The hidden complexities of consumer products
Consumer products are leading a trend in how the process industries approach their products. Continuous production, or even large-scale batches aren’t meeting the evolving consumer preferences. We talk with Alastair Orchard about what they are doing to change this dynamic.
Nick Finberg
00:09
We are back for another episode of the Future Ready podcast from Siemens. I have Alistair Orchard here to finish our discussion on what industry might look like in the near future with a shift towards greater software control on the shop floor. Um, if you haven’t listened already, you can check out part one linked in the description. But if you’re ready, Alistair, I think we might jump in. There are a couple luxury vehicle companies looking at software to find for their highly customized product portfolios that I know of. Um, but you interact with even more industry segments than I do. What are you seeing elsewhere?
Alastair Orchard
00:45
Yeah, I actually find CPG companies somehow fascinating. They have these they have these vast portfolios and if you’ve ever seen some of those maps of the portfolios, you as a consumer you only you see a small portion of this, you actually don’t know how big those networks extend, but it kind of it leads you to this idea that they’re highly agile companies. But if you zoom in on how they actually operate, then Each of those portfolio elements is actually kind of just made to stock on some dedicated fixed hardware-based line in a in a central location. And what they do as a company is just multiply that. a thousand times and you end up producing at scale. But that model really does have negative consequences, you know. So, I maybe I can go through a few of those that come to mind.
Nick Finberg
01:40
Yeah, that’d be awesome.
Alastair Orchard
01:42
So, we already mentioned the fact that it takes just far too long to introduce a new product to into an environment like that. And it’s the hardware that is at least partially responsible. Because if you have a new idea that requires a new process, then the time to market includes the lead time on building a new line, sourcing the machines, commissioning the line. And that’s usually 18 to 24 months So if you’re if you’re stuck with that kind of lead time then you’re not an agile organization. Then if the product launch isn’t successful, then you’re kind of stuck with this capital equipment debt. You’ve spent 24 months building it and now it’s stuck there. You’re not using it You can’t fund the depreciation of those assets, and you can’t re repurpose them because they’re all tied together in an extremely rigid way. Then and this is one of my favorites, and it takes a little bit of kind of imagination here, but because the production line is in some physical location and it’s a fixed location, on average it’s always going to be far away from consumer demand. Okay? Which means that when I make that product there you know I’m forced to ship what in CPG is almost all water vast distances again on average to a point of need. And that’s incredibly inefficient. And actually, be because lead times are long, what these companies end up doing is creating very large batch sizes to kind of backfill the supply chain so that they never risk running out of that product in you know on the on the in the stores um
Nick Finberg
03:30
They also but that in itself is a risk because then you have stock that you possibly can’t get rid of with your partners.
Alastair Orchard
03:37
Yeah, this is called bound inventory, which has a massive cost. I won’t say which CPG companies, but we’ve worked with some CBG companies and actually 40% of the entire supply chain cost. And that that includes raw materials, the energy use for production. the manpower within manufacturing and the depreciation of the entire capital equipment for production. 40% of that entire supply chain cost is in transportation a warehousing, just the backfill those things. And then yeah.
Nick Finberg
04:15
Oh wow
Alastair Orchard
04:16
They also operate on this concept of demand planning, which is essentially a guess about where product is going to be required. And these are extremely sophisticated systems, but the consumers are fickle, right? And it doesn’t take so accuracy is usually between about 50 and 60 percent which as you said means that on average it’s kind of a flip of a coin. So, there’s often problems with oversupply in places, undersupply in others. Yeah, it’s kind of terrible. And the other thing to bear in mind is resilience. This is really important. Because lines are so specialized Even if I have factories everywhere, I’m really only able to make a certain product in a certain place. And that means that if that line goes down, if there’s an outage then it can trigger real sort shortages of the of that product on shelves in supermarkets. And actually, there’s a kind of converse view which is If your if demand for a particular product decline, then again, I can’t repurpose the line. So, I end up with this underutilized equipment. that I’d ideally like to use to cover excess demand in other areas, but I can’t because it’s tied into a specialized line. Yeah.
Nick Finberg
05:43
Wow, that that scale is just so alien to me. I’m so used to talking about production on the scale of millions of units, but consumer products like you’re talking about are multiple orders of magnitude greater. What kind of scale actually is it that we’re trying to talk about with regards to flexibility? or seasonal pay changes or personalized formulations.
Alastair Orchard
06:05
Sure. Well, we can go into a couple of use cases here. The first thing I’d say is that Not everything needs to be personalized, right? So, we think of this in terms of supply chain optionality. If I have a very stable product that I manufacture at scale with low margins, and it’s like my It’s my cash cow, then I’m not going to shift to a radically flexible model anytime soon. But yeah, but any kind of business where that’s that is seasonal or where demand is peaky or where personalization might be interesting then the old model is a broken model. but we get we can look at this maybe in maybe at three scales. So, you mentioned seasonality and personalization. I’ll come to them in a second, but there’s one area which is actually a bigger problem. And it’s one of the one of those areas or one of those use cases which most people aren’t aware of. And this is raw material substitution. So, Let’s say I’m a CPG company, I’m using this a raw material, maybe it’s a pretty common one, it’s pervasive across all of those thousand products that I make. Very often these are commodities, they’re also natural products, and so supply is not always fixed or reliable. So, you know, what happens if it doubles in price? And that’s not impossible by any means. Or let’s say the EU bans it, right? That happens. Or suddenly a raw materials kind of vilified on social media. Okay, it sounds like an annoyance. They have to update their recipes. They have to replace the inventories with this new with this new raw material. Maybe pricing has to be tweaked, nutritional labels have to be changed. Obviously, software can help in all of these cases, but actually the real pain is in manufacturing because the effect on manufacturing of even a small change to a raw material can be totally devastating because again, it’s where the rubber hits the row, let’s say. So, let’s say we have to swap out something like palm oil. This happened not so long ago, not too many years ago. And so, we choose something else, coconut oil. It’s That’s nice. Well, if we do that, it actually changes the physical properties of the product. So, I don’t know if you ever bought coconut oil. like at room temperature it it’s hard, right?
Nick Finberg
08:46
Yeah, like this the solidified state is a lot different between the two. Like you have to keep it warm. for mixing processes, I would assume.
Alastair Orchard
08:55
Exactly. Depending on the temperature, viscosity is different. So, it can be harder to pump and to mix. As you mentioned, this the melting points higher so you have to adjust the temperature you know the processing temperatures and the way that you handle those materials maybe it even comes in a different form of container And so that changes the way that you handle materials, how you store these materials. Maybe it even requires some other process step that you didn’t have before, some preconditioning step. And the way that this is handled traditionally is that these companies manually re-engineer every line in order to adapt that hardware, that hardware to these changes. Okay. And but it gets worse than that because there’s no uniformity of equipment throughout these large companies. They’ve been built up over decades, these companies. They’re quite often the result of multiple acquisitions. And so, I don’t you said mixer, the mixer in factory one, five, nineteen, and hundred and forty-six can actually be f physically very different, different sizes, different impellers, different shapes, different materials. And that actually that means that the companies have to individually tune the parameters for things like mixing speed and duration for each of these product or raw material substitutions and it has to be done by trial and error at each site. Now kind of do that a thousand times a year, which is actually very typical, and you’re in this never-ending battle. of trying to modify inflexible hardware systems. And that’s just as that’s to just to stay s in the same place. Without going into too much detail, but as you can imagine, software really helps. And so, we’ve we bake these software-based techniques into our digital threads, our enterprise recipe management and adaptive manufacturing digital threads so that we can automate that entire process. So, a change in raw material could be propagated automatically and instantly throughout the enterprise. So that’s really cool. I know I spent yeah; I spent a long time on that one.
Nick Finberg
11:23
Oh, that is super cool.
Alastair Orchard
11:24
maybe I’ll go quicker on the seasonal changes. Seasonal changes are really You can think of that as just planned disruptions to the portfolio. So, add cinnamon or something at Christmas, that’s just like a material substitution. And but they play out in the same way. They’re just as devastating and impactful, even if you plan for it. you have to go through all of those same changes it’s just as complex to handle. And so, software helps in the same way. Packaging is maybe a nuance. So, if I if I want to put Father Christmas on the label of a can of shaving cream or something, that’s not a big deal. But actually, if I want to if I want to offer the market these Christmas packs, his and her shaving kits or whatever, then There’s no software-based adaptability traditionally in these factories and so they have to set up manual repackaging of every one of those billions of products. So yeah. So seasonalities, just like raw material substitution, it creates havoc. Software can really change the game here. And you’d have thought that personalization just took this to an extreme. It’s just a kind of logical conclusion of everything that we’ve done so far and talked about so far. and that’s true, but it does have it does have some consequences. As I scale down to lot size one and I’m manufacturing something personally for you then one of the first things is I am I’m having to make a batch which is much smaller than I’m used to. And here’s where I may have to start rethinking equipment. Because if I have a mixer, a 2,000-liter mixer or something, two-ton mixer, then I just can’t mix your favorite shampoo individually there. So, I actually I actually have to start rethinking physical equipment sizing. And actually, if I’m making multiple personalized batches, then the changeovers which in a traditional fact facility are minimized because you have to basically shut the line down, sanitize everything between these changeovers. If I’m doing that for every bottle of shampoo, then I’m in real trouble. So, what we actually see is a shift to pipeless production. We’re removing some of these hardware constraints and this allows us to drive production in a fully adaptive way. It kind of It sounds scary, I guess, but these again are techniques that we’ve pioneered together with CPG companies and they’re actually available in our in this case our Opcenter platform. And then once these once the factories in the in these companies’ production networks are adaptive then these companies can really radically rethink their business models, the way their supply chains are managed. So instead of making stock in that in the wrong place essentially and then shipping it to where you hope demand will be in three months, which is what we discussed before, you can actually produce the products that the customer needs where they need it, when they need it, in the quantity they need it. So, this is on-demand manufacturing and that help helps actually avoid
Nick Finberg
14:44
Carbon emissions it negates the shipping all that water that you were talking about earlier.
Alastair Orchard
14:57
hilarious conversations with customers who were trying to offset their carbon emissions, right? And that basically meant generating a whole bunch of carbon and then planting forests to try and absorb it and it and it just seems the wrong-headed approach. If you can avoid carbon by many in the first place, then that’s much better. Nowadays things like tariffs are maybe the new sustainability. So, if I’m making stuff in the wrong place and I’m having to cross a tariff border in order to deliver the product to the customer, then this massively affects the economics of my business. And so again, if I can manufacture beyond that tariff wall, then I’m negating the if the impact of those tariffs. secure supply chain is a is a big one. So, if I’m relying on very long, very complex and fragile supply chains that can be affected by you know a canal blocking or something on the other side of the world, then I don’t have a secure supply chain. And so, it’s all about reshoring manufacturing and producing on demand, producing at that point personalized products, and actually, this as you can probably hear I’m quite passionate about this because I’m actually in the middle of a new project that I’m kind of personally doing, and this is a distributed manufacturing platform. that networks and orchestrates this kind of network of adaptable factories.
Nick Finberg
16:25
Okay. Well, it’s definitely clear we could talk about this for a long time, and I think I can wrap my head around it maybe. but we should bring this back to what’s happening today and what businesses can start to implement in the next few years as we move towards that that vision of customizability or achievement changeovers like you were talking about. What are some of the technologies, workflows, or mindset shifts that our audience should understand to enable that flexibility that we’ve been talking about?
Alastair Orchard
16:56
sure. Well, I think come companies have been doing this for a while now. I’ve been working with companies to digitize their processes for more than a decade. So, they’re Most of them are already on the on that journey. It’s actually been that’s turning that substrate from analog or hardware into digital, as we said. It’s not always actually been easy to articulate the value of that. Because it’s actually been in preparation for this moment. Back to human evolution. It’s like the biological effort required to kind of evolve the human brain. But now we have this software-defined world that’s programmable. And if you’ve I don’t know if you’ve experienced how AI can write code. It’s shocking. So, if you’ve tried it, then you’ll appreciate the kind of power that having a programmable enterprise kind of unlocks. Anyway, hopefully. What we’ve talked about will kind of help listeners open their minds to all this stuff, to the power of this kind of software-defined future. But there’s nothing like seeing it in action. So, one of the jobs I do is a kind of digital transformation or software-defined x evangelism. so if any of the any of the listeners want to reach out, they can do that directly to me or through you or to the local Siemens account team. And We can kind of get together.
Nick Finberg
18:27
Yeah, we’ll have we’ll have a link in the episode description for the audience.
Alastair Orchard
18:30
Perfect, perfect. But maybe you asked for some ABCs or basics.
Nick Finberg
18:35
Yeah.
Alastair Orchard
18:36
So, all right, well maybe There’s a lot more self-service stuff nowadays. That’s one of the benefits of AI and software to find is that you can do it yourself, but especially the beginning, find a trusted digitation partner. and start thinking about modeling these digital twins of product and production facilities. This is kind of key, as we said, a shift left into the virtual world. It’s just a better place to do discovery and design and optimization, all that stuff. Better that certainly better than scratching around in in the dirt like animals. You know, simulation, it’s predictive. Which means it’s like a superpower, and then you have AI that can accelerate anything that’s software based. So that’s a kind of good starting point. philosophically, just be on the lookout for these hardware defined or rigid kind of constraints. Try to use software to decouple them wherever you can in your factories, in your business processes, because the more degrees of freedom that you have the more options you have. Yeah. And then if you start to deploy AI agents, they need to be able to pull levers. And if the only lever they can pull is the press the button of a giant black box that’s all hard coded, they’re not going to be effective. Okay? So, you we need to be able to deploy these agents that then have the freedom to manipulate software-based systems in the digital thread and within factories. Talking about digital thread, that’s probably another critical thing. A software-defined factory is only as flexible as the data that flows into it from the design house. So set up an innovation digital thread. That’s going to help you automate and accelerate the introduction of new ideas, um, modifications to the to the bomb and have them flow through into production. Okay, so that’s again part of this software-defined philosophy. Then this may be a question that we get asked quite a lot. You know, I’m not a startup, let’s say I’m a manufacturer, unless I’m a Unless you’re a startup, you’re not going to be working only with new equipment. So, you’re going to have brownfield legacy equipment and you need a brownfield strategy. And again, software can really help there. You can layer what we call a data fabric across the machines in a in a facility This wraps them in a modern software-defined layer. It it’s a semantic layer actually that not only provides data access to these legacy machines. But it actually allows AI to reason on the semantics and behavior and statuses and characteristics of these machines. So, these are pretty good places to start, I’d say.
Nick Finberg
21:45
Okay, awesome. Thank you, Alistair. It’s been it’s been great talking with you. And there’s definitely more I want to have you back to talk about. Um, but at this high level that we’ve been talking about today, is there a final idea that you want to leave the audience with.
Alastair Orchard
21:58
Well, I enjoy I really enjoy myself. thank you. final thought. I would just say Now’s not the time to think conservatively. The value and the power of this software-defined approach is basically unlimited. I was I’m here in Chicago. I was actually at Google yesterday and we’re working on a moonshot project with X team and that blew even my mind, okay? and it would and it was the same old stuff. It was software-defined and it was AI driven. What’s coming is really going to change everything. And it’s premised on this idea that we’ve been discussing today, the fact that everything’s programmable. So, I know it’s a well-worn adage, but the advice to think big, start small and scale fast has never been more appropriate.
Nick Finberg
22:51
And thanks to the audience for tuning in for another episode of the Future Ready Podcast. We’ll be back soon to bring you even more interesting guests from around Siemens to talk about topics spanning software to find. artificial intelligence, the comprehensive digital twin, and so much more. Make sure to subscribe so you don’t miss it, and you can find resources in the description to keep the learning going today. See you next time
