Podcast EP 108 - Data-enabled services for machine builders - Florian Weihard, CTO, rhulamat Automation Technologies & Board Member, ShopWorx

EP 108 - Data-enabled services for machine builders - Florian Weihard, CTO, rhulamat Automation Technologies & Board Member, ShopWorx

Nov 11, 2021

In this episode, we discuss how telemachine builders and production line integrators can differentiate themselves by providing data enabled services on top of their asset bases. We also explored the challenges of both gaining access to factory data and of gaining customer willingness to purchase software from non traditional software providers.  

Our guest today is Florian Weihard, CTO of rhulamat Automation Technologies and board member of ShopWorx. ShopWorx is an IIoT platform designed to make manufacturing shop floors more efficient by providing real time visibility and control. 

IoT ONE is an IoT focused research and advisory firm. We provide research to enable you to grow in the digital age. Our services include market research, competitor information, customer research, market entry, partner scouting, and innovation programs. For more information, please visit iotone.com

Subscribe

Erik: Welcome to the Industrial IoT Spotlight, your number one spot for insight from industrial IoT thought leaders who are transforming businesses today with your host, Erik Walenza.

Welcome back to the Industrial IoT Spotlight podcast. I'm your host, Erik Walenza, CEO of IoT ONE, the consultancy that specializes in supporting digital transformation of operations and businesses. Our guest today is Florian Weihard, CTO of Ruhlamat Automation Technologies and board member of ShopWorx. ShopWorx is an IIoT platform designed to make manufacturing shop floors more efficient by providing real time visibility and control. In this talk, we discussed how machine builders and production line integrators can differentiate themselves by providing data-enabled services on top of their asset basis. We also explored the challenges of both gaining access to factory data, and of gaining customer willingness to purchase software from nontraditional software providers.

If you find these conversations valuable, please leave us a comment and a five-star review. And if you'd like to share your company's story or recommend a speaker, please email us at team@IoTone.com. Finally, if you have an IoT research, strategy, or training initiative that you would like to discuss, you can email me directly at erik.walenza@IoTone.com. Thank you.

Florian, thank you for joining us today.

Florian: Hello, Erik. Thank you for having me, looking forward to our discussions.

Erik: Florian, really interesting topic today. I think also not just from a technical perspective, but also an innovation perspective. We're going to be looking at basically how a system integrator has collaborated with entrepreneurs to develop a new technology company. So if I look at your CV, it looks like you're starting from somewhat quite traditional mechanical engineering, and then getting into the IT topic, and now you've managed to combine those quite nicely in your CTO role. But can you just walk us through a little bit how you managed to end up today as the CTO of Ruhlamat?

Florian: Originally, I'm from Germany, and I studied in Bavaria, mechanical engineering, research and development direction. And then while I was studying, I started my internships and finally did my thesis at a German construction equipment company. That was a quite interesting because heavy machines that was very interesting for me as a design engineer, and at the same time a super international company it was 95% of the business outside of Germany. So I also was hoping to also one day see the world, but turned out not to be so easy because a mechanic engineer should sit in front of his computer designing something instead of traveling the world.

So it took me some years, and then I ended up in Shanghai in China and build up for the company some engineering offices in China and Malaysia. And then 2015, I was thinking of, there should be also some other things in the world then construction equipment and pivoted into the automation manufacturing space, and started as an engineering director in the automation equipment space.

And from 2015 onwards, I also saw that industry 4.0 topic was really evolving and getting a lot of attention, so I started to study what's possible there, what kind of technologies are available and what does that actually mean for machine builders who do automation equipment. Then in 2017, I started to work for Ruhlamat, my current company as a CTO, and that's traditional machine builder; at that time, not much digital competences and accordingly, not much a big bigger footprint there.

But at the same time, we also realize that really leveraging data as a machine builder is essential for the future, that will be the major value stream one day for automation companies so we realized we need to do something. The journey started that we found our digital company, we started to develop software and today we can offer something.

Erik: How did you come to this realization? Because on the one hand, it's something that a lot of people talk about, but if I look at machine builders in the market there's maybe a lot more talk than action. How is it that Ruhlamat relatively quickly went from identifying that there's something important here around building data-driven solutions and business models to actually establishing a scalable business here?

Florian: Yeah, that comes actually because before I even joined Ruhlamat, I actually also with some friends together started angel investment fund, and through this fund, we invested into startups which are having solutions around industry 4.0. And out of that, I built up quite a network, which we could then at Ruhlamat leverage. And finally, we built in October 2018 startup, which was kind of a joint venture between three parties, Ruhlamat is automation company, Entrib Analytics as a software company and our investor behind that putting some money into the game.

And then we could start immediately with a company. We had right away a product which we could use IoT platform, and started to build solution. And I think bringing these three partners together was quite advantage because these are really, let's say, like three different DNase coming together: Ruhlamat roadmap from the automation side, the application side, really the manufacturing plant side, then Entrib Analytics really coming from IT side. And then our investment partners also coming we are from really investment, building a company perspective, but also bringing in a lot of data science capability which we could leverage. So we had really all the competences on the table right away, to start immediately building some nice solutions.

Erik: What we see be an IoT ONE work a lot with corporates that are setting up some type of innovation initiative often here in APAC, and we see a lot of collaboration between larger corporates and startups where corporates look at startups, as you know, accelerators are, you know, allowing them to get an idea kind of through MVP stage. What we don't see very often is a VC in the mix. And I think that's in part because VCs are often hesitant to actually step into this, they say, hey, the corporate is going to control this. It's a strategic initiative. I'm never going to make my money. I'm never going to be able to actually exit this company.

You had trust, and it sounds like you had this background. But you had to convince your investors that it actually made sense to collaborate with the corporate. I suppose you also had to convince headquarters that it made sense to take this approach through a startup. Was that a seamless process? Or how did you actually navigate those conversations?

Florian: That conversations were quite tricky. As you say, that's really bringing totally different worlds together. So let's say like a family-owned, small and medium enterprise business together with really a complete new technology. In the end of the day, we really talk about the transformation, which also in our Ruhlamat Automation Company in our team, there was not really this awareness that this is needed. And luckily, the Ruhlamat top management, although our owner trusted me a lot so that we could get that started.

And the same thing as you said from the investor perspective, because for sure, venture capital funds are not so attracted by such a constellation because there is no clear exit scenario, and so on. But on the other side, it also came with a lot of advantages and disadvantages we can really take advantage of today. I mean Ruhlamat is not just one of the shareholders. It's actually a fantastic platform, which you can leverage in terms of being really a customer of this new company. And the biggest customer also a fantastic lab for new solutions, new technologies, the fantastic network also, to find our customers.

So in the end of the day, that is also a kind of a rocket start which you can bring to such a startup. And over the years, it turned out to be the thing, which was really helpful. Because I remember very well, when we started that we were not the only ones doing that, but right now, if you look out there, a lot of them are no longer around. Also, I have learned a lot of lessons over the last years; things which I thought that will be easy thing turned out to be really very difficult.

Erik: But yeah, I imagine there was a lot of negotiation and communication to make the three parties comfortable with each other. It's also interesting that you're running this out of China, because if I think about a family-owned German organization, I typically think of some medium sized town in in Europe where a group of people run kind of orchestrate everything that's happening in the world, especially around anything related to R&D and then China being more of a sales center. Seems like this is not the case here, where here you're the CTO, and you're sitting here in China. But was this also kind of top down corporate strategy? Or was it more you personally being the right person in the right place and then just driving this? How much was this a corporate strategy versus scars aligning around you and the network that you bring, and this particular interest that you had in pursuing with this approach here?

Florian: I think it’s important to understand in this context how Ruhlamat works. Ruhlamat is actually a very untypical German company, because we have around 1,000 employees. We do automation equipment for tier one automotive industry mainly. And our headquarters is in Germany. But at the same time the really center of the operations I would say is in China, because it's by far the biggest operations. We have here around 700 people, 230 engineers in the R&D, and actually also serving the global market out of China. So it's a big center here.

Sometimes some say we are a kind of a co-petition, because every plant has its own product, its own strategy. And for sure, we are trying to have synergies globally and working together wherever we can. However, we also are very decentralized, and everybody can maximize his profits however he wants. And this setup made it possible that we in China basically could completely independently decide what we wanted to do.

And I think with me, starting in Ruhlamat also this idea of industry 4.0 only really came to the company, and with my role as a CTO and being in charge for the innovation and technology roadmap and strategy and everything, it became a kind of corporate strategy, not meaning that the whole company understood it and supported it, not at all. I think it took us another one and a half, two years until it really everybody started be aware of these technologies, the importance and also understood actually the first time how it really adds value to the automation business. And part of this really transformation, which kind of was seeded by this company, our digital company, but took a while to really evolve.

Erik: So I think we're talking here about ShopWorx, and as I understand, if we look at it from a value proposition perspective, you have actually a couple different paths. You have the machine builders, and then you have the end customers who would be typically a tier one automotive supplier; could be probably other industries, who would be using the solution. What are the pains that are being solved here? What's the value proposition?

Florian: To explain our business, there are two perspectives right now. So one is ShopWorx is actually IoT software platform, and based on this platform we built in the beginning. Also, our partner already, before we actually found our company really plant MES manufacturing digitalization solutions. So this is one area where we classically, for example, build plant MES for small medium enterprise, something like that.

And then what we started actually with our company in China together with Ruhlamat is then building really a software platform which is especially made for machine builders. Ruhlamat was basically the customer for that, so enable Ruhlamat to build a line management on control system as a basis for all the equipment, which then on top can really leverage these real time big data possibilities, data analytics, and so on.

And the funny thing is actually, in the beginning, they went some separate routes, but they are coming back now quite intensively. Because once the automation lines are delivered with our ShopWorx system, then very often customers realize, oh, that's nice, so what else can I do with that? And then all of a sudden, you find yourself in a discussion about a plant MES, or data warehouse project or extreme analytics project.

So you ship a machine with a ShopWorx platform together with a Ruhlamat machine, for example, and then the customer ask, by the way, I have three other lines there from some other companies, can you also connect that? So then I have a central management system.  And then I also have some other machines, can you also help me with that? And then you also connect some to edge analytics project connecting some other machines making sense of these data as well.

And even the final step we are doing right now is you're making that a real end-to-end solution. A lot of customers have a lot of lines, and they're all using the same system, then you can connect that system back to the OEM using a kind of a service portal, so a cloud of the OEM, which can then connect to the line and can offer additional digital services, maybe around maintenance, for example, from the OEM site to the final customer. So this is the second business model, which is now coming up.

Erik: So you provide data ingestion from different sources, analytics dashboards to visualize, is that kind of the core horizontal system, and then on top of that you can build applications such as maybe an MES or machine vision solution or so forth? What I basically think about this as a horizontal platform for managing data across the production line, and then either the machine builder or maybe the end user could build vertical applications on top of that platform?

Florian: Yes, kind of, so maybe coming again from the solutions, which we made for machine builders basically. So if a complex automation line is delivered on site, it's a line can consist of 20-50 different stations, quite a complex process can fill a whole workshop. You need to have traceability on this line that's an automotive industry and must have. You need to manage your recipes. You need to manage your production orders. You need to manage your materials. You need to manage your rework, your quality information. So actually, our system gives a complete solution, which makes the whole line control line management for the line.

And on top of that, and that's why we call it auto line MES, it gives a lot of MES features, for example, the order management, which typically also standard many customer sides on the higher level in a plant MES system. But a lot of other customers, especially smaller ones, they don't have a separate plant MES, so they manage directly also all these lines through our line MES.

And this system can be a server at a line, but you also can scale that then up. If you have a second or third line, you can integrate that into this server, and with that, it's becoming a kind of a plan solution right away, because it's jumping out of this single line. And then step by step you can scale it up also is more modules also linking up to more machines so that it even can become a plant MES by itself, which we also do quite often.

Especially new factory setup in China, where we are delivering the first automation lines, very often we immediately become also the plant MES through that because it makes no sense to put in there another layer. So that's a quite elegant solution to handle all these tasks. And with that, for sure is coming all the visualization part, the dashboarding part, the report part about OE, about quality, productivity and so on. Then there are additional solutions like maintenance management.

So the scheduling of the maintenance, condition-based maintenance features, the whole documentation around maintenance, we also have a module which is a spare part warehouse where you can manage also the company's spare parts, and the consumption of that having a minimum stock there, and so on, so complete maintenance management system is typically also built around this automation lines.

On top of that, once you have that a server there, you are accessing really all the production data, you are also able to build analytic solutions, targeting, for example, improvements on the quality side or productivity side that is a nice leverage then as well to make the lines more efficient, solve huge problems with a digital approach instead of mechanical.

Erik: I guess there's this perspective from Ruhlamat, where you're deploying ShopWorx through Ruhlamat’s machines, and then that allows you to have a larger footprint in the factories of your customers. Then there's the second track here, which would be other machine builders, adopting ShopWorx and integrated into their machines, and I imagine in that case, they would be looking at this as an enabling technology that would allow them to build potentially new revenue streams and so forth in addition to I suppose being able to better manage their assets and maybe provide higher service levels.

What's the perspective look like from a machine builder that is a customer of Ruhlamat? Is the value proposition oriented around them building new revenue streams on top of their asset base through this, or is it more of differentiation from other competitors by being able to provide kind of free services that'd be combined criteria?

Florian: I think, for every machine builder, these kind of solutions are really, absolutely essential, if they want to be premium automation machine builder, for example. Because using the data from the equipment, production data is essential to build better machines. I can give you later on when we come to the use case to some examples around that. So the problem is, if you do not by yourself as machine builder, if you're not able to provide a solution based on your production data to your final customer, others will do that because the final customer will not say, okay, if the machine builder is not doing that, I don't do it.

So they will just work with somebody else. And then somebody else will work with the data from your line, and basically, automatically also would take over the user interface to your customer. Because once you work with the data, you usually work with a software platform, and all these user interfaces become an additional layer between your final customer and your machine. And then step by step you're degraded from a premium machine burden to a simple mechanical parts supplier because really the big picture is done by somebody else.

And at the same time, it's also a big challenge and that is something where we really can help because in the end of the day, you as a machine builder like, for example again taking Ruhlamat as an example, we are more a system integrator than really a product company, which means we build highly customized lines for customers also with them together, which also means our customer considers the process which is actually running on these machines as their IP. So they would never ever allow that you have real time production data access to that, because they want to keep this knowhow for themselves and want to protect it as much as possible.

So that means that you as a machine builder actually not really having access to data from your own machines. Because the data you generate when the machine is in your own workshop, when you debug the line, that's data trash there, you cannot work with that. You really rely on accessing the data at the production side during the production, and then you need to work with your customer together because you cannot just get the data without giving the customer something in return, a big benefit.

So you need to find a way to really work with the customer together to develop high value solutions with this data, which then in return gives you the opportunity to also access and work with data. So that's a kind of win-win situation, but at the same time, also a chicken egg problem, because how you want to give your customer benefit if you don't have data?

And this is something which all is kind of integrated in our ShopWorx line MES solution, where we actually can deliver a benefit in advance already, which is so interesting to the customer so that we also can work together on the data and really placed the system on side. That was a big leap and that changed the whole thing. And today, actually, there is no line anymore which is not going out with this system. And there's no customer, which is not asking for that, it's really a must have today. Some customers, but that is maybe 1 out of 10, they have own solutions, for example, and they don't rely on our system, but 9 out of 10 really go with our solution.

Erik: This makes sense as a machine builder because you're then building an integrated line and so you then have a platform that can aggregate data points from across this line. If I'm a machine builder that's building some maybe critical machine but it's only one machine in a line and there's things upstream and downstream that produce data that might also be quite relevant to me, what's your experience there? Because I feel like that's the most difficult position to be in a factory where you're selling an isolated piece of equipment into a line, and on the one hand, there's some value in being able to analyze the data coming off of your equipment, but you're kind of a stone in a lake. Eventually, the plant is going to want to have a platform that integrates across the equipment.  So eventually, you need to either be that platform or somebody else is probably going to take this away.

Or even if you have a very strong value proposition, you probably need maybe some data from upstream, downstream to help enrich the information on your equipment so you can understand maybe why we're producing a certain level of quality because the data coming off your equipment might not be enough by itself to be particularly interesting aside from predictive maintenance or something like this.

Because I feel like in the industrial ecosystem, this position of being a machine builder that’d selling one or two pieces of equipment into a larger production line is kind of the most difficult position to be. And it's a little bit the same as being a tier two supplier into automotive where you have a couple critical pieces in an engine, but by themselves you don't have the leverage to do very much with the data coming off of that. It's with the integrated that has more leverage there. Do you see any interesting strategies for these players? I guess they are a customer group of yours, but how do you see them find the leverage in this position?

Florian: So if a customer like Tesla, for example, they have so many resources and such a clear strategy on that, for sure, they would not take hours. They would say give me your interface, I pull the data, I do everything myself. But that's not the typical customer actually. So if you look at our tier one and maybe tier two customers in China and also a lot of international players, you would be surprised how much they still struggle with really getting the data together, building a strategy on that.

So what we saw is that what we are offering right now that is a fantastic shortcut to get that started. My feeling is that a lot of them, they also have learned a lot of lessons in the past from being very dependent on one central software company doing all that for them. So they very often really pushing also in this direction to decentralize that a bit, having different suppliers in their factory, having an ecosystem of solutions. And this is something which this role we can play. And we can play basically a kind of a rocket start there, because we have our system, and we link up a very complex system, which itself already has so many parameters, and where we easily can also integrate other equipment maybe up or downstream so that you really can make sense out of that data and use that.

I can give you a one example, which shows a little bit what the leverages and at the same time how sometimes this is seen maybe a bit too broad from the data side. It's a data science project so I cannot really share too many details in this case, because it's really about their production process. But this is a big customer we built for them over, I think, a decade roughly 140 lines. So we are their number one supplier. They have around 130 plants in the world, and automotive tier one supplier.

And these 140 machines which we built, they were not bad. They went through quite some redesign over the years to improve the machines to bring the cost down. However, only after 10 years when we started in the first time really work with the data of these machines together with the customer we saw that there's so much way for improvement. For example, for this customer in one plant, we linked up actually the whole plant, the process is up and downstream to our ShopWorx server to have all data together in order to find a solution to solve a problem which they had with two specific defects from these lines. They can happen maybe one time out of 10,000 pieces produced. And in the end of the day, if they go out to a final customer, they would cause a recall.

And what made it very difficult to find them is you need a destructive test: you cannot see it and you cannot measure it. So there's always a chance that this goes out to the customer and causes a multimillion dollar recall. So what we did with our data science team and the customer together and our team, we tried to develop a machine learning solution, which can identify this part, predicted that it is bad just from the process data from the process correlating that.

To our surprise, it turned out finally we needed only six process parameters to really predict it. We produced every day to gigabyte of data just from the process of this line. And in the end of the day, it turned out we only need six out of that. So for sure, that was a learning curve, analyzing all this data, doing a lot of experiments. And finally, this little handful of data, some sensors in our machines could solve this huge problem. And today, we can 100% predict these defects and completely avoid recourse. So that's a really a multimillion dollar problem solved. So it's an unsupervised machine learning model which we have developed here.

And that shows a bit the leverage you have with data on the effectiveness, on the productivity, on the quality of the equipment, without that you cannot solve this problem. But solving this problem you don't need a complete plant solution. And if you wait for that, maybe you need another 5 years or 10 years until you can solve it. So with our approach, you can start right away: you have all the data there. And if we miss something, we just edit. I mean you can add a sensor here or there. So you can build your data inputs as you need it and then make a brick big progress without looking at everything.

Erik: Let me ask a few kind of rapid fire more on the tech side. So are you cloud first? Do you also deploy on the premise?

Florian: So, ShopWorx is a hybrid platform: so we can do both we can do on premises, we can also do cloud deployment. It turns out that especially small medium companies which looking for plant MES, and don't have so sensitive processes, they prefer very often to have a cloud solution because it makes their IT infrastructure very slim. And in China, even we sometimes have the funny constellation that they don't want to ask their headquarter. So every server in the plant, triggers a lot of questions.

But all the other customers which really have very complex production processes, like a typical tier one automotive customer, they need on premises server. They pay a lot of attention to really protect data. The discussion about cloud is today not possible. So this is all on premises solution.

The central strategy for us is having a self-serve platform. Because imagine you are a machine builder, and you want to use ShopWorx for your production line, and you want to build for your customer a solution. So you need to be able by yourself to build the line with this platform, to configure it and so on, rather than having somebody else a software company in the project, which is working with you, you would lose a lot of confidential data, a lot of information.

And at the same time, this is also for the final customer important that they can do things without always coming back to us asking for customization, because simply they would share too much confidential information by doing that.

Erik: What would be the skill sets required from the customer from a data science perspective to build a machine learning algorithm on the platform? Is this does this require a relatively high skillset? Or is it somebody that goes through training and has a basic understanding of how these things work, and they could work on the platform and produce effective algorithm?

Florian: So this is a machine learning model, which was developed by a data science team. In this case, it's our data science team, but actually, it doesn't matter. If it would be the customer or whoever who developed it. So our system, in this case, on a ShopWorx app stream analytics what we use there, you can deploy this machine learning model by yourself in the app stream analytics platform, and train it on the platform without coming back to us, and can then manage your devices there and deploy it on certain lines, retrain it there with some data sets.

You still need to build the model by yourself, for sure. But for reasons of confidentiality, you would anyway and not outsource that, you would do it by yourself. But the rest of the whole platform to really roll it out and manage it, this is a set of thing.

Erik: But how scalable do you find this? Especially when we talk about machine learning algorithms, one of the key questions is really, how scalable is the algorithm? Meaning we first have our team of data scientists that build the algorithm that is relatively high cost, right so there's not going to be much of an ROI if we have to do that every time we wanted to play this on a new production line. But ideally, the algorithm is scalable to some extent, so that when we deploy it on it on another factory that we have somewhere else in the world, as long as the parameters are similar, then it's maybe 20% of the effort to scale. How do you find or maybe you have a mechanism for assessing how likely an algorithm is going to be able to scale without a significant cost burden because that's really going to impact the ROI of the use case?

Florian: There are a lot of like standard analytic things you can do, which are always the same. If you talk about automation line, you want to optimize your cycle time, for example, that is something you have bottleneck somewhere in a line, you have processes. So this you can analyze, and whatever process is running on the line does the same thing. So this is something which is scalable, and which can be a standard solution actually.

What I was talking about before a predictive quality model, which forecasting a very specific defect, that's a highly customized thing, or a highly custom made thing. It took us a year to really develop this machine learning model. And here also, in terms of scalability, it was not easy, because imagine you have this customer, maybe 700 of these machines across the world. And in all of these machines, operators can change process parameters. Whenever you change a process parameter, you need to retrain these models and need to test it again. So you need to have also a good platform to really make that possible decentralized somewhere. Or operator, after changing something, he still can retrain the model and deploy it again and make sure it works as well.

So therefore, from the model side, you need a really advanced model which can handle that changes. And on the other side, you need a good technology roll it out. ROI for that, for sure, you need to know that the problem you're solving is really big enough, that is a multimillion dollar scale, and that you can really scale it across the world in all your plants because otherwise, it makes no sense to invest a team of data science guys together with also because you really need a cross-functional team with your production team, with your machine builder and so on all together. But the leaps you can make a really huge.

So I think in future, we will see that more often. But the basis is to bring a platform there, to make data available, to do the first steps. And without that, you cannot jump to this next step. So I think it's a step by step process for our customers.

Erik: So I think one of the challenges in applying data science to industrial is exactly this challenge of scalability, as opposed to consumer where you're on massive platforms where people have very homogeneous needs, and then datasets are pooled. So you can build very effective algorithms that work for a lot of different people, solving similar problems. Hypothetically, you could do that if you had more coordination, which I think we don't have in industrial environments, meaning, let's say one machine builder on one plan to build an algorithm that effectively solves a problem.

They could potentially bring this to other plans from other customers. But often, that's not possible because the customer won't allow the data was used to create this algorithm. Do you have conversations ever with customers? But do you see a future where there might be more collaboration where if somebody allows their data to be used to build algorithms, they might receive a discount on the data science service of building the algorithm in the first place, because they're basically contributing to the R&D of a solution that can then be more productized? Do you see any kind of interest from the end users in this? Or is the mindset still from your experience very much focused on data security and IP protection, even if the risk is more hypothetical?

Florian: The first thing I always tell our sales, if somebody is new coming in, please categorize your customers because the biggest mistake is always to put them all into one pot and say that's my customer, that's what he needs because actually, it's not that homogenous. There are other very strong big tier one automotive customer, like of Ruhlamat can be a Bosch, a continental. Such companies, I mean, they have a lot of resources, they also take that very serious with confidentiality. There is not much possible here especially if you're talking about really bringing solutions and leverage them somewhere else as well.

But on the other side, there are a lot of customers who have now also increasingly high pressure to bring down the costs. And they need to really make progress with their production. And at the same time, they also have limited batches which they can invest in new lines. And they are very open. They have solutions which you can leverage across customers, especially Chinese customers are far more open to that. German customers, a lot of our tier one customers, they are very careful, and we respect it. This is our responsibility as a customized machine builder to really build a solution for them.

But here we have also some customers where we really have built up a trust over many years. And with them together, we can really do not just projects which have a kind of a one year focus and are really projects which right away are needed in production but we can really work on strategic R&D projects together, which have a three years horizon where we already today think of how we could solve certain problems with a data analytics idea and we make some proof of concepts here and there, some smaller R&D projects always with the goal to really bring that to a production line.

And what happens then is within these customers, so because these are also big organization, there's very often more than 100 plants, so they scale it also globally. So actually, let's say like this market within this customer is also not so small, so you can then achieve certain scalability and return of investment. But I also want to warn, especially the machine builders to always see software and data solutions as a cash cow, especially from day one. This is also something which in my company in Ruhlamat, took us a while to understand it.

So these solutions enable you to build a better machine and you will get a better price for your machine, you will have an advantage over your competitor. So it will strengthen your core business. It will strengthen your margins, very often not directly through on the software price tag, but just making you more competitive and more profitable. So I think for this machine builders, this is really also a tool to have very strong competitive edge, which becomes increasingly important, especially in China, as you well know.

Erik: The business model for ShopWorx, is it pure SaaS? Is it by the machine, by the seat? How do you price this? Because I guess there could be quite a few different configurations depending on who the customers.

Florian: Yeah, so there is one basic price which is actually for the platform itself. And then they can be on top a customization, which is maybe required in some cases. So, most of the customers, they still today they have a budget for the automation line and they make a budget for this part separately, either for a planned MES or for really line MES, maybe linking also up several lines. So there's a clear budget. There's a clear specification so it's very traditional. So you just your license to the customer and with that then you have a kind of annual maintenance contract, making really sure that you can keep it up to date and that you can react very fast.

Because a tier one supplier who is using our software, for example, on the line to manage the line and handle the traceability, if the software is down, the production is down because no traceability means no production. So, you can produce a lot of scrap, if a software is down, causes the production to be down as well.

We with time moving forward, for sure, SaaS becomes a very interesting topic. I personally love it also because as soon as you talk with the customer about SaaS, it shifts away the attention or the discussion from the technology to the benefit. Because you don't want to have a recurring payment for something which doesn't repeatedly currently or all the time add value to you. So it’s a very nice thing to see how this shifts the whole discussion. And that shows also that a lot of customers are actually in manufacturing plants, in automotive industry not yet ready to talk about something like that they cannot really projected to the way they make their budgets.

But it shows already that some areas like optimizing a production line, optimizing a cycle time, things like that. That is a value which makes you more competitive over time. And the longer the line is running, the more data you have, the more you can optimize it. So these are kind of analytics apps, let's call it like that, which you can roll out on an existing platform, which you then can move indeed to a SaaS business model.

Because also, from a scalability perspective, again, this still needs deployment, it still needs hardware, it still needs an expert like the automation company to really deploy it. It's not Office 365, which you download, you started and you use it so you still have a component there, which is not so scalable. You sell that with a line. You sell that separately as a software solution. But on top of that, then you have a scalable platform where you can upgrade things without additional hardware deployment, and then it becomes interesting.

Erik: Is there anything that we didn't cover here that's important for us to touch on?

Florian: We covered a lot already, I think one of the biggest challenges we still have is really and this is what also the machine builders have is make the whole thing really scalable. There's still a lot of customization needed in this space manufacturing. Even you think, especially when we worked in the beginning with small medium enterprises on some IoT projects, plan MES projects, you think now you have covered everything, there cannot be any customization anymore. But there's still something different. So make the technology really ready to handle that without huge customization effort. That is the thing which we have a big piece already solved. Definitely for the automation companies, the machine builders, because this can be quite standardized, but there's still a long way to go to make that on a higher level also scalable.

Erik: This also is my feeling right now. And we talked about this adoption of, let's say, for lack of a better word kind of next generation software for industrial is solving the problem of scalability. It's often not necessarily bringing the most cutting edge technology that's often actually not required to solve the problem. But it's how do we do this in a way that doesn't require a huge system integration budget, and yeah, can scale given the expertise that a company has? Because these tend to be the unglamorous problems that need to be solved and are I think, right now, in many cases still left unsolved.

So it sounds like that's a place where you've made some good progress. I guess this will be a focus for the coming years. Aside from scalability, if you look forward in the next two or three years, are there any other big areas of development that you're going to be prioritizing?

Florian: I think in the end of the day, for us, there are two directions, which are extremely important. So one is really building analytics solutions where we really can leverage the data. Because the first phase of our company when we developed this was actually building the platform on site where you can have the data, where we bring them together, and where you can roll out the first analytic solutions. But there's just so much more to do to really improve the line.

And the other direction is making the whole thing really end-to-end solution for a machine builder. So if you have your line MES servers on site, and you have several of them at several plants at several customers, bringing all that together, then to a central service platform where you then can based on this data of digital services around your machine, really selling the spare parts, maybe automatically triggering that through condition-based predictive maintenance and so on, so really making the whole maintenance service business model for machine builder. That is a huge topic and that's something where we are working intensively right now.

Erik: It sounds like you have your work cut out for you. I wish you and ShopWorx team a lot of have success in the coming years. And thank you again for chatting with us. Last question from my side is what's the best way for our listeners to either get in touch with you or with the ShopWorx team?

Florian: So I think the easiest way to get in touch with me is on LinkedIn. Erik, so I think everybody who's following your podcast should also be able to find me, and also always happy to learn a lot. Talk with other machine builders, that would be fantastic.

Erik: Thank you, Florian.

Florian: Thank you, Erik.

Erik: Thanks for tuning into another edition of the industrial IoT spotlight podcast. If you find these conversations valuable, please leave us a comment and a five-star review. And if you'd like to share your company's story or recommend a speaker, please email us at team@IoTone.com. Finally, if you have an IoT research strategy or training initiative that you would like to discuss, you can email me directly at erik.walenza@IoTone.com. Thank you.

test test