Podcast #183 Waking Sleeping Giants

TRANSCRIPT

Ken: Good day, and welcome to episode 183 of our Momenta Digital Thread podcast series. Today I'm greatly pleased to host Joe Perino, Principal Analyst, LNS Research. At LNS, Joe focuses primarily on industrial transformation and operational excellence for oil and gas, process manufacturing, and other asset-intensive industries. He also provides collaborative coverage across the industrial Internet of Things, Big Data, Cloud, advanced analytics, Edge computing, the digital twin, robotic process automation, blockchain, and asset performance management. Joe started his career as a process engineer in the refining, chemicals, and pipeline sector. He spent several years in Sales and Business Development with process automation and technology suppliers before moving into Product and Industry Marketing Strategy, Corporate Development, M&A, and management consulting. He has helped process manufacturing companies with strategy development, digital transformation initiatives, and market development throughout his career. Joe holds a Bachelor of Science in Chemical Engineering from the University of Notre Dame and a Master of Science and Finance from the University of Houston - Clear Lake. Joe, welcome to our Digital Thread podcast.

[00:01:53]

Joe: Thank you, Ken. Glad to be here.

[00:01:55]

Ken: I'm glad to have you. For those of you who didn't get a chance to hear what we had talked about before, which is pretty much all of you- we've had some technical difficulties. This is almost the third trial for this podcast. While we're going to talk digital transformation, unfortunately, the digital- at least of Microsoft Teams has not lived up to expectations for this call. You'll have to forgive us if we sound a little bit not directly on the podcast. On the flip side, we should be well-practiced in our answers. With that, you know, we call this the Digital Thread podcast. Naturally, we'd like to ask about YOUR digital thread. In other words, the one or more thematic threads that define your digital industry journey.

[00:02:35]

Joe: Ken, I say there are probably three threads. First, I have always spent my career in the process and energy industries. Being in Houston, that's easy. Second, I've always stayed in touch with technology and kept up with technology. Third, circumstances in my career dictated that I had to reinvent myself, learn to do new things and develop new skills, and be good at more than one thing, which I have done. For your listeners, I've had 15 jobs since university and have been laid off 9 times. Yet, I am still here and not ready to quit yet. You can only control what you can do. You can't control everything else that goes on around you. You must be willing to shift careers if required. Mine has zigzagged quite a bit.

[00:03:20]

Ken: It's interesting; when we think about the digital transformation of industry, you always think of this as an agile sprint, at least done in the best practice way. I think sometimes our careers are the same way. Many of us are taught to- we write our epitaph or write the Wall Street Journal article that describes what our future is and somehow work backward from that. I found that anytime I look back, I can see the thread that makes up where I'm now, and I can see those digital threads- that's how we came up with that name. Going forward, many times, it's a matter of trying these agile sprints and seeing what works and what doesn't until you find your place- which it sounds like you have at LNS.

[00:04:00]

Joe: Well, yes. Looking back, I can connect the dots; it makes sense. A lot of my experiences are built upon each other. I can certainly tell you that when I got out of school and started to work as a process engineer in Cleveland, Ohio, and then in Houston in the plant, I had no idea I'd end up where I am- in particular, becoming a consultant later and doing as much speaking and writing as I do now- and then looking back at it. But you're very true, and the story's not over yet. I am 67, and I'm not ready to quit; there's more to be written.

[00:04:30]

Ken: Well, we'll talk about what you've written because it was inspirational, certainly for us. If you're going to top that, there are many great things to look forward to. Looking through your bio is interesting. Of course, it was a mouthful introducing you right up the front, but if I look at your bio, it is a two-decade- who's who of oil and gas process leadership. I see you at I2 and Schlumberger, CGI, KBC, and even Well Aware, a San Antonio-based startup that we followed in the past. If I had to boil these two decades down into your top three insights working in these frontlines of process control, what would those be?

[00:05:06]

Joe: Well- and I don't think these are a secret to the listener, but the first is- I've been around long enough to witness the transition from a very hardware-oriented situation in the automation space to one that will become dominated by software. It's not that we don't need good hardware or that it isn't a profitable business. Still, success in non-control-related software is a real challenge for the automation companies as it is also for the hyperscalers.

The second insight is that some companies are chasing software without a real strategy. Lacking one will be very costly because of the sky-high software valuations that are out there. Several firms are rushing to acquire things, and they wonder what to do. In fact, from an LNS viewpoint, we see two types of companies; one that asks us, "Hey, Joe, we're about to do X, are thinking about doing X, should we do x? And if we do X, how do we position it?" and then there are the others that go out do X and then come and ask for help after they've done it, which is not the right time to do it.

The third insight is that we're still very early in the adoption phase of artificial intelligence and machine learning. I think it's still fairly overhyped in the market. Most end-users still really aren't ready for a significant rollout of this technology. It's probably not necessarily due to the technical issues, but much more about change management and readiness in the company. So those are my three.

[00:06:26]

Ken: Very good ones. And certainly, we've been to the centerpiece for 'Waking the Sleeping Giants' discussion today. Before I do, though, I'd be remiss not to talk about LNS. All your experiences culminated in joining LNS, industrial research, and advisory. I guess you joined them in 2018. What attracted you to the company and, perhaps generally, the advisory space?

[00:06:51]

Joe: Well, actually they found me on LinkedIn in July of 18. They had a contract recruiter working out of the Philippines that reached out to me. The reason for reaching out to me is that a very senior industry analyst, Dan Milkovich, who has spent many years with Gartner, and about five years with LNS, was going to pull the plug and retire in about a year. They were looking for an industry-experienced person to replace Dan. They found me, and they added me on the December team, and then Dan coached me along in the first year and as he phased down to retirement. I had not heard of LNS before because we're only about 20 people.

What attracted me was the chance to bring the culmination of all my past experiences to bear as an analyst, having been in the shoes of our two primary groups of clients. One is the software vendors, hardware vendors, the automation firms, and the software people like AVEVA and AspenTech. The other is the end-users we are increasingly adding to our list. I've also been a service provider, consultant, and end-user before, so I've been on every side of the equation. What makes LNS different is that our focus is on industrial transformation and more on the operational side of the business as opposed to the CIO. Although we do cover CIOs. The other thing that makes us different is that, unlike many of our supposed competitors, we are not the echo chamber for the marketing department of vendors. While we do write about vendors a lot, we provide a lot more insights and advice on the inside of what they should be doing, as opposed to simply echoing the message of the marketing departments. That's proven to be a little bit of a shift in strategy that Matt and Mehul did two or three years ago, and it's paid off for us. We think we're unique here. Sometimes, we find ourselves competing against a McKinsey or an Accenture more than we compete against an ARC or a Gartner.

[00:08:45]

Ken: For anybody who knows- the man, the myth, the legend, Dan Milkovich, those are both physically and figuratively large shoes you had to fill.

[00:08:56]

Joe: Well, I don't know that I filled them all. But I've replaced him anyway, and I hope Dan is having an enjoyable time in his retirement. He's still out there because he's still part of this other ex-analyst group that occasionally writes here and there. Still, he's certainly not doing anything full-time in the space anymore, as far as I can tell.

[00:09:13]

Ken: Yeah, he's a great guy. Certainly, a long track record. So all of this for our conversation- culminated in a recent article you published in the LNS Industrial Transformation blog titled "Waking Sleeping Giants… Releasing the Kraken." The title alone is inspirational enough or at least provocative enough. Our Momenta team found this quite insightful, given our digital industry investment focus. You chronicle the digital transformation journey of the top automation vendors: ABB, Emerson, Honeywell, Rockwell Automation, Schneider Electric, Siemens, and Yokogawa. What inspired you to write the article?

[00:09:49]

Joe: Well, I had originally developed the zone of transition diagram for one of our clients because we were discussing where the automation companies are headed and their challenges. It reminded me of the Gartner Hype Cycle curve and the change management curve. So, I said I wanted to write a blog about that; Matt said why don't you add on and talk about that relative to what the automation vendors are doing in this space because they're challenged. I took that on, wrote that, and I expanded it. Fortunately, when we can do blogs at LNS, we can use a catchier title- which is why you've seen the word 'Releasing the Kraken,' which, by the way- is the software Kraken. We wrote that, and I got quite a bit of view on that, not only on our own website, but I think about 1000 LinkedIn views of that article.

[00:10:43]

Ken: I can imagine. Mike Dolbec, our managing partner for Ventures- But Mike and I took this as part of our venture team meetings over an extra couple of weeks. We had each of the members of our venture team play effectively CEO for each of those companies you outlined. We had them go and look up the strategies, then we came back and looked at the corporate dev head's perspective of the same companies. Who are they likely to acquire? You want to be ahead of that curve for venture investors. We took this to heart because I think one, you picked good companies. You were quite insightful and probably not too complimentary. And I mean that positively. You took a critical view of the journey they've still got to play. And you said it earlier; this is still at the front end. And many people think, oh, digital transformation, IoT, industrial IoT- these are like hitting the trough of disillusionment. Some of these are still relatively early for many hype cycle perspectives. You pick some large companies that are slowly feeling their way into it, some better than others. Which I think you've chronicled well.

[00:11:50]

Joe: Each of these companies- is open to acquisitions, and they have some activity going on there. We'll note that next week it's Emerson and Aspen Tech's turn coming up next week. But all of them have taken a little bit different approach. One of the lessons that some of them still need to learn is that if you're going to sell enterprise software, you need to have an enterprise sales force; you're not going to do with automation people. Hence, I think Emerson realized that and is doing what they're doing. Honeywell- I worked for both Emerson and Honeywell in my career. Honeywell took a long time to create the connected enterprise team to tackle that. Others have a different approach, but they're all faced with increasing amounts of software. They're all faced with open process automation systems in progress as we speak, coming. They're also faced with the decoupling of software from hardware. I think we're at an inflection point in history, the big one being the start of DCS in the late '70s when Yokogawa and Honeywell came out with it. It's matured quite a bit since then. The systems are very good, and they're very reliable. But now, things are opening up, and the ISA '95 architecture is flattening a bit.

There's a new challenge, and this will be something that's going to get phased in. I think, over the decade and into the 2030s. It's a real challenge for these guys to tackle this. Let me give you an example. Many of these vendors have a huge installed base of technical debt, meaning installed users. Some of them have- for example, ABB has five control systems, one of which is the Taylor Mod 300, that Taylor built for Dow, that dates to 1984. Some of this is still out there, as is TDC 2000. So, what do you do? In a lot of these systems, you can migrate forward, but you can't migrate forward to the current capability. There's going to be a lot of 'rip and replace' that I think will occur in the next 15 years. These companies don't want to lose the install base. But there's going to be a time where they can't get parts. It's too expensive to support. End-users can't get the newer functionality they want. I think we've entered a new inflection point phase that will move faster than the development of DCS did from the late '70s to the '90s and the 2000s. Let me stop there.

[00:14:12]

Ken: You hit some great points. The one that's probably the kind of extra outside in is cybersecurity and the potential risk of these older systems, especially given what's happening in the Ukraine and Russia. I agree with you that there will be a lot of pressure to migrate these systems or modernize many of them. Let me ask about that. You introduced something called the zone of transition earlier. In my perspective, it's a digital maturity model, as you say, the Gartner hype cycle or maybe the 'crossing the chasm' change management cycle. Why do you think we need a specific model for the process automation space?

[00:14:49]

Joe: The zone of transition applies to all firms. The reason that I think the automation companies face this is that their business model was based on proprietary hardware and associated software that was embedded in it, which created a vendor lock-in for a long period of time, sort of an early version of ARR. Once you sell a system, you're going to get a certain number of services and spare parts and all that, which you can count on. The problem they face with digital disruption, software decoupling, and open systems is that this model will break down. It's not going to break down completely tomorrow. However, they're going to have to figure a way to still win in a more open, decoupled world and pull their base along with it, so that they don't lose market share and don't see their margins drop.

Each of these vendors is coming at this from a slightly different angle. Some of them have better financial numbers than the other guys do. They have a real challenge. Now, if we compare that, it's not all bad news. IT companies such as Cisco have maneuvered their way through having proprietary hardware, and now, they have interchangeable hardware like routers, and other things, with competitors. IT companies, and people in IT, can swap out hardware. Yet, they've still managed that, stayed profitable, and stayed relevant. That's the challenge they face- IT has done a better job of maneuvering through. I think the automation companies will have some bumpy times ahead while maneuvering through this.

[00:16:20]

Ken: As we listed those companies earlier on, who do you consider the leader in the pack? What have they done that bears repeating by the others?

[00:16:29]

Joe: Well, it depends on how you define the leader. Do you find it by market share? Profitability? Who do you think is technically innovating faster than other people? Those things vary by all these companies. Frankly, most vendor innovations don't provide a lasting competitive advantage, as functionalities are easily replicated one way or the other. For example, Emerson was first to come out with remote IO, and now everybody has it. How long did that take? Less than two years, maybe less than a year before other people did the same thing.

There are a few technical innovations I like that stand out to me. For example, I like Honeywell's ability to virtualize controllers in their Hive offering. They're selling a fair amount of that. One set of IO can talk to any controller, any controller can talk to any IO, and the controllers back each other up. They're not hard-wired backup; they're in a nest of controllers that can back each other up.

One thing I like to point out to vendors is that since most greenfield projects will utilize remote IO now, I always ask them why they keep designing rectangular-shaped controllers reminiscent of old single loop controllers when they're mounted in racks that sit in an air conditioning building? Instead of designing controllers that use blade servers and make that look like a data center? They're still stuck with that. They can't get away from the old way of building things.

If they were more innovative, they could do that and take advantage of what IT has taken advantage of and still build something reliable, fast, and meets the needs of the process control world. There are some challenges for these people. I don't see the speed of innovation that I'd like to see in other places in software and, frankly, on the IT side.

[00:18:13]

Ken: Speaking of IT, there is certainly another group of giants that didn't get mentioned in this article- the so-called hyperscalers. AWS, Google, Microsoft. What impact do you think they will have on the traditional IoT giants?

[00:18:27]

Joe: Well, they're already impacting the giants. A lot of the level three software like advanced planning and scheduling, MES and MOM, the engineering software, CAD, PLM simulation, even control system testing, plus all the analytics- that's all shifting to the Cloud and some to the Edge as well.

I don't see the hyperscalers in the plant at the control system level necessarily saying level two, where the controllers are. They are there at the Edge. What keeps them from playing there is probably better places to go, the lack of domain knowledge that they must leverage and use the know-how to apply their products. All automation vendors and others will depend on one of the three or more than one of these hyperscalers.

Their business model is simple. We want you on our platform. All roads lead to Rome; that's what they want. They're going to do whatever they can to attract applications to their platforms. They're not trying to tell the control system vendors- move your control system to our platform. But the development tools, the testing tools, and all the other level three software is moving to their platforms. The game, of course, is to develop such a strong position that you dominate in terms of volume and cost advantage. That's the game that they're playing.

But on the other hand, over time, these guys will become commodities in what they do. The services already are because there's not much difference between Amazon and Azure in terms of components that sit out there. There is a difference in how they go to market, how they partner, what kind of deals they get, and how they incentivize their partners. Those are differences that are real. That is where a lot of the battle is.

[00:20:06]

Ken: Every industry has its digital disruptors; they call it the Elon factor. Like Tesla for automobiles or SpaceX for space travel. Do you foresee seeing such a disruptor for industrial and process automation? And if so, could you guess in terms of timeframe?

[00:20:23]

Joe: Well, we see that disruption now, I mentioned, the decoupling of software from hardware and open systems. I think that's going to take, at least to the end of the decade, those two things to mature enough to where people are going to say, yeah, we're going to just put in an O-PAS compliant system. It is going to take a while. We don't know which vendors will be leading in that area; probably, Schneider and Yokogawa will move sooner than others. That's my guess, based on their activity and what they're doing. None of them have announced that they're going to go there.

The other area that's probably ripe for disruption is legacy MES, Manufacturing Execution Systems. There's a lot of them out there; they're rigid, they don't scale. That is being disrupted now by startups in the industry that are componentizing or making it composable to have your manufacturing execution system; however, you want to define it. That way, a small plant can start with five or six functions, and they don't need to buy the whole enchilada, so to speak, with the cost and the two-year implementation program. That's an area that's going to be disrupted.

The third area that you'll see disruption is that we already see correlative artificial intelligence. They use the word correlation. What we haven't seen yet- which is coming, is causal AI. Causal AI reasons and makes choices that humans do. It is mostly at the university level now, but I know of at least one end-user using that technology in their customer-supplier management end of the business. It's supplementing and fixing a lot of things that the ERP system won't do right because the ERP systems of record are still rigid, transactional step after step workflow systems. When you run into a problem, humans must take over and fix the issue.

The idea with causalAI is that the machine will tell the human what to do and fix the problem so that the transaction can continue moving forward. You will see more of that coming out of academia/university and into the systems theory as we move forward in the next five to 10 years.

[00:22:28]

Ken: Well, that's what I'll have to note. We're avid investors in the space, so beyond causal AI, what other trends in startups are you watching these days?

[00:22:37]

Joe: I'm watching three trends that I particularly like to write about. One is Operational Excellence 4.0, and it's tied to industrial transformation. The second one I just mentioned- autonomous operations, which I've written about- and the role of correlative and causal AI. There's a system of system challenge to get to autonomy. Autonomy is a lot more than just having your control system run itself. The third one is what we're watching is this future work issue. By that, I mean the issue around attracting and keeping talent- the great resignation. People don't come to work. They don't stay at the companies anymore. What do when your most mature worker is only four years on the job, and you don't have the 30-year people anymore? That's the biggest problem that manufacturers face; it's this people issue and the skills issue that goes along with it. These are the three big trends.

[00:23:32]

Ken: The third one is very interesting because we do exec search work. We see that firsthand.

[00:23:39]

Joe: Now, if we go to the startup space, we track four areas. I'll mention some names because some of these people are our clients. Of course, we're watching analytics vendors, and I'm doing an analytics solution selection matrix guide right now. That includes everybody from C3 to a small firm like Quartic. We just watched Seebos, for example, be acquired by Augury.

The second thing we're keeping an eye on related to future work is connected frontline workforce software. There are about 20 vendors in that space. They are working on software that facilitates factory workers' safety and effectiveness. I'm not talking about operator rounds; I'm talking about use cases in asset management, EHS, and safety in all these things; they are all facilitated by one piece of software that handles all the collaboration around workers working together and working as individuals. This is a very hot area right now. And there are a lot of players out there and a lot of whitespaces to be gained. Two companies we're watching very closely are Augmentir and Redzone.

The third area is these vendors that I'll call manufacturing performance systems. They're using data models of the production system to inform and optimize. They're doing something we might call wrapping and extending around existing MES or even replacing an MES in at least a small space. There's sort of the Namur approach to wrapping and extending. Whether you're complying with Namur and using OPC UA or not. These are firms like BrainCube, SightMachine, Tulip, TwinThread, ThinkIQ, Oden, Machine Metrics, and people in that space.

Finally, a fourth one is none of this happens without data access, data management, and data democracy- this is a big deal. How do you architect the data at the plant and then the corporation and be able to scale it? This takes us into the data ops space. I've written about data hubs. IT calls all this data fabric. The players there are mostly startups: Cognite, Element, Uptake Fusion, and Kepware- part of PTC, HighByte, Inmation from Germany are in that space. Even a new startup spun out of Trendminer called TimeSeer, which focuses on data quality. This is all about bringing the three types of data: structured, unstructured, and time-series together and putting it in a way such that applications, whether it's a BI tool or a digital twin, can access that data in a standard way and consume it. This is something that all the end-user firms are challenged to architect right now and figure out how to do. Do they DevOp it from scratch? Do they buy components? Do they buy most of it rolled together in one solution? How do they do this?

Then an offshoot of this is this contextualization space in which you decide whether you're going to do it the traditional way, which is a hierarchical canonical mapping of data together, or do you use object technology? Or do you use what's pretty hot now- it's been around a while, like 15 years, and that's graph technology. That's a hot space. There you have vendors like Neo4j, Cambridge, Redis, and others to keep an eye on. Most of those, again, are startups; they have VC money behind them. I think you're going to see a lot of activity there. Graph technology is just beginning to enter the industrial data management space. I'll stop right there.

[00:27:03]

Ken: Wow. I wish you wouldn't stop. But of course, we're coming up on the end of our time here. Thank you for the plugs, by the way, for several of our portfolio companies, including HighByte and Sight Machine. In closing, I guess what are you reading or watching these days that inspire you?

[00:27:19]

Joe: I am watching and reading many things, and I follow LinkedIn quite closely. LNS is taking advantage of that fact. Matt has a blog almost every morning with a snippet of insights or research. Some of the other salespeople working for us and other analysts do the same thing. Occasionally, I'll pipe in.

From a personal side, I like to watch sports, and I also like to read about history- US history and world history. I try to keep up with a bit of science on the climate change side because we're following that very closely now because of the sustainability and ESG movements that are out there now that we're tying in. The business team's challenge is that many companies have had to make commitments to NetZero, but they haven't figured out how to get there. It's one thing to say we'll be NetZero as an industrial firm; it's another thing to figure out how to get there. We're following that closely too.

[00:28:26]

Ken: Very good. You answered the question the way I had intended to, so thanks for that. Joe, thank you for sharing this time and these great insights with us today.

[00:28:34]

Joe: Well, Ken, thank you again. I appreciate the time with you and your listeners. And you can reach me again, of course, on LinkedIn or through LNS, and I'm happy to answer any questions your listeners might have as a follow-up.

[00:28:44]

Ken: Perfect. This has been Joe Perino, Principal Analyst at LNS Research. Thank you for listening.

[The End]