Strengthening Supplier Relationships with Digital Twins

Speaker 1: I know a lot of the time we feel that intentionally or maybe unintentionally, that we keep our suppliers at arm’s length, and really having that trust and transparency can be difficult when navigating these new unknowns that we’ve been talking about. So we’re gonna kick this next session off with a panel, but first we’re gonna hear from one of our panelists, Stan Kroeker, current department lead engineering for CAD PLM and an aPriori admin at AGCO Corporation and learn about how his team is utilizing their digital twins today. Welcome Stan.

Stan Kroeker: Okay, so let me give you just a brief introduction to AGCO itself, who we are as a company, how we use aPriori today and looking forward to how we use aPriori with suppliers in the future, and one more. Okay, so AGCO is a global supplier of agricultural and farm equipment. We have design centers and manufacturing set up around the world, and our goal is farmer-focused solutions to sustainably feed our world. We do that through machinery, we do that through control systems and software, and all sorts of things like that, and we are a global presence for the farmers and becoming a trusted partner for the farmers. And so some of the things that we do is really anything involving grain or seeds. We plant the seeds in the ground. We have computers that then monitor the planting rates, we tend the plants as they grow, we harvest the crop and as we’re harvesting the crop, we’re using software to take and monitor how much that particular piece of the field generated, how much grain was there, and then we score and condition the crop in our grain bins, and then use the grain bins and other systems to then feed and nourish livestock all the way through the whole life cycle.

So again, some of our smart farming innovations as we harvest the crop and measure how much grain we’re getting, when we apply fertilizer next year, we can use that map of the grain to then apply fertilizer at different rates throughout different parts of the field. I don’t know how much you know about the farming business, but one of the main inputs cost-wise to a farm is the fertilizer along with the seed, and so managing how much fertilizer you put down as in if you have high yielding areas that don’t need as much fertilizer, you can cut that cost back and it goes right back into your pocket. So next slide. So digital factories, we have done a lot of work within AGCO, we do a lot of building of our own products, and so a lot of what we have done is make very detailed digital factories of our factories. So we walk through the factory, we look at all the processes that are in place, we walk through the ERP system to understand what are the pieces in the ERP system that we put time into, ’cause time then goes into the money, and we do this for all of our different process groups, sheet metal bar and tube, roll forming, all the different processes that are in our factory. And the biggest thing that we do with digital twin technology and our digital factories once we build them is our manufacturing engineering groups then use these digital factories to calculate what the standard cycle time is to make a part.

How AGCO Uses Digital Twin Technology in Real-Time to Make Supply Chain Decisions

So I have a piece of sheet metal that’s in Creo, I run it through aPriori, I get a routing that will become the routing on a shop floor, and I get cycle time, the number of seconds that the part is at each one of those work stations. That cycle time becomes what goes into the business system, so I’ve replaced outdated and not near as accurate spreadsheets and user knowledge with essentially this digital factory. So we continue to do that digital factory work within our factories and move it to other places, different factories like that, but one of the big challenges is how can we take and put this same concept into our real-world suppliers? The idea being that if we can take and look into their factory… Let me say that a little different, if we are willing to partner with our suppliers and our supplier is willing to partner back with us and let us look into their factory, can we, with the help of our partner, take and build a digital factory of their factory? So this is the Zero RFQ kind of idea where on my aPriori system within AGCO, I have a user that can run supplier A as one of the digital factories and get a quoted cost. If we can reach this level of real-time data detail with a supplier, then the quote time locally to send out the the quote, the quote time at the vendor to receive the quote, act on it and send it back.

All of that time can be put into either other quotes on the vendor side or into other work on our side within AGCO. It’s a difficult thing. So the next step would be to identify some key vendors and build a partnership with them, ’cause really what you’re wanting to do is almost a joint venture, if you will, where you’re looking to really build a partnership where it says, we trust you and you trust us and together we are going to make sure that each one of us has got a strong position going into the future. It’ll take us showing aPriori to them, getting them to understand the process, building a model of their factory, and then working through the details, a small team to a little bigger team to a little bigger team until we double up and get to a good size. So that is my process that we do today and what we’re looking at going forward, and I think now we’re ready to bring up the rest of the panel.

How Can Digital Twin Technology Impact a Strong Supplier Relationship?

Speaker 1: All right. Now to bring up the rest of the panel, so we have Sam Freesmeyer, former VP of Engineering at AGCO Corporation, and now consultant and aPriori Executive Advisory Board member, and John Pilla, former CTO and VP of Engineering at Spirit AeroSystems and now consultant and aPriori EAB member. So welcome panelists. So the last presentation that we took a look at, Jodi made some really great points about being present and building trust with your strategic suppliers, so I’m curious to hear your take on creating those strategic supplier relationships, and does having a digital twin help to support building those relationships and have that big impact?

John Pilla: So I just came from a supplier discussion a lot along the same lines, but one of the keys is really that your suppliers have to trust you. So what I think has to happen is you go to your few biggest suppliers first and you start talking to them in terms of shared destinies and how you can build factories together, a digital factory of their place if they’ll trust you. But the whole point is that you’ll all come out with better profits when it’s all said and done, and that is not easy. There might be some mistrust based in the past, it might be some bad blood, you never know, but these things have ways of… That can be worked out if you can work with them, show them that you’re really interested in reducing their cost so that you can share in the increased profit, not just trying to put the screws down and get a little bit more money out of this negotiation.

Sam Freesmeyer: Yeah, it’s a pretty fundamental paradigm shift, if you think about starting from a place of distrust and basically an adversarial relationship almost, or an arm’s distance relationship, and it comes from a lack of a common data set.

Common information on what it is you’re trying to accomplish. And so you hear a lot of talk about decision-making and partnering, how do you get both you and the supplier on the same side of the table looking at the problem instead of looking at each other and trying to figure out how to extract a little bit more out of the other guy. So there’s a sort of natural movement to more partnering, but it requires that base of common information, and in the absence of that, what you do is, as a purchasing agent, you’re gonna send out four or five RFQs, you’re gonna look for the responses, you’re gonna wait for that to happen, you’re gonna compare all that data that comes back, you’re gonna pick the best package, you’re gonna move forward, because that’s the only information you’ve got is what five guys can do compared to each other.

But what if you had an objective standard that you used instead of just five guys competing with each other, and you dig down into what can be done with the should-cost analysis as your starting point. If you agree on the processes you’re gonna use to build Stan’s part, this is the most optimal way to build this part, three or four, five guys all have the machines to do that, they should all come up with a similar number. Well, how about you save all that time you spent producing RFQs and waiting for RFQs and evaluating RFQs because it’s expensive to both you and your suppliers. I advise a job shop in Iowa that does exactly this kind of business, small machine shop, and president of the company responds to every quote and he gets one out of five and that’s not adding value to his business.

And if he could figure out automation, and a way to work with his customers such that they just send in an order based on a common standard, that would save him a lot of money and certainly gonna save us a lot of money in dealing with suppliers. So there’s that transformation that has to happen.

SK: So if I could take that into the design world use case, as long as we’re taking and building our digital thread, the supplier that you’re partnering with has got some sort of skill set that makes them a key supplier for you. And whether it’s a particular aspect of their machine set or their process optimization or something like that, whatever it is that makes them good enough to be part of your key suppliers is something that you can then leverage on the design side, and so you can work your designs to leverage their skills, which then further increases both the productivity and the benefit to both sides.

SF: So then the question becomes, how do you start to break down that barrier, how do you first reach out to a supplier and say, “Hey, we wanna do this.” We’re all from the engineering side, so that’s purchasing’s problem, right? [laughter] No, but if you think about what purchasing is trying to do, our partners in procurement are looking for how to build relationships, they’re looking how to take wasted time out of the things they do, disruptions, and out of the things their suppliers do, so they’re gonna have an idea of who are the people that have capabilities from the previous parts they sourced. What they might not have is a good understanding of exactly what’s needed on this new part we just designed, and this is where a should-cost analysis gets real useful.

SF: And I think in one of Stan’s slide, he referenced the idea of sharing some of the information that we already have, so if we go to a supplier and say, “Okay, right, let’s forecast, let’s talk about the cost of these parts you’re making for us, here’s a should cost that shows the processes how they should work, what are your profit margins? What are your speeds and feeds, and all the other information you need,” you don’t have to have his digital factory set up to show him a should cost and say, “This is what we’ve come up with, how is your process different?” And start to create a motivation to say, “In the future, when I bring you parts that I want to source, I’m always going to come with this supply chain digital twin information, here’s an opportunity for you to save yourself some effort by just building your own digital factory and we’ll run off of yours instead of off of mine.”

JP: And that should include the detail inspections needed, then the finishing is needed, because often those add costs and we kinda forget about them after machining a apart.

SF: Yeah, and I think that’s really important, because now you’ve changed the focus. Instead of how can I get more money out of him or he get more money out of me, it’s a question of, “Here’s a set of operations that one of us is gonna have to do, do we need all of those operations? Can we figure out a design that doesn’t need maybe all of those operations, or can I find a lower cost way to do those operations, or maybe there’s a different set of operations that will lead to a better result.” It changes the nature of the discussion from adversarial to collaborative.

JP: You might say we’ve called out shot peen on here with this kind of coverage, and the guy making the part says, “Hey, can I flat peen that instead? It’s faster and I already have capability,” and then you’re engineer goes, “Well, sure. I just didn’t know that.” I mean, these things sound simple, but they’re in the details.

SF: Or I’m going for a surface finish for appearance reasons that I don’t really need for the function of the part. There’s a thousand design decisions like that that the engineer might be able to recognize and say, “Gosh, that’s kind of an expensive operation, why are we really doing that? Maybe we don’t actually need that.” Or the supplier with a greater knowledge of having made hundreds of similar parts, will have an opportunity to provide that feedback ’cause it’s not just quote the bid, it’s look at these operations and what can we do more efficiently?

Zero RFQ and Supply Chain Management Strategies

S1: Fantastic. Yeah, so you kind of started to talk about my next question here, the past of Zero RFQ is a long path sometimes, and there are many short-term strategies and then some long-term strategies, and you started to talk about sharing the digital factory and being open and transparent. Can you go through some additional short-term strategies that you’ve used in the past and some of those long-term strategies that you would tackle next?

SF: Well, I think one of the short-term strategies, and we already started talking about this a little bit, is how do you develop the trust and remove bottlenecks? And so we’ve done some NPI projects where we bring cost into the NPI cycle, and then you’re going all the way back to what is it the customer needs? What is it that new product introduction? What is it the customer needs? The fact that I’ve always done it this way, can I design something that’s different that’s gonna give me a real breakthrough in the cost or in the processing? Is there a new material I can use? A lot of times I like to bring suppliers into those NPI discussions. Now, these may or may not be the final suppliers to the part, but it will be suppliers that you’re trying to build a relationship with, and so you bringing them into a couple of these design cycles, you’ll let them contribute, you might even find that they’ve got ideas that make you want to go to the suppliers more often, which is their motivation to be there in the first place. So I kinda like those brainstorming sessions as a place to start bringing suppliers in even if we’re not really ready to move to that phase yet.

SK: I think that kind of collaboration and automation either across departments or between the vendor and the company with the design, brings in other ideas, different concepts, everyone else’s experience and becomes part of the story of that part. And you can then take and leverage that extra expertise.

SF: I think it also lets you start to look strategically at make buy decisions, because you start to look at the processes that you’re good at doing in-house, there are processes that you’re not good at doing or processes that you don’t really have the right machinery to do well, so you’re using a less optimal approach to getting that part made. As you start thinking in terms of processes instead of parts, it lets you make better choices out of which parts you wanna in-source versus part you wanna outsource, and so you can start to address your make buy decisions more strategically. I’ve got a supplier who does a really good job on this type of parts, I’m actually gonna move a lot more of my parts out to this supplier so that I free up capacity to work on stuff that I haven’t found a good supplier for yet.

How Digital Twin Technology Can Help Fill the Procurement Expertise Gap

S1: Alright. So last year, we actually did a study in partnership with industry around manufacturer and supplier relationships, specifically within procurement. And of the many challenges that we uncovered contributing to delays with the sourcing and procurement process, we found that both manufacturers and suppliers, 66% have trouble finding that specific expertise internally and 63% lacking the resources. So what tactics would you recommend manufacturers or suppliers to do internally first to help set themselves up for success, and do you have any examples of tactics that have worked in the past?

SF: Yeah, I’ll take once, I think that’s pretty broad questions. There’s a lot of ways to look at that. One of the things that is missing in a lot of our design engineering facilities is the manufacturing input, what are the processes we’re actually gonna use to make this part? And we hire a lot of engineers straight out of school, we put them to work and until they walk out on the shop floor, they’ve never seen a shop floor, they don’t know what these various processes are and how we use them and what type of characteristics they’re gonna give to the part. They’re in a virtual world of 3D and they need that input from manufacturing. The problem is, it’s a problem of numbers. At least in the organizations I’ve been involved with, there’s probably one manufacturing engineer for every 50-100 design engineers, so they can’t sit by everybody’s desk. They can provide some input when it’s needed, when it’s needed is usually three months after the part was released. So what happens is you get input long after you can effectively use it, so the ability to help guide the design process with a body of knowledge that’s immediately available to each engineer.

SF: Digital twin technology gives you the ability to replace some of that experienced skilled labor with a tool. Not all of it, you’re still gonna need those guys doing that work and fine tuning the digital factory definition so that it really reflects what it is you’re doing. But that is a way to bring some of that manufacturing expertise to the floor or to the design floor.

Build Your Data Analytics Now for Artificial Intelligence in the Future

JP: To take that a little further, I’m excited about the possibility of artificial intelligence. If you look at the aPriori suites today, there’s already some in the design for manufacturing, right? It says, hey, change this radius. You’ll save a cutter pass. You’ll save so much time, so much money. But the true artificial intelligence application, it eventually will be, you know, one part will get designed and then the artificial intelligence will design seven or eight that are a lot like it but just different enough that you don’t need seven or eight engineers doing ’em. Not that we don’t want the engineering jobs, but what we heard Peter Zeihan say this morning was that, guys like me, we’re retiring and the young guys like you, you know, there’s not enough of you. So we’re gonna need some artificial intelligence to fill that gap.

SK: Well, and the other thing was that that intelligence that’s built into the aPriori product for the design for manufacturing, scores, that is tunable, right? So you could have an out of the box design for manufacturing for a general factory, but when it comes to your factory or your vendor’s factory, if you have a vendor that has a capability to drill small holes very deep, well, their answer to the DFM score can be different for that kind of a feature than for another factory. And so that kind of custom ability, the ability to make the digital factory really exactly match what you’ve got on the shop floor and what your capabilities are, whether it’s internal or external, is it’s worth a lot.

SF: That actually raises a point that I think we need to keep in mind is if there were one best answer for how to do engineering, we’d be all done. We could go home. That’s not the way engineering works. And you know, as I look at it in the AgTech space in particular, you look at the major developers of agricultural equipment, John Deere, CNH, AGCO, Kubota, these guys that are the big global companies, what are they focused on as far as technology? They’re actually looking to startups for a lot of the technology they’re bringing in-house right now. Well, why is that? Is it because they don’t have engineers or is it because they’re not smart enough? Or is it because they’re maybe too smart? You know, this is how we’ve always done things. We’ve institutionalized it so much that, you know, what happens in a lot of these startups is the stupid ideas.

SF: You know, that’ll never work, until it does, right? And so we can’t… I mean, that’s what innovation is. It’s somebody, you know, too lazy to do things the way it’s always been done, right? So they go out and they invent a new way of doing it and it’s fantastic. Well, this innovation in the manufacturing space is the reason that a tool is not the only answer. A tool helps you institutionalize the body of information so that you can recognize an innovation when it comes along. Supplier comes to me and says, “Yeah, I got a great new way of doing this product for you. It’d be 15% below your lowest cost.” What’s my first assumption? He’s sandbagging. He shows me the processes that I hadn’t thought about using. He’s brilliant. So, one of the reasons to kind of institutionalize this whole issue of understanding the cost of a process well is so that you can recognize a useful innovation when it comes.

How Digital Transformation Initiatives Positively Impacts Your End-to-End Supply Chain

SK: Well, in addition to that, if we think of aPriori as a simulation tool, now I can start to ask what if? What if this, what if that? So I had run a little exercise a couple of years ago with aPriori, we had some senior management in engineering that wanted to look at slot and tab assembly for weldments as opposed to tooling. So what does that do to the overall cost? Well, it turns out if you’re cutting slots and tabs along the edge and in the middle of parts, you’re using a lot more laser time. So I’ve got an increase of 20% on my laser time. Do I have capacity for that? I don’t necessarily have tooling costs that goes into that, but that brings up the other question is, okay, so if my lasers are busy enough right now that I don’t have enough capacity, well then I need to plan for that in order to put that kind of thing.

SK: And so I can now start to do what if scenarios and figure out capacity and that kind of thing. And our manufacturing engineers do quite a bit of that, especially when you’re looking at bringing in new equipment or something like that. But it also works on the design side. Do I cast it? Do I weld it? Do I, I mean these are all the standard things that aPriori has taught you about how the software works. But if you look at that from an analysis standpoint, then you get to ask some more interesting questions yourself.

S1: Yeah, absolutely. And that kind of goes back to the creating that optionality. So using those digital twin insights to really unlock more information. You alluded to that with evaluating if I should go with a new machine or looking at capacity.

SF: Well, and going back to Stan’s example of, you know, if I wanna increase the usage of my laser for tab and slot design, what can I outsource for the next six months to make space for this little experiment? You can have a lot of these thought process out front. You don’t necessarily have to find a supplier with the same tools. You’ve got the option to say, can I find a supplier to offload some of the other stuff that I would’ve done with this tool, but I could use another process to create? Or they’ve just got spare capacity.

S1: And if you have their digital factory, you can look and see how much spare capacity they would have to take on a new project as well. Great. So what are some key performance indicators that manufacturers should be tracking to help mitigate risks with their suppliers?

SF: Well, I like lead time, and fulfillment of lead time commitment. So there’s two different parameters there. Is it six weeks or eight weeks? And if I said it was six weeks, was it six weeks or eight weeks? So it’s two different parameters. There’s one, what’s the best lead time I can get? And secondly, how reliable is that? ‘Cause oftentimes, you know, some of the bigger costs we incur are not the cost of the parts but the cost of the parts being late, and then we’re having to start expediting. So that’s one KPI that’s pretty important to me at least on the prototype side, on the production side, again, you know, purchasing worries about that. But again, it’s the same problem, right? It’s the same problem. How do we make sure that the parts are coming on the best lead time and that that lead time is reliably kept?