue models, and moving from Capex to Opex pricing. You start working with machine learning, artificial intelligence. Whatever is then possible and makes sense.
How does machine learning and AI play into the picture?
The danger of what we see at the moment in the market is that a lot of people want to jump immediately to step three. Those are hot topics, and they can sell these to the management teams, you know. There's a lot of failure inherent in that approach because if you don't stabilize your whole environment working through the first two phases, you can’t try to do predictive maintenance. For example, on low-quality data, it isn’t sustainable and not real. It fails. And that's what I meant in the beginning when I see projects failing.
People are trying to leapfrog into the new hype stuff before they control the new data flow and the new capabilities they have as a foundation. It’s like the children’s story of the three little pigs and the wolf. If you start with AI, you're building a house of straw.
How would you recommend developers start working with IoT? And with Cumulocity? Is there a way they can download some of this and play around with it?
That's what we encourage a lot. We have a free trial version you can sign up for in two minutes. It’s a completely open platform, open APIs with open documentation, and can be deployed in various infrastructure service environments. It can be deployed on-premise too. I remember a few years back, some of our competitors sent us emails congratulating us on how detailed our documentation was. They probably thought we would then close it, or whatever. I think they were getting pressure from their customers to open up the documentation the same way. But we kept it open because I strongly believe that the way we have to embrace the future is by being as open, as free, for our clients and partners to use our software as possible.
That's also why the API structure is still open. It's not just that we are opening a few API’s, 100% of our APIs are open and available. So, that's very important.
Do you feel that Cumulocity customers are going to find value-add the deeper they get into the product?
Absolutely. That's happening. That holds when we talk about the three phases, you know? For example, you typically are interested in visualizing your data information. Dashboarding is the simplest capability we have. But in the beginning you think, oh, that's super — a connectedness. I have a dashboard, and I can do stuff. But then the second step is the block flow automation. Then you appreciate that we have embedded a block flow automation engine to optimize your data.
After a while, clients usually tell us that the dashboarding tool has the smallest value, although they selected the platform in the first place because of its outstanding dashboarding capabilities.
In closing, what are the main takeaways for leaders venturing into IoT?
Apply the three-phase approach we developed the hard way through our failures to your organization, and you’ll have great success.