ADAM SELLS 0:10

Welcome to our podcast. We're here from Goodfellow. So yesterday our CEO, who I think is on the screen still, uh did a keynote speech about scaling. And about how scaling doesn't necessarily mean that you need to start again from scratch. So what we wanted to do today was delve in a little bit deeper to some of the topics that Simon was discussing yesterday. So first of all, um we'll do some introductions. So my name's Adam. I've been with Goodfellow for a little over 13 years now. Um and my responsibility at the moment for all of our custom um offers to customers, so custom projects. Benedict, if you'd like to introduce yourself.

DR. BENEDIKT MOSER 0:49

Yeah, I'm Benedict. I'm a material scientist by training and I'm a senior technical consultant uh at uh Swiss Technology Partners. That's the uh basically the Swiss competence centre from Goodfellow for uh materials consulting, testing and analysis.

THOMAS GREAVES 1:04

Thanks, Benedict. Tom? Thank you. So I'm Thomas Greaves. I've been with Goodfellow for about 15 months now. I'm the regional sales manager, so my responsibility covers Latin America, Europe, and the APAC region. Uh and I also work closely with Adam in our custom product range as well as our uh ongoing developments worldwide. And Arunema.

Dr. ARUNIMA NAIR 1:24

Hi, I'm Arunuma. I work as a technical medical specialist and Goodfellow. I have a background in medical science and engineering.

ADAM SELLS 1:30

Thanks, Arunema. Yeah, and what we wanted to do today was bring together a mixture of uh technical, as you can see, uh, and and commercial because when we're talking about scaling, um it's it's complex. There's often very complex technical systems and complex commercial things to consider. So we'll be talking about things like risk, so whether that's geopolitical risk, supply chain risk, risk in uh specification gaps, continuity of supply, um quality, traceability, and then the impact that sort of deglobalization is having um at the moment uh which which is uh seems to be accelerating as well in terms of the requirements that we're seeing from the market. So, Tom, what actually tends to break when uh we want to scale?

THOMAS GREAVES 2:16

So often when you're scaling and Goodfellow works closely with a a large number of startups, primary research, academia, as well as then industry primes taking developments forward, it's quite often um the specification that breaks rather than the material itself. So when developments are starting, you're working closely on a new um a new fundamental project, everyone's working closely on those um material developments, the spec is is evolving through time, but it's closely monitored by a certain select few people, and then as scale happens, progress, development, growth, then uh the unknown unknowns creep up where if a material that has certain properties at a small scale was insignificant, then at scale becomes a real challenge to those businesses. And trying to monitor that and correct that over time sometimes feels like you're starting again trying to then recapture what those specifications were. It's almost sector agnostic in some cases, whether you're working in fusion technology, semiconductors, um, or medical devices, you see the same trends but with different materials.

ADAM SELLS 3:23

Yeah, and I think um it depends on on the company as well, doesn't it, and their experience with with scaling. We we work with uh companies that have done this time and time again, and new startups that are kind of scaling for the first time. So I think to that point, um what's underestimated by companies, and I put this to the panel at Tom, I don't know if you had some some initial thoughts.

THOMAS GREAVES 3:45

So I think we've we've got a good example of um a company that we work with in the electronics sector, um, initially starting as a as a as a founder-led innovation. That one of the strengths of Goodfellow and its um acquisition businesses has been to always work closely with the customer relationships from the start. And this business had a revolutionary new idea and approach to semiconductors, and they tried a number of different businesses and were unable to achieve um the goals. They came to the Goodfellow team and we persevered with them and with that challenge, and over the last five, seven years that we've been developing, they're now progressing into a different level of scale, um, which then commands uh a number of different challenges from recruiting the right employees and technical staff to making sure that you've got continuity of supply of materials, especially when that is dedicated into one geographical location. It's how do you then, if there's an global issue related to those material sourcing, how do you pivot and keep the customers moving forwards?

ADAM SELLS 4:45

So, what do you think they potentially underestimated then, or what were the unknown unknowns, do you think, in in some of these examples?

THOMAS GREAVES 4:52

So I think it's definitely in terms of the time and also cash flow is a is a real challenge for startup businesses, making sure that not only the cash flow in terms of the development costs, but also then when the next round of funding or product cycle will be paid for. There's always an end customer, and you've got that journey to repay the the costs with the development. So often the cash flow is probably something that is very well underestimated within within this time frame.

ADAM SELLS 5:21

Yeah, I think and that's the complexity with all of this, right? You've got all of these different factors that you're trying to balance. I suppose taking it back to the the specification and the material, Benedict. I guess all of this commercial stuff doesn't matter if the specification is not right.

DR. BENEDIKT MOSER 5:39

That's absolutely correct. But uh of course the uh the failure of the specification is basically triggered commercially because uh as long as you're on a prototype level, you might buy your material from a trusted source, uh, you buy it from Sigma Aldrich, or you buy it from Goodfellow or whoever, uh and money doesn't count because the material cost is not very high at that level uh when you produce some prototypes. But then all of a sudden, if you go to scaling to mass production, uh purchasing comes in, uh you buy 20 tons of the material, and then money counts. And then you you seek other sources, you buy from different sources, and then you find out whether your specification is correct or not. And that's actually what you said at the beginning: the specification breaks, not the material itself. Uh, because then you find out that your specification does not really cover the requirements that you have, and that's the moment that you need to start really understanding your application. What do you actually need? Um, and then having the right specification uh helps you even buy material cheaper, but still uh that fulfills what you need.

ADAM SELLS 6:47

And and how does that link in with the actual analytical techniques, right? So it's one thing to set a specification, but then it's another thing to then pin down exactly how you're gonna measure that specification.

DR. BENEDIKT MOSER 6:58

Exactly. That's the that's the second point about uh then quality assurance. Once you have a specification that works, then you need to understand how do I now ensure that all the material meets actually that specification. Again, as long as you buy uh prototype level stuff, uh you probably don't need to test it, or you test it anyways because it goes through the lab basically. Uh once it goes through the production line, you need to make sure, okay, uh, or you need to have the decision whether you're relying on uh certificates that your supplier uh sends you, or if you're in a critical application like aerospace, space, deep sea, where failure is really not an option, um, which means you probably would uh ask an additional independent lab to actually uh retest what you're buying. So not rely on the outgoing specification of your supplier, but doing actually an incoming and specific specification of the material.

ADAM SELLS 7:55

What materially changes when we go from, as you said, a handful of samples to weekly production scheduling and things like that? What what are the what are the kind of key things that people should be considering?

DR. BENEDIKT MOSER 8:07

It's the uh the quality assurance that you need to set up. You need to set up a system of uh what you want to test, who is going to test that, uh, how often you're going to test, which is basically based on a on a risk analysis. And uh and of course it's based again on the specification. So the specification is really the central point that you have for your material. Um if that is not correct, you can do testing whatever you want, and you will anyway fail at one point. Yeah.

ADAM SELLS 8:38

So Simon's medical example yesterday gave an example of a specification gap, right? Where a product had been developed through RD, um, scaled, and then due to the natural variation within the manufacturing process, there was a la a drop of performance or a lack of performance. What can we do to try and anticipate these things that that are not uncovered by the RD process and the initial scaling to reduce the risk of these specification gaps appearing later on in the production process? Maybe even we see examples where it happens two, three years into the product being used.

DR. BENEDIKT MOSER 9:14

Yeah, it's it's a very typical example and it it can happen to anyone basically. Um I mean we had we had the same uh same issue years ago with an aluminium company, or actually it was a packaging company buying aluminium foil uh as a as a lid foil that later in the product actually was perforated. And uh like 20 years after uh after the product went to the market, they wanted to add a second source basically for the aluminium foil and realized that it doesn't work, and no one knew why. So that was really after 20 years, there was basically a specification gap discovered because it was not just the aluminium, the chemical composition that plays a role, chemical composition and uh and thickness of the foil. There were many other uh important properties that that first supplier just had in his foil, but no one knew, and the second supplier didn't, although the specification was uh was still fulfilled by the second supplier. So that basically can happen to anyone if you did not early enough start testing different sources because only that basically gives you the feeling what are the important properties. Um you cannot do that just on an uh on a piece of paper and think about the properties. You have to test it at one point.

ADAM SELLS 10:37

Would you say it's about bringing enough data together uh so you have a uh as wide a picture as possible? Arunema, sometimes the when we look at the root cause, it can be incredibly small changes in the composition, right, that you find uh contributing.

Dr. ARUNIMA NAIR 10:52

Yeah, it it uh like Benedict says, uh it's about the traceability. So when we certify a material, we actually uh just don't give the chemical composition, we give the uh process history or surface finish or the gas atmosphere that material has been made. So uh when uh when a customer wants to buy materials from us, and then he will have all these data available to him so he can compare between the batches or the supplier or when uh the changes to a new region, they can compare this. So uh they can track these changes and they don't have to start again, but they can track these changes and go back to where it has drifted, and then they can just redo it from there itself. Yeah, I would say.

ADAM SELLS 11:41

Yeah, so the key there is that that we have traceability, right? And then you can track the changes, so maybe something small changes it may indicate something that we should look into before a problem happens.

Dr. ARUNIMA NAIR 11:52

Yeah, yeah.

DR. BENEDIKT MOSER 11:53

And I think at the end it's also uh a question of the robustness of your process. How robust is your application of a material um towards any changes uh in that material? And I mean, coming back to your original question, you probably should do that uh that change of supplier or widening your source uh in a very controlled way. You might have a supplier, everything works, you can run that for 10 years, and then you should not just switch the supplier because then you might really uh go back to field one and uh do a redevelopment. But you can start evaluating other suppliers, and uh then you can slowly in a controlled manner uh improve your specification and uh and find out what really matters and uh then widen to several suppliers. You just need to do that in a controlled way uh because with always with having in your head that you might actually fail with a new supplier.

THOMAS GREAVES 12:57

Also, to build on that, when companies are innovating and starting fresh, it's all about having the right partnerships. And one thing that Goodfellow has adapted over the years is then to have the right partners in place to sort of minimize the amount of testing that's required, but also make it effective for the application. So obviously, you want to do a robust test framework and make sure it's comparable with the development, but also you need to have the right partners guiding you on which sort of properties that need to be looked at, and also application testing. We see that time and time again, that there's a specification based on a material properties, but they're not fully thinking through the application, and this is where some of the unknowns then creep in.

ADAM SELLS 13:36

How often do we think at the RD stage, and when we're talking about these suppliers and uh we're thinking about the geopolitical impact of samples that are being used for the RD? How can we steer or what advice would we give to people trying to think ahead to the production scale when we're kind of getting those initial samples for the RD, depending on where they're coming from geopolitically?

THOMAS GREAVES 13:59

So I think from our from our perspective, we often work um heavily in compliance. We have a number of um export-controlled materials, dual use, um, materials that are coming from um particular geographies that either have um only a few sources globally, or there's um certain regions that we need to make sure that we're securing supply carefully and and following um that that um environmentally um friendly sources. So when you can work closely with Goodfellow and working with the supply chain, you then can manage and mitigate the different supply um routes by looking at if it's sourced from western regions or from Chinese um from APAC regions. So it's all about balancing that and doing that due diligence up front also helps because if you are focusing in a scale-up position, you're just looking at the small quantities of material, you're maybe not thinking ahead about the the restrictions, and that's where again working with the right partners can help build you and give you that confidence through the development.

DR. BENEDIKT MOSER 15:02

If we look at this question from an R&D perspective, if you already from the very beginning think about sourcing, think about manufacturability on a scale, then you might actually kill the innovation at the at the beginning, because uh then you you always you can always say, well, it's not possible to source it, etc. etc. So you shouldn't think about that from the very beginning, because you need to do innovation. You need to do uh and these guys that do the innovation they need to be free in their thinking and uh have crazy ideas. But then at one point, like maybe past feasibility or something like that, you should have a separate track that uh that covers all the questions that might come with scalability. Uh, questions on how you can source, how you can manufacture on a on a larger scale, etc. etc. So you should do that early, but you should take care not to kill the innovation uh at the very beginning.

ADAM SELLS 16:00

Yeah, I think that's a good point. That's gonna come to you, Arunima, from your experience when you're in academia doing research. You don't want to be too kind of um stifled by all of these additional complexities. You want to be able to focus on the science, right?

Dr. ARUNIMA NAIR 16:15

Yeah, yeah, that's true. Like so in academia and the lab scale research, it's just on the material development in a small scale, and then in lab lab, everything is controlled or like uh clean environment, and when it goes to uh prediction, that's the problem where there is the problem. So, yeah.

ADAM SELLS 16:35

And and and for us, a lot of the small samples of materials that we provide into academia, you know, we're customers are using the website, they're buying it, we're shipping it, we don't know, and and we don't ask necessarily all the time everything that's being used. But as soon as we try to get a feeling for the fact that this might start scaling, that's when we'll try and be proactive in terms of looking at that product. Is what's the manufacturability? Can it be produced in volume so we can start guiding the the customers? Um so um in practice, um, what does it take to provide the same material reliably, repeatably into um into a manufacturing example? I think one of the other um case studies that Simon talked yesterday was was the electronics example where it's scaling very quickly, there's a big demand in the market at the moment in semiconductors. How do we in practice do that? And especially when you're looking at multiple production sites as well, how do we control that?

THOMAS GREAVES 17:38

That's a really interesting application because it's one that has really spanned going from this a very small scale to to quite significant um production volumes. Um so, in the example that that Simon's given in semiconductors, you're really working with the the fundamental challenges of engineering where they were working with with the Goodfellow team on just technically how to approach and manufacture components, and they were iterating um relatively quickly but trying to fail fast and moving forwards. And as they they had almost set forth on a journey uh preempting what the market would need. So it wasn't necessarily a product that was uh ready for the market now, but when certain technology was brought along, that's when it was really um uh able to then start scaling, and that scale then came with a rapid progression. And so the the technical challenges moved from being able to um machine parts and iterating through um complexities from a technical perspective to how do we uh maintain volume, how can we uh commit to uh providing samples and product product production on at such a scale that moves from different regions as well? So that's the real challenge that we've had to um adopt and work with.

ADAM SELLS 18:53

Yeah, and adapt with the customer as they move from essentially an R&D organization into a large production operation whilst navigating complex global supply chains, um different volumes coming in.

THOMAS GREAVES 19:06

So we've even had to you know look at you know, can the current sites maintain that capacity? Do we need to expand? How quickly, when should we start to expand that capacity? Who's gonna uh fund the capital and infrastructure requirements? How are we gonna train staff quickly that are in a real skilled operation? You know, there's these almost commercial aspects that at the inception of the of the product um seven years ago were were not even considered. So it's also about working closely with the customers and having that close relationship.

ADAM SELLS 19:37

Yeah, and that's something that we won't have time to talk about today, but uh but yeah, uh having enough skilled resource to come in and and work in these advanced manufacturing spaces is definitely a challenge in in lots of areas. But before we uh close and go to some questions, I wondered if each of the panel could give give us one piece of advice uh if they were talking to somebody that was planning a scale-up, start with Benedictin and come this way.

DR. BENEDIKT MOSER 20:00

Yeah, maybe uh uh again coming back to your last question about uh what is a critical point. A very critical point in maintaining uh the quality is also the traceability of the materials where you buy them from. Because the the the worst thing that can happen to you is that you're not looking for a new supplier, but then you know that you're changing something, but that somewhere in your supply chain you don't know something has changed, and all of a sudden it doesn't work anymore, and you don't know why. And of course, no one has changed something. That's usually what we find uh when we do the failure analysis, because then that's that's when customers come to us and ask, Why does it not work anymore? Uh and then we try to find who has changed something, so no one has changed something, of course. Uh the specification is not right, uh, that's normal, that's the specification doesn't cover everything, and that's where the detective work starts and finding out what has changed that it doesn't work anymore. And that's the most dangerous thing, and you can only avoid that by uh working on the traceability, which of course is not for free, that costs something, um, and that depends also a lot on the industry where you're working in. Uh, when you're in aerospace, uh you have the traceability, but when you're working in uh in consumer goods or whatever, uh usually your supply chain doesn't care about that.

THOMAS GREAVES 21:27

And for my side, I would say don't let perfection um get in the way of progress. I know it's a common phrase, but I think when you're scaling up and you're innovating and developing new products, it's easy for companies that have got cash flow and they've got existing um product lines, they can shoulder the burden of development. But when you're developing new products in a in a scale-up process, being able to remain agile and and reconsider the process and and pivot certainly helps. So if you're focusing on perfection, trying to take every single box off in terms of the performance of the material for the application, um, you need to make sure that you can pivot to to to enable products to be developed in a proper process. Thanks, Tom. Arunima?

Dr. ARUNIMA NAIR 22:11

Uh I would say that don't wait until the end. So uh the lab scale research and the prediction technology research should progress parallelly so that one can help the other. So there won't be any surprises at the end when you are scaling up. Yeah.

ADAM SELLS 22:26

Are there any questions? Hopefully it was uh we covered everything then. Um I'm sure not. I'm sure there's lots more. Uh, if anybody wants to come and find us afterwards, then obviously please feel free, we'll be happy to carry on any discussions. Thank you. Thank you. Thank you.