Podcast episode #5: Unlocking the power of data

ZARI VENHAUS: Welcome to the next episode of "The Making of What Matters." There's a lot of talk today about digital transformation, and I'm very excited to have our Chief Digital Officer, Aravind Yarlagadda, here with me to talk about IoT, digital transformation, what's happening in our industries, and some of the work that we're doing at Eaton. Hi, Aravind.

ARAVIND YARLAGADDA: Hi, Zari. Thanks for having me over.

ZARI VENHAUS: We're glad to have you here. So you are the chief digital officer at Eaton and have been in this role since the beginning of 2020. And we brought out a chief digital officer so that we could really focus on digital transformation for ourselves, but also on how we can bring value to our customers using data. Could you talk a little bit about the importance of data and some of the things that are happening in our industry today that make reliance on data so important?

ARAVIND YARLAGADDA: Today's world, everybody talks about digital transformation. I mean, every company is going through digital transformation. Every industry is going through digital transformation.

The context of data is really important. When a company talks about digital transformation, it sends a message to its own consumers and users. What it really means is, there is a lot of enterprise data. There is a lot of device data. There's a lot of personnel data.

All of this data needs to be put in context of what it means from an internal productivity standpoint, whether it is factory, it's manufacturing, whether it is employees. I mean, their own productivity, their own processes as how you deliver value to customer how they deliver value to external users and in terms of how they execute their work processes day to day. So that's one prong of how you view data's value.

The second is, there is a lot of data that's trapped in a customer's environment. There's a lot of data that's being created that's not being homogenized on a regular basis. The value to a customer that Eaton can offer is, can we help the customer make sense of what the data really means in the context of their operations? And what kind of data is really valuable to the customers, whether it's Eaton device data, whether it's third-party device data, whether it's our software data, whether it is data from a variety of services that we offer.

So putting that all in context, processing that specific data, offering those insights, and using those insights to actually feed analytics packages or software suites that can deliver ultimate operational value to customers-- that, to me, is the real value of visualization through data.

ZARI VENHAUS: Can you talk a little bit about some of the macro trends that we see happening in industry and how we think that's influencing our focus on, but our customers' focus, on digital transformation?

ARAVIND YARLAGADDA: Zari, this is a very important question. And I want to touch upon it very sensitively. The COVID-19 pandemic, while it is unfortunate in terms of the impact it's had on the social life and, in general, the lives of people, it has accelerated digital transformation significant. It has at least accelerated digital transformation by five years. Let's say-- I mean, if I'm at an industrial company, if I have a ten-year journey or an eight-year journey for digital transformation, I would want to get that done in three years now. I don't want to wait for five years.

Number one, we've proved that we can work effectively and efficiently with digital tools, wherever we are in this world. This pandemic has forced us to work in that fashion. It has also forced different kind of economic changes on supply chain, on our factories and the way we work day to day in terms of how we process information. Yes, it has stretched the day more. But at the same time, it has made us more effective and efficient.

So this pandemic, COVID-19, is a major, major driver when it comes to acceleration of digital transformation today. And it is here to stay. I mean, the digital transformation impact that it's had, it's here to stay.

The second trend that I see is the retiring workforce, the aging workforce. While the domain-centric aging workforce is retiring, you need to manage the transition very carefully in terms of training the younger employees in terms of how they adapt to the domain. Make them love the domain. Make them love the domain, and help them bring their own tools and skill set to the table-- digital skill sets to the table and enhance the domain-centric value that we provide to our customers. That's the second trend.

And, of course, everybody talks about data and assets and all that, but what is really driving this? If you talk about-- if you take a look at the grid, there's aspects of the grid that need to be added. I mean, specific regions, geographies-- the United States as an example, there's current pieces of the grid that will need to be hardened so they can be more resilient. So digitalization can actually help modernization of grids and our modernization of assets in the grid, or milking the existing assets, milking the existing equipment so that they can derive more longer-term value to the end user. So that's one trend. It's basically aging assets.

And last, but not least, the way we actually work, digitalization is actually driving more cooptation-- not competition, but cooptation. Yes, I mean, how do we drive the price point down of a specific offering and value that we provide to the customer? But at the same time, it's forcing us to actually partner with a lot of other technology providers out there and deliver the best package solution.

ZARI VENHAUS: And if we think about 100 years-- over 100 years of power management expertise, so really understanding the applications, where our products are used, how they're used, but also the relationships that we built with customers to really deeply understand their applications, a part of the value we can bring is taking all of that knowledge and that expertise and building that into algorithms, machine learning, artificial intelligence. Can you talk to me a little bit about some of the things that we're working on where we're focused in data science and analytics?

ARAVIND YARLAGADDA: Mark Kelly and team from the Center for Intelligent Power, they've developed a library of algorithms called Power Genome, Power Genome Algorithms. These Power Genome algorithms are nothing but data science learning models applied to specific domains.

Some specific examples that I can provide in the context of data centers-- battery life, battery health indicators. I think they have developed certain power genome algorithms that can indicate, and that will save labor hours in terms of management of the data center facility, and also labor hours translated into financial savings for the end user so they can run the data center better. Predictive maintenance of certain equipment in our own internal factories-- I mean, these are some examples that the Center for Intelligent Power has actually encapsulated domain knowledge within our data science algorithms.

And I think the point that I'd like to make here is, artificial intelligence and machine learning, heavily overused terms. All it boils down to is those knowledge graphs that we build, those learning models that we build, that we need to learn about the patterns in the data. And once we learn about the patterns in data, correlate that to specific domain problems, whether it's in aerospace, whether it is in vehicles, whether it is in electrical, or whether it's in industrial environments.

Correlate that to certain domain problems. Correlate those domain problems to specific equipment that actually feeds the domain content or deployed in that specific domain enviroment. That is a part of what Mark and team are actually doing with the Power Genome algorithm.

ZARI VENHAUS: And so when we think about how much data there really is, isn't there an aspect of this, too, that's, with all of this data, what is the data that's really important? How will companies start to think about parsing through the data and picking out the data points that are actually going to help them make better decisions?

ARAVIND YARLAGADDA: First of all, we as Eaton will need to understand these higher-level problem statements, from energy management, operations management, power management aspects. We need to understand the value of the problem statements they have and dig very deeper into what aspects of fidelity do you need to make your job better, to automate your processes?

And then dig a bit deeper into, OK, are we providing the right data that can map to those specific value propositions and those problems statements? And have that conversation with the customer, either through a design thinking approach or through continuous conversations with wise customers. So those specific data signatures map to those problem statements and value propositions and, ultimately, can help us, Eaton, understand how we process the data and what kind of data will need to be fed into those software suites, whether they're third-party software suites, whether they're customer-developed software suites, or our own Eaton software suites.

ZARI VENHAUS: So if we talk a little bit more about data and how we're using it internally, I want to talk about Industry 4.0, around how we're bringing automation and digitalization into our factories, into our manufacturing footprint. Can you talk a little bit about how we're using technology and data to help drive better decision-making in manufacturing?

ARAVIND YARLAGADDA: We've taken a very logical, grounded approach when it comes to factory digitalization with the deployment of Industry 4.0 technologies. So step 1 is really understanding, within the four walls of a factory enterprise, do we have the right processes deployed, and are we doing-- are we executing those processes right? Do we have waste in the current process that we need to take out, and do we need to lean out those processes?

Second, to automate these processes that we have shrunk or optimized, we need certain factory automation tools. So step number two is basically factory automation tools that we need to put in place.

Once we do that, step number 3 is understanding what model. It's getting deeper into the four walls of the factory, understanding what model lines do we have, what equipment do we have, and what data are we collecting. So we need to understand how do we automate these model line deployment, how do we create the specific digital code, if you will. And the definition of a digital code, basically, is a set of fundamental digital technologies that are deployed to automate certain factory processes. And then the next step would be, how do we harness this data, preferably through Brightlayer, to actually drive more insights for factory efficiency performance.

ZARI VENHAUS: And it's so important to start at step 1, because I think people hear a lot of this idea of, we'll just bring out another tool. We just bring in software. But a tool isn't going to solve your problem. You need to understand the process first, really know where those pain points are to choose the right tools.

ARAVIND YARLAGADDA: Absolutely, absolutely. You have to take it from a bottom-up approach, rather than a top-down. Most times, I think people consider-- people use digital and software interchangeably. Software is workflows, control, et cetera. And analytics is part of software, and data drives analytics.

And that's where digitalization comes into the picture. Digitalization is about the leverage of data for automation of processes. It's the leverage of data, the cleansing of data, to actually provide better insights to drive decisions.

ZARI VENHAUS: So let's talk a little bit about Brightlayer, our digital foundation. Can you talk to me a little bit about the value you think that Brightlayer can bring to our internal processes, but also to our customers?

ARAVIND YARLAGADDA: You've heard the term "digital thread," right? And how do you view the digital thread within the context of a factory, within the context of a work environment, the context of a customer? What it really means is, there are a lot of devices, IoT devices, assets, equipment, operators, et cetera, that sit on a factory floor, that sit in a power plant, that sit in a customer's environment. Powering those devices with smart content so they can communicate, so they can exchange information, data primarily. So that's step one of the digital thread, or notch one of the digital thread, if you will.

The second step is basically, how do you extract that specific data, the data that is of value, and apply certain algorithms or certain domain-specific algorithms to provide the right insights? That's step number two. And then how do those insights actually power software packages, whether they're Eaton software packages or third-party software packages or customer-built software packages? That's where the value point is.

The consumption point is the software or the analytics package. Whether it's through a mobile device, or whether it's through a desktop, or whether it's through a browser. That's the consumption point. And that's where Brightlayer, through our data insights, which are mostly asset-powered-- and we want to play to our strengths, Zari. Our strength within power management, our strength has always been about selling gear, about encapsulating or domain expertise within that gear. We're taking it one level above and encapsulating that in our data science algorithms.

And last, but not least, I mean, ultimately, you need to deliver an experience, a discovery experience for the end user or the customer in terms of what problem are we trying to solve, and then, do I have the right products that I can assemble as a solution that I can use in my environment? And do I need to collaborate with Eaton or Eaton's ecosystem of partners to develop certain applications using the Brightlayer platform, whether it's for cloud connectivity, whether it's for the edge connectivity, whether it's for developing a set of learning models based on the data that I have. So that's the experience hub.

So I'll repeat the four steps again. Basically, making our products more digital and smarter and connected, step 1. Step 2 is driving that data and insights that are of value. Third, feeding those insights into software packages, whether they're Eaton or third-party. And last, but not least, probably the most important, how do we deliver the right experience, the right digital customer experience? So that's weaving the digital thread. And Eaton adds value at every step of the way.

ZARI VENHAUS: This was a great conversation, Aravind. Thank you so much for joining us on this episode of "The Making of What Matters."

ARAVIND YARLAGADDA: Thank you, Zari, for having me. I really enjoyed the conversation. Look forward to more of these conversations and more to learn.