The future of decentralized manufacturing – Podcast Transcript
Following the industrial revolution and the rise of globalization, mass produced and consumer packaged goods (CPG) have been manufactured on the concept of centralization; a factory in a central location produces a product before shipping it to warehouses the world over. While this approach has many benefits, in recent years it has become increasingly clear it might not always be the best way.
In a recent podcast, host Conor Peick is joined by Alastair Orchard to explore the future of CPG manufacturing from shifting left to networks of local production facilities before taking this concept right to the edge with a pop-up factory that can be shipped in a container.
Check out the full episode here or keep reading for a transcript of that conversation.
Conor Peick:
So welcome into the Future Ready Podcast. My name is Conor Peick. I’ll be moderating our discussion today and I’m joined by Alastair Orchard, who is the VP of Digital Enterprise at Siemens. And yeah, we’re here to talk about the CPG pop-up factory concept, adaptive manufacturing, and Lots of exciting topics. So Alastair, thank you so much for joining me today. Really excited to have you in.
Alastair Orchard:
It’s a pleasure.
Conor Peick:
Yeah. So just to start off our conversation, I would love to get your thoughts on how you see manufacturing evolving in such a way that production is no longer tied to a single location. Maybe that it can be a little bit more dynamic and more adaptive, especially around the globe.
Alastair Orchard:
Sure. You posed the question as though it was purely future oriented. But actually, I think this has been this started some time ago.
Conor Peick:
Okay.
Alastair Orchard:
Remember COVID?
Conor Peick:
Yes.
Alastair Orchard:
So it was during COVID where we first saw supply chains kind of disrupted and interrupted. We saw Western manufacturers with their facilities started repurposing lines. If you remember, they stopped making car parts and started making PPE.
Conor Peick:
Yeah, right.
Alastair Orchard:
And everyone started kind of thinking, the next one was sustainability. And people were kind of saying, maybe we are making things in the wrong place and just kind of shipping them across oceans. Maybe we could cut carbon cost and complexity permanently if we made things where they’re actually needed. And then much more recently, we’ve got tariffs. And physical stuff is is taxed when it goes across a tariff border, but at least up until now, data isn’t. And so we can kind of use the Siemens a digital thread concept to shift left. Do you know what that shift left concept means?
Conor Peick:
Yeah, My understanding is it’s moving, at least in the engineering context, moving some decisions earlier in the life cycle. You use simulation to test and validate.
Alastair Orchard:
Exactly. And so in a manufacturing context, you instead of sending a kind of product design down the digital thread and saying the factory is going to work out how to make it because that factory is now going to make a million of those over the next 10 years. If we’re doing If we’re doing distributed manufacturing, then it’s almost always lot size 1. So you can’t afford to get it wrong 100 times. It has to be right first time. So you shift the decision about manufacturing upstream, just like you said. And the designer of the product actually uses those simulations to design out the manufacturing process and create a kind of blueprint for manufacturing. And that can then be beamed anywhere behind tariff barriers, for example. Yeah, so I think it’s already started. And of course, I think it’s accelerating. You had Pepsi in here, right? Yeah. So the main reason I’m at the fair is because we’ve been collaborating with them on a big showcase. And they’re quite vocal about the fact that they see this as maybe the next big frontier for them.
Conor Peick:
Right, That ability to respond to demand as it maybe changes around the globe seems really, I mean, potentially a really powerful thing for a CPG company. So I mean, I guess maybe we can just talk a little bit about how unlocking that ability to produce where the demand is at, maybe particularly where there’s potential for a spike in demand due to some big event or whatever it may be. How does that maybe change the way that a consumer goods company, that they think about things like forecasting for demand, making sure they have enough stock on hand, product availability and being responsive to the market in that way?
Alastair Orchard:
I think it will change how they’re thinking, but it’s not going to happen overnight. I also don’t think it’s going to be a complete shift. So lots of CPG are kind of high volume, low variability products that are pretty easy to predict their demand profile on. And I would say that they’d continue to be working, manufactured pretty traditionally for the foreseeable future, maybe indefinitely. But like you said, anything personalizable, anything ultra spiky, so highly localized, anything seasonal, anything with a very short shelf life or market life. So those are really good candidates for on demand production. I think making just in time instead of just in case, it really avoids all those shipping and warehousing costs. for something that is usually just almost all water. It’s 90% plus water. Some of our CPG customers have told us that they spend up to 40% of their cost of goods sold on shipping and warehousing, which is insane. Yeah, So the other thing you mentioned though is NPI, new product introduction. And that’s a really good point because these companies take a huge amount of time and effort in introducing a new product, like 18 months, and even just substituting a raw material can be hugely disruptive to the supply chain. Whereas if you’re starting to think about a distributed network of these highly adaptable factories, then actually every order is a new product introduction. And what that means is you can turn your kind of value chain into a weapon. You can weaponize it, you can make it a competitive advantage. So they’re really thinking about this.
Conor Peick:
Wow, it’s fascinating. And I mean, especially with a lot of these brands as they become or already are or become these sort of global brands, I mean, thinking about something like PepsiCo, for example, they have, you know, there’s different preferences and different tastes in different markets around the world. So is this, you know, you talked about this network of highly adaptable factories. Does that make it easier for them to meet demand in these different markets with taking advantage of the adaptability that they have.
Alastair Orchard:
It certainly does. And actually, that’s the use case that we’re doing here with Pepsi. We took an example, which is the sports sporting competition that will take place in the US this year. But that will create a spike in demand for Gatorade is what we’re talking about here. And that’s worth talking a little bit about, I think. because clearly a company like Pepsi needs to, PepsiCo needs to meet that spike in demand. But the interesting characteristic of Gatorade is that it’s available in multiple colors. And because team jerseys are different colors, then depending on who wins which round and which teams will meet, that will have a kind of non-linear or non-uniform effect on which Gatorade colors are actually consumed. And it’s something you just can’t predict. You have no idea. Brazil are going to meet Portugal on the 15th of July in Seattle. You’ve no idea until the day before. And no one’s going to be drinking blue Gatorade under those circumstances. It’ll be this massive peak for red and for yellow. And so the main way to, or the idea of decentralized manufacturing is to place these adaptive production nodes around the country, actually one close to each one of the stadia. And then real-time demand signals, which in this case are the results of football matches, they feed into the system, which then adapts. Demand is then kind of allocated and distributed to the network and red and yellow Gatorade is made on demand. And so, I mean, it’s a kind of toy example, this one, but it can be then adapted for any kind of drinks or food, but even things like personalized medicine.
Conor Peick:
Yeah.
Alastair Orchard:
There’s, I don’t know if you heard of the concept of kind of basement manufacturing for for hospitals, why not produce personalized medicines on demand? It’s exactly the same on-site. Now, the IP comes from many potential pharmaceutical companies, but then they have to be personalized for individual patients. And these things have very short shelf lives, often they’re required almost immediately in order to save a life. And so, yeah.
Conor Peick:
And of course, you don’t need a huge amount of stock for a medicine that’s personalized potentially to one single patient. Exactly.
Alastair Orchard:
That’s also the case for CPG, again, because the biggest cost is shipping the water around. And so that’s localized. water is normalized in CPG, so you have to do something. You can’t just take it out the ground and you have to normalize. You either have to filter it or balance the electrolytes or whatever. Once you’ve done that, which can be done locally, then you’re good to go. And it’s just a question of some additives and flavorings and the special sauce is not difficult or expensive to move around.
Conor Peick:
Relative to the water, it’s quite a bit less. Yeah, interesting. So when we’re talking about these, portable factories, what do you think is sort of the minimum portable factory kit, you could say, talking software, the hardware within that portable factory? Yeah, what do you, what’s the minimum amount of stuff you need to make it portable, make it relocatable, but obviously able to meet the demand that might occur?
Alastair Orchard:
I think it’s smaller than you might imagine. Out on the shop floor, we have this pop-up concept. So it’s actually a container.
Conor Peick:
Oh, yeah.
Alastair Orchard:
And the entire factory fits into the container. And really, what would the smallest footprint be? Maybe, well, a single robot can weigh, dispense, so do the dosing of powders and liquids. It can also change its end effector and mix. It can fill. Maybe you need another, you need an AMR, so a kind of AGV with an arm on it to do the replenishment and maybe the outbound logistics. But that’s actually a tiny footprint. Obviously, the volume that something like that could do is pretty small as well, which is why that would represent a kind of extreme end node. So you put this in a point of sale, for example, you’d have something like that in a Walmart. And it would be manufacturing products from multiple companies based on either kind of personalization requests or just to fulfill empty shelves or whatever. Maybe something bigger would go into an Amazon fulfillment center or something, but really you’re either multiplying that single robot out or using a bigger robot. So it’s really quite small.
Conor Peick:
Interesting. And then some amount of software load in there as well, presumably to receive the data you talked about earlier.
Alastair Orchard:
Yeah. The main data that we talked about is the recipe. In discrete manufacturing, that’s called a bill of process. And let’s stick to CPG. Think about a cookbook, right? So you have, let’s say you have to make a chocolate cake, you have a list of ingredients. That’s the kind of bill of materials. But then you have the instructions. And the instructions essentially say, okay, take the first three ingredients, the dry ingredients and mix them together and then add the wet ingredients one at a time and then mix until fluffy and then preheat the oven. And these are the manufacturing operations that can be used to drive that set of hardware and equipment to produce product A, B, G, F. And because we don’t have a production line, the production line would be, is what really kills flexibility in a manufacturing facility. It’s the kind of old way of doing it. You take that recipe, you give the recipe to a machine builder or a line builder and say, either build me a machine to do this or build me a line to do this, fill it with pipes that rigidly connect the operations together. Now, that’s usually quite efficient. It runs quickly, but it can only make one thing. And we want our factory to run everything. So the recipe, in our case, remains in software. And we have a software execution engine. It’s part of our op center, Siemens op center. X portfolio. This one’s called BOPX, actually, which stands for Bill of Process Execution Engine. It’s an adaptive production orchestration engine. And I’d say it’s one of the most critical elements here because it ingests the order and the recipe. And then it does what’s called a real-time capability matching. So every operation, just like a human would do when cooking this chocolate cake, every operation in the recipe, it says, all right, what do I need to do? I need to dose. and it’ll then communicate with the equipment saying, look, can any of you help me with this operation? And the equipment’s autonomous and to a degree intelligent. We use standards like MTP, modular type package. They self-declare their capabilities and so they can kind of answer back saying, you know, I can help you out with that. And it’s this kind of real-time negotiation which gets the cake made or that gets the Gatorade blended and mixed and packaged.
Conor Peick:
And is the right color.
Alastair Orchard:
And in the right color. Then there’s much of the software remains outside of the factory, so on the cloud. But yeah, I guess we’ll talk maybe about some AI and other kinds of things later on. So there’s more than just that footprint. Obviously, each one of the pieces of equipment has to be then modular because they’re kind of standalone pieces of equipment that are separately automated, very often now with software-based automation. So software-based automation that can be updated on the fly. So we can send set points down, but we can also send all of the G-code, so the actual instructions for movements to the robot on the fly. So it’s called adaptable because we can really change the complete behavior of the cell in order to manufacture different products and even a mix of different products. So in a typical adaptive factory, what you’re not doing is making production runs. Like Gatorade is typically made in Atlanta. They’ll make red Gatorade for a week and then shut down. They’ll clean everything and then they’ll do green Gatorade for a week and then shut down. And they’ll do that irrespective of demand. And in our case, will be making red, green, blue, all simultaneously. And this is actually another reason why it’s called pipeless production. Pipes are not only rigid points of connection, but they also get dirty. So if you eliminate them, you don’t need to do cleaning. You just need to clean the container that contains the batch. So yeah.
Conor Peick:
Interesting, but unfortunately, we’ll have to end it there. I’ve been your host Conor Peick joined by Alastair Orchard on the Future Ready podcast. Tune in again next time as we delve into the future of agentic driven factories.
Siemens Digital Industries Software helps organizations of all sizes digitally transform using software, hardware and services from the Siemens Xcelerator business platform. Siemens’ software and the comprehensive digital twin enable companies to optimize their design, engineering and manufacturing processes to turn today’s ideas into the sustainable products of the future. From chips to entire systems, from product to process, across all industries. Siemens Digital Industries Software – Accelerating transformation.
