Podcast #142 Industry 4.0 Pioneer

Good day, and welcome to episode 142 of our Digital Thread Podcast Series. Today I’m pleased to host Vatsal Shah, the founder and CEO of Litmus the leading unified Edge-to-Cloud platform for the Industrial Internet of Things. Momenta’s a proud and early investor in Litmus.

Vatsal leads the management and engineering team as co-founder and Chief Executive Officer at Litmus. He has extensive experience with industrial engineering, electronic system design, enterprise platforms and IT ecosystems. Vatsal earned his master’s degree in global entrepreneurship from EM Lyon, and his bachelor’s degree in electronics engineering from NIRMA University, India.

Vatsal, welcome to our digital thread podcast.

1:19
Thank you for having me Ken, it’s exciting to be here.

1:23
I feel like I always say this in all these podcasts, but it’s long overdue! You were one of our earliest investments, and I know we’ve had you on some webinars that we’ve done, and others, but I don’t think I’ve ever had the opportunity to truly just interview you directly, so I’m glad we were able to find the time for that. And you certainly have a fascinating background, so this should make for an interesting discussion.
So, I always like to start this by asking the question, what would you consider to be your digital thread? In other words, the one or more thematic threads that define your digital industry leadership.

1:59
Litmus has always been founded on meeting real customer needs. I think that is our key value proposition and our leadership in the industry. Big companies see an opening in the market and they try to fill it. More often than not they will just extend their existing product, or product portfolios to fill that gap. Small and nimble companies like us, we have to find specific needs with the customers, understand from them, and develop a product, develop solutions to fill the need.

We knew early on that manufacturers needed a way to digitally transform without rebuilding their entire shop floor, and we focused our business on allowing them to do that in the easiest way possible. That meant starting with building hundreds of different drivers to capture operational technology data, followed by adding analytics, followed by adding various different machine learning components, integrating them to Cloud and enterprise ecosystem that is so prevalent in the market nowadays.

Our customers need access to their operational data, and they need to get value out of it to improve the way that they are doing business, and we enable them to do that in a simple and secure way.

3:16
Now you started this digital thread as industrial as an industrial engineer at Arvind Mills and Rockwell Automation, which obviously gave you a very strong platform to, as you say, solving real customer needs. What attracted you to the industrial space, and what were some of your early learnings in it?

3:36

So, long ago I started my career as an industrial engineer, my job was to design those PLC Ladder Logic, design SCADA utilization, and as a network, engineer bring many different types of automation vendors together on a single application. You might still remember M2M world, and all the different vendors were trying to create a joint solution which fulfilled customers need. One of the very large projects that I worked on was one of the oil and gas initiatives, we had to integrate Amazon, Yokogawa, Allen-Bradley, systems together and it took three or four – close to three months to design connections between all of them. Yes, we were using fancy OPC’s, and we were using many different native drivers. I still remember we were using that visual basic environment which those operational technology vendors used to provide. That fundamental data communication challenge was well-identified early on.

In the 2013-14 timeframe the industrial world started talking about data and data-driven decision-making. Yeah, extension to those MtoM solutions, but one thing was these OT systems would not change overnight, and broader cloud and big data system wanted that data in a specific format in a specific way, with enforced security on top of it. We knew that there was a huge gap between operational technology, and the IT core system, this is when we founded the company. Our first slogan was ‘Things, Business, Connected,’ we wanted to go to that middle layer that links the OT and IT world.

5:29
It’s a very interesting time in there too, and the transition between MtoM and IoT, which I remember very well. I always pitied all those poor companies that put MtoM in their name somehow! In the same way I kind of pity those who have IoT in their names now!
You co-founded Litmus, at the time it was Litmus Automation, though as you say in 2013 your co-founders, of course, John Younes who is your COO now, and Sacha Sawaya who is your CFO. Originally what problem were you setting out to solve, for whom, and how did you evolve that over time?

6:10
In 2013 we started pretty generically. If you remember, in our early investment which is to Momenta, we were solving a broad IoT loss, Industrial IoT challenges. We were connecting any different types of operational technology and any different type of IT ecosystem on the other side. In our solutions portfolio, we had smart manufacturing, smart transportation, fleet management and connected cars. Again, way too many things for a start-up often.

So we went to customers, we got in front of the customers as soon as we developed the product, and consistent feedback was, ‘Go back to your roots,’ and the early products that we designed were really focused on smart manufacturing. The early version of the platform, this is 2014-15 type of timeframe; it was just a cloud platform, and we were giving away our edge product for free. We used to call them Gateway Agents, so current Litmus Edge was a free product called Gateway Agents at that time. Those Gateway Agents had PLC drivers, all the different drivers that we developed, it had a small analytics engine and integration to our main cloud product which was Loop Cloud.

We got in front of a lot of large manufacturing customers and asked them if we could help them bring their manufacturing data to the cloud, and they told us, ‘Your free Gateway Agent sounds promising, can we use it, and not use your cloud product, can we just use your free product? We got that question more often than you think, and that’s when we did the first pivot.

We converted cloud for a strategy to edge first product, and that’s the starting point of the Litmus that you see right now. And on top of it, the challenge that we understood early on is everybody was focused on machine learning and artificial intelligence, and vision processing, every start-up was getting funded around that, the specific spectrum. But those would not be successful if the data is not claimed from the operational technology side, we are dealing with 1960s data, we are dealing with a 1970s system, 30 years back the internet was not even there on those operational technologies. So, if we can solve that mess on the operational technology side, we can build upon that portfolio, we can build upon that foundational layer and add more integration, add more analytics as we go, and that’s exactly the plan that we followed.

By the year 2018 we had the highest amount of protocol drivers in the world, we have this intellectual property which allows us to launch new drivers every week. And again we went back to customers and they kept on asking, ‘You’ve solved our data challenge, now allow us to analyze that data, allow us to do this machine learning, allow us to bring data into our ecosystem,’ and that’s exactly how we built our product. This is, we evolved over time, we started with the foundational data layer, we kept on adding edge computing, edge processing, and integration components on the top of it. That’s what we define as an industrial edge computing system.

9:42
One might say your clients actually pulled you to the edge, interestingly enough in that regard, because obviously OT - Operational Technology has always been very edge-specific, and so it’s interesting that now you’re able to pull in your cloud offering, your full-stack offering behind that as well. So an interesting pivot given how many platforms started off, either on-prem or cloud to begin with as well, and didn’t quite make that jump that you guys did clearly to the edge.
So, you guys have been clearly positioned at this Industrial Internet of Things space, really since your founding; and I do remember some early conversations at, I think it was Xerox Park if I remember, in Palo Alto. I know you use the term industry 40 quite a bit to describe the solution space in your own set, what does this term mean to you, and how does this differ from past efforts at industrial automation, or even what you might consider to be a human-machine interface or HMI for the industry?

10:48
Yes, a perfect segway. The thing that we have to understand from the legacy world, or operational technology world, or just modern automation system, there are a few key things that you can spot very easily. The first one, they are tight vertically integrated solution, the company that designed PLCs, they designed their SCADA system, they will design their HMI, they will design their historian system, they will provide applications and solutions on top of it. They limit the capabilities of their next generation of products on their existing product portfolio. Why would they support their competitor's PLC system? The multi-vendor solution was and is not their first priority.

The second thing that you can spot is there are lots of isolated point solutions. You’re going to find legacy systems were designed to fulfill a specific need which is right now. They were never designed to extend as you explore, they were never designed to innovate on top of it. So, this is the reality, this is the word that we see on the shop floor. So when we entered we had to start with a non-disruptive approach; everything that is already there on the plant floor needs to stay as it is.

What we started is we want to liberate the data out of the operational technology system, and then on top of it we created the data-driven journey for the users, so we allow them to normalize the data, so all the data from Rockwell or Siemens looks, all the same, we allow them to push data to Azure, AWS, Google at the same time, we allow them to enable those data lake services, so for us, the focus is data-driven intelligence.

For legacy systems, the focus is to solve a challenge that they have right now. This is why we are winning against them, we are winning against data systems, legacy, OPC server systems, just because customers are focused on data, data security, everything in terms of how they can utilise that data, at the edge, at the cloud. Those will be the key difference.

13:09
It’s interesting, you used the term legacy, and I was in a plant not that long ago and used the term to refer to an old PLC system, and the gentleman turns around and said, ‘No, this is our heritage system.’ And it does bring up a good point, there’s always this thought more on the IT side of innovation that says just rip and replace, right? And you can do that certainly with PCs and other things on the IT side of the business; OT, these are very long life cycles for these things, and once you’ve validated and are regulating any kind of manufacturing process, very-very difficult to swap those. One of our earliest thesis where you guys fit in really well on the investment side was this idea of overlay systems, as we joke, brown is the new green; brownfield is where the reality is in OT. So the ability to translate that abstract and ‘bring it' into the IT world, especially with the hyperscalers as you mentioned, is really-really critical in that regard.

I like your idea of data-driven intelligence, great term there.

14:28
Yes, and I think that’s one of the things that is consistently missed in the industry, which is everybody wants to focus on the end result, but the journey is far more complex because operational technology doesn’t want to contribute to data-driven intelligence that easily. So, there needs to be a common understanding for the convergence to happen.

14:56
What have been some of your notable use cases and wins at Litmus?
 
15:03
In terms of use case, we had one very interesting story happen last year. We are a product-led company, so our smallest customer and our largest customer they’re using the exact same version of the product. We are amazed when one of our customers push boundaries and realises as many use cases as they can on top of the foundational data layer. One of those customers was Niagara Bottling, exceptional leadership on their side, and the team – our team is executing it very well – the Niagara Bottling team. What they did was they started with the base data foundation layer, that means went to the shop floor/factory floor, installed the Litmus Edge system, opened up two different capacities; data capacities, you are pushing everything to the cloud environment, Azure environment, confluent environments, AWS environments and more.

On the other side, they are utilizing our machine-learning, or artificial intelligence capabilities, and run times at the edge. So, all the different operational technology team members, IT team, data scientist, asset reliability engineering team, they are creating different algorithms, they are creating different flows, they are creating different Dockertized applications, and they’re running it at the edge on the top of the exact same data source. So let’s say, if one team is focused on energy monitoring they will create the digital twin out of our product, and they will only collect three variables from that pool of data. The other team might be interested in exploring 15 data points. Another team might be interested in exploring ten thousand data points. So it’s phenomenal to see how they push boundaries of the product, that would be one of the fantastic use cases.

In terms of the wins themselves, if you remember in 2019, Momenta was advising us to implement the OEM strategy. The OEM strategy worked out very well for us, we started working with end-user customers in parallel to a gateway or IT OEMs, and then we started working with automation OEMs, and now we are working with hyperscalers. So that OEM strategy took a long time for us to build, we learned a lot of things on the way, but as of this moment, we’re getting standardized across large amounts of these OEMs. That means anytime they want to deploy something at the edge they are going to refer it to Litmus Edge. Nobody wants to clean up that mess, and we have done it so well that we’re getting the in-bound request at a volume that you can’t even imagine. That would be one very notable win for us.

18:03
Yeah, well said, and you’re absolutely right, it’s very-very long sales cycles to get into an OEM situation, but once you’re there, done right it’s an annuity stream because you just continue to generate additional sales at very low marginal costs. So for you guys particularly, because of your OT placement, I think that really made a lot of sense, and it sounds like you’ve managed that extremely well.
So, you provide what you call a unified Edge-to-Cloud platform for the Industrial Internet of Things, what does the term mean to you, and really what are the advantages of it?

18:42

The short response to that would be we call it a unified Edge-to-Cloud platform, because of technology and business model tech that we have implemented over time. We have done so many different types of use cases, proof of concepts, pilots with the customers in the early few years of our customers. We wanted to learn from them, we wanted to explore a new industry, we wanted to find new use cases, we always wanted a new customer. One thing that we learned out of it is technology complexities, they are the real deal-killer. For every small, medium, and large project that has failed, technology complexity would play one of the roles inside that.

So, what do customers care about? They don’t really care about all the technology pieces, but they do care that you are compatible, or you are utilizing those latest and greatest technologies. The latest and greatest in terms of security, scalability, you’re using let’s say future-proof solutions like Kubernetes, you’re using a containerized ecosystem so they can maintain it very well in the future.

So, before this technical complexity kills our project we had to start a unified approach in terms of how we deploy our product. So rather than selling technology pieces, rather than selling products, we started selling the unified Edge-to-Cloud strategy; that means you start with something at the edge, then you extend it to the central system, that central system can be hybrid, it can be in the cloud, it can be on-premise, it doesn’t matter. But this unified Edge-to-Cloud strategy was implemented in a way that customers use our product, rather than selling technical pieces, we started selling the joint solution. And in 2020 we improved our business model to the sited-based business model. This business model has been working out very well, once again the idea was not to sell products, the idea was not to sell the technical pieces, the idea was to sell the journey that customers are exploring. We started with foundation, you are growing so you are on a growth plan, and you are on a scale plan, and our products are combined together – edge product and edge management product combined together under that business model umbrella. That’s how we define unified Edge-to-Cloud in terms of technology as well as a business model.

21:16
I like it, it's actually as you said, it’s a solution journey in some sense, and the technology is kind of secondary to supporting that. But it is as we joked earlier, clients are pulling you to the edge, but do clients still need to bring that information to centralised systems. So, you’re able to start where the client wants to, and end where they need to in that regard.

Let me ask, how do you know when an organization is really ready to adopt your product, and what best practices have you seen in helping them realize that potential value?

21:55

One of the successful approaches that we have seen is a small start and slow and steady progress. There are so many challenges that you will find on the operational technology side, network infrastructure side, security side, cloud, and management side; if you try to solve all of them at once, those projects get bloated out of budget, they are going to get bloated out of technical complexities, human resources investment, and more. So the small start program, means start with the reactive analytics, before you can enable reactive analytics you have to go and implement the foundational data layer. So customers start by installing Litmus Edge, customers connect all of their assets, they’re still not utilizing the data, but everything is connected.

Now they will take the first step which is reactive analytics. If this happens, do something like this. Once they are familiar with that, they will go for condition-based monitoring, if this-this and this happens allow me to prevent the next step. Then they will go, this is purely statistical analytics and operational technology team members, as well as OT executives, they should be able to design it. Then they will go for the next step which is, ‘I want to predict before it happens,’ so they shift from reactive to condition-based, now they are going for predictive. In terms of predictive, they will utilize the power of the cloud, they will utilize the stored data that they have to understand the patterns, understand the anomalies, understand the behavior of their assets, and predict some specific challenges, and they’re moving more towards prescriptive analytics as they go in the future.

So this is the journey. If you ask somebody, or if somebody asks us, ‘Can you prevent my machines from failing by using artificial intelligence in the next two hours?’ We’ll probably walk away from that deal, it’s not possible, realistically it is going to fail. So it’s a step-by-step journey, it’s not something that we can realize at day zero, but at six months of effort, a year of effort, you will fall in that predictive-prescriptive journey. That’s where over 60 percent of customers are. So this is what we have done, start small, follow a journey to get to the point that you want.

24:27
Well said actually. I always summarise it as think global, act local, grow organically! So start with the big vision in mind, but do something small if you will, and then grow from there. As you say, follow or lead your clients down their solution journey in that regard. This next question probably apropos to that, I’ve heard many analysts and I’ll call it IT technology vendors bemoaning the fact that the adoption of the Industrial Internet of Things has been slow, some will say it still has not caught on. What’s your perspective on the speed of adoption here, and what would you forecast for the next few years in terms of the adoption of technologies like yours in the industrial IoT?

25:18
Industrial IoT is a journey that customers will take. In that journey there are so many different things that they have to solve, coming out of those 1970s/80s functional manufacturing systems you are entering a whole new data-driven world. So, it’s a journey. There are obvious challenges like data challenges, so many different vendors, some of the large companies they grew with acquisitions over time, so they’re having to find isolated siloed data everywhere. That data challenge needs to be solved early on.

The second one is network problems. Bandwidth is simply not sufficient, everybody wants to - ‘Let’s just push everything to Azure,’ yeah but, who has that bandwidth? Like the operation technology systems we see they are on the serial systems, RS 485 10 Mbps connect systems, they are getting replaced and they’re getting more cloud-ready. They are implementing those bandwidths at the operational technology level.

The third one will be security problems. Without industrial IoT, if you go on a shop floor what they do is, they do the parameter security, which means you enter on a plant floor, they will have a firewall in front of it, they will have a firewall on the top of IT network, OT network, and they think everything is secure. But when you are enabling hybrid connections, when you’re enabling your operational technology systems and they are connected to the cloud directly, all of your system, all of those PLCs DCS, edge computing system, they are on their own to protect themselves. So, the shift from unsecured communications on the plant floor, to shift to every system is well-protected, that’s not an easy problem to solve. And the last and most important one that I would say is, there is just too much noise with the repurpose solution.

The repurpose solutions which are legacy OPC windows, legacy SCADA platforms, they are just redesigning their websites and repositioning themselves as an Industrial IoT solution. They think the data problem can be solved by using those Windows agents or Windows services, and unmanaging past structure. So every time modern companies, companies like us and many of the modern software companies take three steps forward, these guys force us to take two steps back. You’re going to find security and lots of vulnerabilities identified, projects are getting killed, projects are getting bloated out of budget, and far more.

So, there are multiple network challenges to the data challenges, security challenges, and just too many repurposed solution, they are holding us back up to a point. It doesn’t mean it’s very slow, you might not realize that everybody is doing machine learning right now, but people are at various stages of solving these four data challenges. And in the next two years, or in the next few years type of timeframe we will see that there will be vendor consolidation, a lot of different… there’s operational technology sites, as well as the data provider, network provider, so you are going to see a whole lot of consolidation and there will be core system consolidation as well. It will still yet to be seen, but most likely the vendor ecosystem consolidation will be driven by hyperscalers. So your product is very well compatible with the Azure ecosystem or AWS ecosystem, Google ecosystem, and more. So the next two years will be exciting.

29:17
It will, and because one of our practices focuses on the executive search of really digital industry leaders, we have certainly noted the amount of hiring the hyperscalers are doing, between all the three largest ones. And so I think they all see the opportunity as well, and it sounds like you’re well-placed in terms of working with them.
You have been named CEO of the Year on at least two occasions that I remember, but most recently by the IoT Breakthrough. To what do you attribute these recognitions?

29:53
I attribute the recognition to my very supportive family, while I’m doing this entrepreneurship, my co-founders who are as dedicated as you can imagine, and the whole team behind me. The Litmus team, it’s one of the most amazing group of people that I’ve worked with. I started the company with a simple vision, but a lot of these different skill sets, a lot of these amazing people that joined over time, are executing that vision every day. So together we are implementing the products we are designing, we are getting in front of the customers and we are improving the industry as a whole, and that is the reason why I’m getting the CEO of the Year award. So I would definitely attribute it back to them, the guys who are doing real work.

30:46
Yeah, in fact, if there were truly CXO leadership team awards, I could see you, John and Sasha getting it together, you guys have been the three musketeers since our earliest meetings, and I’ve always been impressed by both the intellectual horsepower, but also the moral alignment in everything that you guys do. So it’s great to see.
I always like to end up with questions, what inspires you, or how do you find your inspiration?

31:23
I’m still one of those people who find inspiration from the customers, and I’m lucky enough to get in front of customers as many times as I want! Every time we get in front of a customer, or a potential customer they will teach us a new thing, they’re exploring the challenge, or they have a specific problem that they would like to solve, and they do inspire us to do better. They push our boundaries, and they are extending our products in a way that we can’t even imagine. So that would be my first point of inspiration for the job that I’m doing right now.

In terms of the broader ecosystem, whenever I get the chance I listen to various podcasts that’s just a typical commute thing that I do, all the way to specific industrial communities which are popping up everywhere, and some fantastic people there. More and more the blogs and Industry 4.0, communities in European Union, they’re just doing fantastic work. So those are the key points of inspiration. And now, you’ve got a loyal subscriber on the Momenta podcast channel. I explored it just recently and then I got addicted to that, so you’re on my top list.

32:58
Well, I appreciate that very much, and it’s great that you’ve been able to participate in this. So thank you again for spending this time with us today Vatsal.

33:09
Thank you very much for having me on the podcast. And for the listeners, we are one of those customer-obsessed companies, let us know if we can help you in your industrial IoT journey, and please check us out at Litmus.io Thank you, Ken.

33:26
Yes, thank you as well. So this has been Vatsal Shah, the founder and CEO of Litmus, and if I may say so, you think you’re getting a technology company when you buy in, but you’re really getting a digital industry solution Sherpa to help you along your digital industry journey in that regard.

So thank you for listening and please join us next week for the next episode of our Digital Thread Podcast Series. Thank you and have a great day.
 
[End]