Rockwell Automation, Inc. (ROK) Earnings Call Transcript & Summary

Hello everyone, and thanks for joining today's webinar. Before we get started, we have a few housekeeping items. The audio for this event is streaming through your speakers, so make sure your volume is turned up and speakers are turned on. Our webinar platform performs best on Chrome and Firefox browsers. On the lower left-hand side of the presentation, you will see a Q&A box. We encourage you to enter any questions you have there, and we will answer them at the end of the presentation. If you're having trouble connecting to the webinar, please take a moment to refresh your browser and disconnect from your VPN. If you're still having trouble, please clear your cache. We have instructions in the handout section of the webinar platform. All of the panels in the webinar platform are adjustable, so you can resize them simply by clicking on the corner to adjust or hit the maximized screen at the top right-hand corner of each panel. This will effect on your experience and not anyone else. Today's event will be recorded and will be available right after it's completed. You can access the recording utilizing the same link that you used to access live event. After the webinar, we will also be sending you an e-mail with the resources from today's event, including the slides, handouts and event recording. Additional information regarding today's topic can be found in the handouts panel of the webinar platform. With that, I would like to present our today's speakers.

All right. Well, thank you very much. I appreciate the introduction. My name is Oliver Haya, and I'm joined by Kevin Olikara. We are here to talk to you today about, how to use industrial data and how to put it to work using a new offering from Rockwell Automation called FactoryTalk DataMosaix. So without any further ado, let's dive right into it. Fundamentally, you need data to solve your challenging problems. And this is very much a people-focused problem. Different people have different needs for data. We look at a lot of different personas, lot of different users of industrial data, and you all have really different problems. Reliability engineers may have an interest in trying to find the root cause of equipment failure, how to predict unplanned downtime, how to prevent it or how to fix it faster if it happens again. Process and manufacturing engineers may be focused more on yield optimization or improving quality, looking at some of the other process variables or environmental factors that may be contributing to issues in production. Data scientists have potentially the most advanced skills with how to utilize the data and how to build predictive models or machine learning models, really how to build the best quality, how to build the most cost favorable product. And across all 3 of these, you have different skill sets, you have different capabilities and different problems that you're trying to solve. Each one of these personas is going to utilize different data, they're going to utilize data differently, and they're going to use different tools in order to access the data, to create applications, to do exploratory analysis and to try to improve their production processes. We want to start off with a little pull to understand who you are and what you are working on, what problems you are trying to solve and so first we are just going to ask you, what tough problems that you are trying to solve and this might be an select all, so you might be able to select and click all that apply if you cannot just pick the one that is best fit for you, but we want to know what tough problems are you trying to solve with your industrial data. Is this root cause failure analysis? is this process control optimization? Are you looking at how to improve your energy utilization, reduce your energy consumption or be more sustainable otherwise and have used water, air, gas, electricity or steam? Are you trying to analyze production losses to improve your yields, or looking at how to predict quality better? There is plenty of other use cases and options too. Kevin, do you have any of these that you've favored over the years that you like to focus on with data?

Yes. One that I'm particularly excited about is process control optimization or yield optimization. We hear all the time about manufacturers being capacity constrained. And the #1 thing taking away from capacity is scrap. So being able to minimize that scrap is going to pay incredible dividends in terms of maximizing productivity and plant capacity. So I would say yield improvement and scrap reduction has been something that I've been particularly excited about.

All right. Let's see how that lines up with the group. And that does seem pretty well in line. It looks like it was a select only one. So thank you for everybody who participated in that. I appreciate you voicing, I think that, that lines up really well with what we're seeing across the industry, the process control optimization, the root cause failure analysis. All right. Thanks, Kevin. We'll check back in with you in a few slides when it's your turn to talk as well. The next set of discussion that I'm going to kick off here is a quick recap of some of the information we talked about in a previous webinar around an introduction to industrial DataOps. Why do you need data? How do you use it? And then we're going to pivot into more of a discussion around what to do with your data and how FactoryTalk DataMosaix can help you analyze and utilize that data across your enterprise. So let's dig in right now into maybe some of the barriers that you've seen with analyzing data, getting value out of it and being able to use it across the industrial operations you have. I'm actually going to start off with the quote at the bottom here, which is from our smart manufacturing report, we did this -- this was the eighth one that we did. And 40% of manufacturers said that they lack the ability to use data to make decisions to outpace the competition -- I'm sorry, 40% more than last year. It's an increasing number every single year. And the challenge that we've found talking with companies across the spectrum here is that a lot of companies get stuck into point solutions, for example. The reason they get stuck in the point solutions is because it's hard to get data out of silos and to contextualize data. What ends up happening here is that you can solve 1 very narrow problem, you can solve 1 yield optimization in 1 part of a process in 1 line, in 1 factory or you can help improve the quality in 1 place or you can predict downtime on 1 type of machine. But it doesn't scale beyond that particular asset, doesn't scale beyond that type of asset, it doesn't scale across the plant. These point solutions are really good from a proof-of-concept perspective, but because the data isn't readily accessible, it isn't contextualized across the different use cases of the data across the IT data, the OT data and because the data isn't necessarily trusted, it may have poor quality, it may not be delivered in a timely fashion. You don't have an owner of that data. All of those things lead to a difficulty in scaling and scaling applications is really where the value comes from. A proof-of-concept will look good on your resume, but it doesn't measure up anywhere near to the value of scaling a successful proof-of-concept or proof-of-value across your entire operation. And so what we've seen is that a lot of proof-of-concepts get into what we call pilot purgatory. They show the value in 1 place, but it takes so long and take so much effort and people and technology in order to be able to scale it that the ROI gets diminished dramatically. We're going to take the next few slides to dig into why that is and what are some ways to get around that. What we see often is that people tend to focus at the visible questions here. When they're looking at different solutions, what am I trying to do, what data is needed and how do I use it to solve the problem? And those can be really valuable in that proof-of-value stage or in that pilot stage. But all those details at the bottom half of the iceberg underwater there is really critical to being able to scale this across. How do you collect that data? How do you store the data, who has access to it? How do you secure it and govern it to make sure that the data isn't being modified, that it is being quality checked? What extra context does the data need? There was a big fashion for a while of just put a sensor on a motor and send the data to the cloud and we'll tell you everything about your machine. And the reality is that there's a lot more information that needs to come out of the equipment in order to make it meaningful. You need more information about how that machine was designed. You need more information about what batch was being run or what process or recipe. You need more information than just a single sensor. That's why you have to pull all the data together. And that data when it is contextualized will get used in lots of different places. Good data is going to be referenced across multiple applications, and there's a lot more questions to ask here, too. But these tough challenges in the bottom really need to be solved to get out of pilot purgatory and to scale. So one of the things that we recommend is an industrial data hub. This is a way to operate in a more agile fashion to be able to treat data as something that needs governance and needs an owner around it. We view an industrial data hub as having 4 key facets: one is facilitating industrial DataOps. This is going to make it easier to integrate OT, IT and ET data together. And if you're unfamiliar with ET, that's Engineering Technology. That's the P&IDs, that's your electrical drawings, those are manufacturing specifications. Bringing all of this data together with time series data with alarms and events with the engineering data with the IT batch records or the orders that are coming out of your ERP and your MES. All of this data needs to be integrated and contextualized together and you need to treat this data like it is part of the product. And you can make new applications that can be reused or they can reuse data in multiple facets. These new industrial data products are going to be things like predictive maintenance, predictive quality. They might be decision support. So things like diagnostics or other forms of analytics that can help you trace back what happened, why did it happen? And when will it happen again? We find that it's really critical to decentralize the data ownership out of maybe an IT data lake and really start to provide both federated decentralized access as well as ownership by the domain experts. These last 2 bullets on the right here are really critical to giving the right people, the right access to the right data at the right time. And that's because, quite frankly, the folks that have data science backgrounds on the IT side may not know what the data is. You've got to bring the OT knowledge in, you've got to bring the domain knowledge in. And this is why standard DataOps that has been growing and growing in the IT space for a long time, probably isn't enough for an industrial environment. Industrial DataOps is really critical to pulling that industrial data out. IT DataOps is fantastic and it's really critical for IT operations, but industrial DataOps is going to help bridge OT data in and start to connect to IT data. It acts as that industrial data lake potentially, but creating those pipelines in order to standardize connectivity between your OT data sources, your data lakes, your operational applications is really critical. And industrial data is, quite frankly, different than IT data. Time series data has unique facets to it. Where you need to be able to measure the quality of that data more frequently. You need to be able to understand when there's gaps in data and you need to be able to store it in an efficient way when the data doesn't change that frequently. Other aspects of industrial data are the size of it. There is just a lot more industrial data potentially. When we start to pull in large amounts of time series data from the factory floor, that's huge. A typical servo drive can create a terabyte of data every year. And if you've got a plant full of thousands of servos and thousands of variable frequency drives and other PLCs and sensors and actuators, all of that data is going to become very big. So you need to figure out how to manage that correctly, how to create proper hierarchies and organizations of that data. And organizing that data isn't just digital process. It should include more links to the physical world because industrial data ties back to the physical world. We can start to connect this to CAD models, to imagery, to your P&ID to other electrical equipment diagrams. Being able to link the digital and the physical is a key facet of industrial DataOps that IT DataOps that doesn't typically cover. And because we're focused just on the OT space, we can make it easier for those domain experts to self-service on the OT side. This allows the subject matter experts and the domain experts, all those people we talked about before that have that process expertise to start to build applications that are helpful for them, that allows them to start to create new value out of the data. And this is also going to lend itself towards engineering analysis. At the end of the day, a lot of the people that are going to be looking at this OT data aren't business analysts. They are going to have a part in that, but it's going to be a lot of engineering analysts as well. It's going to be figuring out how to do that complex root cause analysis, being able to look at many different facets of information yet. And so industrial DataOps is a critical factor that layers on top of the standard data ops practices that more and more companies are adopting. And when you do lean into industrial DataOps that can really help unlock transformational value. This can reduce maintenance costs dramatically. It can improve worker productivity and it can reduce downtime on your assets. And when we talk about reducing downtime on assets, that is a huge boon to your financials, to your bottom line because those asset downtimes, that's lost production. Maintenance costs, that is lost money. Worker productivity allows you to do more with fewer people, which we're all struggling with hiring enough people to operate the plants and assets that we have. And so leaning into industrial DataOps is really critical to get more information out of your OT assets to make them more efficient. And we've been building a new data platform called FactoryTalk DataMosaix to help do this. FactoryTalk DataMosaix is a Software-as-a-Service industrial data hub that combines industrial DataOps with data governance, access for the right people, connecting data from the ET and OT side on the left with the IT side and the applications on the right. It is meant to be the centerpiece of a connected enterprise production system. And you can pull that data from third-party sources, and we've optimized it for pulling data from Rockwell sources. You can send that to all different kinds of applications. Kevin will talk in a lot more detail about where that information can go and which applications can use that. And it's really a beautifully open ecosystem. I want to talk a little bit about why we chose to do this as Software-as-a-Service first. A few years ago, we introduced FactoryTalk Hub, and one part of that is FactoryTalk Operations Hub. This is a large and growing Software-as-a-Service portfolio. And we've linked it into Software-as-a-Service as a company because the benefits to you can be huge. There is a dramatic reduction in the total cost of ownership when using Software-as-a-Service because so much of the infrastructure and people power that needs to go into building a service like this falls into Rockwell. And that allows us to help you get access to this more quickly. It allows us to help you grow more quickly as well. Because as you start to change what you're working on, as you start to grow out of your solutions or grow out of the space you were in, if you had a custom-built solution in your own cloud, that would be a lot harder to modify and change. This allows you to get to value much faster because you don't need to go hire more and more OT cloud data analysts or cloud ops people. You can use the software with the people that you have and that will help you start to get more predictable costs more quickly. Costs are really interesting to put on the cloud because a lot of people have seen their cloud costs grow and grow and grow year-over-year. And with Software-as-a-Service, you are going to have more predictable cost structures. We're also going to be responsible for optimizing your system performance, making sure that you have the right access to data at the right time and that's our specialty. It might not be yours. And we're going to do all of this in a way that maintains your ownership of the data, your access to the data, and you can take that data with you if you ever choose to stop using the Software-as-a-Service platform. So let's talk a little bit now about what happens when you have this data. First, you need to get the data. Your good news is your plants started rebuilding. Now you need to connect to that data. So how do we connect to all the IT, OT and ET data into a common data hub and make it available. Then we need to contextualize that data with you. The contextualization process is one of the most time-consuming and painful parts of making data useful before it can be useful and you need to add meaning to that data. And so AI-supported data contextualization tools can accelerate that dramatically. To make the data useful. You need to make sure the right people have access to it in ways that they can discover the data, they can use the data and then you can start to assign value to it. And that value comes from the applications, the analytics, the visualizations that come from that data. And that value is really driven by the applications that you built and there's a lot of different kinds of applications you can build. Across the bottom, we have what we call foundational applications. This is monitoring things like OEE, looking at your historians, being able to monitor your assets. These foundational applications are really the starting point to build many other more advanced applications. These foundational applications tend to be more descriptive, sometimes diagnostic, but they describe what happened. As you start to try to look at why something happened, you're going to be able to build more and more advanced analytics, things like yield loss, energy usage, asset performance. And as your analytics become more powerful because you get more data and because you get more people with the right access to the data, you can start to build towards predictive models. Predictive models are a really good way for you to start to test how accurate your data is, how accurate your models are. These predictive and prescriptive models are not only going to tell you what's going to happen but then you can start to build in what should happen after that. And these are going to be yours. These are the things that you can work on with your system integrators. You can work on it yourself. These are the ways to improve your operations. And eventually, as you start to get more and more confident in the data and the predictions and the prescriptions that it makes, this is how you start to build towards an autonomous enterprise. This was industrial DataOps in analytics in a very brief spell. I'm going to hand it over to Kevin here in a moment to start walking through the framework that we're going to use for the rest of the presentation to talk about DataMosaix and how it helps you make data available, meaningful, useful and valuable. Kevin?

Yes. Thank you, Oliver. So today, I'll be walking you through these 4 main pieces of our industrial DataOps framework making data available, making that data meaningful, making it useful and finally, valuable, starting with what does it take to make data available? We hear all the time for manufacturers that their data is locked away in data silos. To get the data that they need to solve problems, people are forced to manually export data from multiple source systems where that data resides. Not only is this extremely time intensive, but it's also error-prone and there's no traceability to where the data comes from. And once the data is extracted, it's immediately out-of-date. That's why the first step to industrial DataOps is to utilize connectors to automate data extraction from the relevant OT, IT and ET systems. This provides common access to data for people across the organization in different functions. To help accelerate time to value, FactoryTalk DataMosaix provides a library of standard data extractors to connect to a wide variety of data sources. This includes several options for connecting to your OT data sources, including Edge Gateways, Historians and OPC UA. Over time, Rockwell Automation will be releasing additional data extractors for premier integration with Rockwell Automation hardware and software products. Sometimes, more customized data extractors are required, such as when connecting to more complex enterprise systems like an ERP. In these cases, templates are available to help accelerate that development. Engineering data like 3D models and technical documentation that describe the physical real-world nature of industrial environments can be broadened as well. As organizations start to scale across multiple locations, an Edge management and orchestration solution like FactoryTalk Edge Manager, can be used to deploy and manage the extractors needed at each location. FactoryTalk Edge Gateway can be used for a simplified access to OT devices and contextualization of the Edge and is included at no additional cost with FactoryTalk Edge Manager. Data governance starts as soon as the data enters the system with centralized administration. This includes setting the access controls and policies to make sure people are able to access just the data that they need, whether that's by function, site or something else. Another key aspect of data governance that DataMosaix provides is traceability, with the clear lineage of where and how the data is brought up. So just as a quick recap, the main objectives of making data available with an industrial DataOps solution is to simplify the data connectivity across a wide variety of industrial data sources. And then enabling the infrastructure, pipelines and governance to do it at scale. So that brings us to our next polling question. Where is the data that you need today reside? And Oliver will be walking you through the options.

So we've got a few different options here. Feel free to select again, if you can -- select more than one, please do. If not, we're looking at Historians or SQL databases. We think that's a really common one control systems or other SCADA systems, is this in your MES or ERP? Is it in your CMMS or your enterprise asset management? I can't tell you how much important information lives on somebody's computer or a thumb drive that gets passed around or just a network drive or other. This is -- we're really looking for what's the most important one for you because I think most of you are probably going to be looking at using data from a lot of these sources. And we want to make sure that we've got the right connectivity and the right priority on how we do that. Kevin, what would be some of the examples of other that you can think of that people might want to pull data from?

Yes. Some other examples include a laboratory information management system. So if you're doing testing of products or quality control testing, I think that's a critical piece of information when doing, say, predictive quality. Sometimes, we're pulling from alarms and events from an HMI. So that's another critical piece of understanding what's going on in the industrial control system. So a lot of places where important data can be gathered from, for sure.

And it does look like a lot of information is living in Historian or control systems or similar databases. And I think that makes a lot of sense. That's where a lot of context is built today. That's where you go from having a sensor and an actuator to having part of a control loop. So that's valuable.

Yes. This is definitely aligning with, I think, what we would expect with a lot of data and Historians today, trying to extract data from various control systems. So definitely in line with what we're seeing today.

All right. Great.