Apple Podcasts Stitcher Spotify Pocket Casts Google Podcasts


In this episode, Dr. Bryce Meredig and Dr. Arnold discuss:

  • The business case for adopting data-driven methodologies and materials informatics within materials and chemicals organizations
    Suggestions for kicking off digital initiatives within materials and chemicals organizations
  • Dr. Arnold’s wide-ranging career in the materials industry, and her perspective on the impact that data-driven R&D can have on the pace of innovation within an organization
  • The similarities between Six Sigma and materials informatics adoption and implementation
  • How materials informatics and a data strategy can help facilitate knowledge transfer within an organization

“Start with where you have problems and opportunities, and ask what data can support delivery on these solutions and these opportunities.” – Dr. Cynthia Arnold

Speaker Bio

Dr. Cynthia Arnold spent an over 30 year career in materials industries, leading technology development and innovation and marketing partnerships for diverse applications including medical, automotive, aerospace, electronics, building and construction and printing. She most recently served as the Chief Technology Officer of the Valspar Corporation and as the Senior Vice President of Global Technology for Sherwin Williams, where she was responsible for a world-wide team of 1100 staff developing paints, coatings and associated technologies and services. Prior to Valspar, Cynthia was the Chief Technology Officer at Sun Chemical, a leading producer of inks and pigments. Prior to that, she held technical leadership positions at Eastman Chemical and General Electric.

She served as a Sloan Executive Science & Engineering Fellow in the White House Office of Science & Technology Policy for several years, focused on industry-government technology alliances. She is currently a member of the Advisory Committee of the University of Minnesota’s Department of Chemical Engineering & Materials Science; a Board member of the Minnesota Zoo; and serves in advisory and consulting roles for a number of materials companies, including Carbon, Cabot, Citrine Informatics, and Milliken.


Bryce Meredig: Welcome to Data Lab, a materials informatics podcast with Bryce Meredig, Chief Science Officer at Citrine Informatics. Dr. Cynthia Arnold spent an over 30 year career in materials industries leading technology development and innovation and marketing partnerships for diverse applications, including medical, automotive, aerospace, electronics, building and construction, and printing.

Bryce Meredig: She most recently served as the Chief Technology Officer of the Valspar Corporation, and as the Senior Vice President of Global Technology for Sherwin-Williams, where she was responsible for a worldwide team of 1100 staff, developing paints, coatings, and associated technologies and services.

Bryce Meredig: Prior to Valspar, Cynthia was the Chief Technology Officer at Sun Chemical, a leading producer of inks and pigments. Prior to that, she held technical leadership positions at Eastman Chemical and General Electric. She served as a Sloan Executive Science and Engineering Fellow in the White House Office of Science and Technology Policy for several years, focused on industry-government technology alliances.

Bryce Meredig: She’s currently a member of the Advisory Committee of the University of Minnesota’s Department of Chemical Engineering and Material Science, a board member of the Minnesota Zoo, and serves in advisory and consulting roles for a number of materials companies, including Carbon, Cabot, Citrine Informatics, and Milliken.

Bryce Meredig: Cynthia, welcome to Data Lab.

Cynthia Arnold: Great, thank you Bryce, it’s a pleasure to be here.

Bryce Meredig: Cynthia, as you know, I always like to start us off with a fun fact about our guests, and in your case I understand that you had an interesting experience with German titles at one point in your career. Could you share that story with us?

Cynthia Arnold: It was a while ago, but I was General Manager of a joint venture between GE Plastics Division and Bayer, now Covestro Group. We were working in the auto industry, and we were talking about how to prepare our business cards. We had a new team coming in, and there was a lot of discussion about business cards, being US information on one side and then our German office address on the other.

Cynthia Arnold: And because I was a female General Manager, the spelling of Geschäftsführer, Managing Director, became under debate. I guess it was a little difficult at the time, as long ago as it was, in order to understand how you have a female version of a Geschäftsführer. So we ran with it, that was a great experience.

Bryce Meredig: Great, and hopefully now that’s something that is a lot easier to do, to have the female versions of those titles printed.

Cynthia Arnold: I hope that everybody understands what that version should look like pretty easily now.

Bryce Meredig: Great. Well, obviously you’ve had a wide ranging career in the materials industry, and you’ve said that you believe that the time has come now for companies to start to use data-driven methods to unlock the value of the materials data that they have. What specifically is the value of materials and chemicals data within an organization like the ones that you’ve led?

Cynthia Arnold: That’s a great question. I’ve been a big believer of being scientific and data-based throughout my career, running the organizations that I have. I strongly believe data within an organization is a huge asset. Unfortunately, most companies use only a fraction of the data at their disposal; it’s used once for the intent for which it’s generated, and then typically stored away and generally not easily accessible.

Cynthia Arnold: There are a lot of reasons why data isn’t reused. It could be that the data don’t have an understandable taxonomy that makes it easy to look up, they aren’t tagged in a way to make it easily retrievable. Or that data are commonly stored in different systems across an organization.

Cynthia Arnold: Some data are numerical and structured, but a lot of data, especially research groups, it may be in complex forms such as micrographs and photos as well. So you have data coming in all these different forms, and it’s the heart of what a research scientist or a product developer uses to really deliver a product to the market.

Cynthia Arnold: Imagine a research scientist or a product developer who does seven iterative experiments to develop a product, and each experiment requires 10 different tests to validate the product performance. If six of those experiments come close to meeting the customer requirements, but don’t exactly deliver on it, maybe you’re lucky, the seventh is going to deliver close enough requirements, and you’re going to run with that to the market.

Cynthia Arnold: So what happens with the other data around the six different experiments that preempted finding the right one to deliver to the market? It’s usually stored away and not likely referenced again, even though requests for new product innovations will likely require revisiting similar performance attributes, or especially building from the physical relationships that were exhibited in the experiments that were undertaken.

Cynthia Arnold: You’re not really learning from the data that was already generated to deliver that good product to the market. So the next time a request comes in for a modified or a new product, you aren’t fully leveraging the work you’ve already done. And this, of course, is not at all productive for an R&D organization.

Cynthia Arnold: It costs time to get to market, you could be a lot faster if you were leveraging the data that exists, but you essentially start over. Your researchers may have learned from that series of experiments, but it counts on them remembering what to reference in the next round of experiments and product developments that they engage in.

Cynthia Arnold: So on a deeper level, this kind of approach leads to developing products that aren’t optimized for performance or for cost. It’s really straightforward: the more data you have, the better the conclusion, so being able to go back and reference that historical data is going to give you a better product. So the value of using this data is to obtain that better product and get to market faster with less work. It’s a profoundly influential value proposition.

Cynthia Arnold: And through the predictive capability of materials analytics, on top of it you have this available larger dataset, you can have higher confidence that the experimental work will yield the results that are required. So you even have the knowledge beforehand of some predictability of the likelihood to meet that outcome, and this allows for a really different type of conversation with your customer.

Cynthia Arnold: Imagine a salesperson comes in and wants a modified or a new product. Being able to reference this broad historical dataset will allow you to talk to them very early about whether that product opportunity is actually realizable or highly risky. Instead of expending a lot of resources and leading the customer down that path, you can really change the engagement with your customer and have a really different dialogue with them up front.

Cynthia Arnold: If your products are more innovative, you should be able to capture a higher margin and higher market share. And at a broader level, when you think across the company at the enterprise level, product management should be really pleased to be developing a stronger portfolio of products overall. This can lead to simplifying the product line, and that has profound implications for manufacturing productivity and procurement benefits. So, again, this is a very broad-based at enterprise level value proposition with what materials informatics has to offer.

Bryce Meredig: How would you explain, or to what would you attribute, the suboptimal data landscape within companies today, the problems that you alluded to at the beginning of your statement there?

Cynthia Arnold: Well, data related to products and materials sits in a large number of parts and groups in an organization. We’re talking directly about R&D, or product development, and of course there’s R&D data, this takes many different forms. There are a lot of physical lab notebooks that are still used for IP protection, in generating IP and protecting that and having a record.

Cynthia Arnold: There are project management systems of all different sorts, some are homemade, some are commercially available. Lab information management systems, homemade systems, spreadsheets, photos, micrographs, there’s such a wealth of information, of data that’s available, just in the R&D technology group. On top of that, customers may have data on the performance of your products, how your products work in their operations, and how they work in your customer service environment.

Cynthia Arnold: So that’s concerning just the product directly, but at an enterprise level, product and materials data are also held in many other places. They’re held in the procurement organization, which supplies the components of your product; in operations, regarding how the product is made; in quality systems; in CRM, customer relationship management systems, that capture what produce attributes your customer values and what they would like; in ERP systems that capture project management data and cost data; regulatory management systems; and on and on.

Cynthia Arnold: I think this gives you a feeling of all of the disparate data related to managing products in a materials or a chemical company. And all of this data constitutes a very deep materials informatics infrastructure. So I think some of the problems have to do with the volume and the diversity of the data, where it’s located, the breadth of where it’s located, as well as data quality, which is an issue in many cases.

Cynthia Arnold: However, companies don’t really need to boil the proverbial ocean to make progress in better using the historical data and the existing data. Doing a data inventory I think is a very good starting point, understanding what you have. Then identify which data will provide the most value, and how it will be used once you can organize it and make it accessible. And this becomes the basis of a data strategy and a data roadmap within the organization.

Cynthia Arnold: And again, it doesn’t need to be 100%. Start with where you have problems and opportunities, and ask what data can support being better to deliver on these solutions and these opportunities. When the rubber hits the road, you can put together a focused team of a few content experts and some systems people. They can work together to help clean up and organize and put the identified datasets to use on real opportunities rather quickly.

Cynthia Arnold: And then, I think if you’re systematic about this and proceed down the line with fulfilling your data strategy really directed at real opportunities and problem solutions, I think you’re going to find you’re going to build your way into a good data taxonomy and framework that will be highly reusable and valuable.

Bryce Meredig: Well, you’ve highlighted some of the profound challenges around data that many materials and chemicals organizations face. What specifically are some of the barriers to improving this data landscape, to getting to a better position? I imagine some of them are technological, probably some are cultural as well.

Cynthia Arnold: I think culturally it’s hard to take the time to look back and say, “We’ve got to go back and address historical data, clean it up, and organize it, because this will help us in the future.” When we were rolling out some systems for Six Sigma back in the late 80s, early 90s, the same thought was about design of experiments. When we were talking about, “We’ve got to generate more data,” it’s, “Why can’t you be quick? We’re just going to rush to the lab and we’re going to whip up a product based on our institutional knowledge, we’re not going to use a lot of references. This will be the way to be faster.”

Cynthia Arnold: And in fact, the cultural change there was quite similar to the one that I think is relevant from adopting materials informatics. It’s the fact that you have to take some time to design experiments and think thoughtfully about, “What experiments are really going to give me the data generation I need to prove or disprove a hypothesis about getting those product properties?” Take your time on understanding your customer needs in setting up the experiment. So that was a cultural change.

Cynthia Arnold: Now the cultural change is, “Let’s go back and organize the data.” We have so much value that exists; rather than essentially recreating it, and trying to reference it just through people’s memories, the best of our people’s memories, let’s really try to put some structure to it and organize it, and use this as the framework in how we change working going forward.

Cynthia Arnold: So again, I think this is the heart of the culture change, is we’ve got to slow down and go back and reuse some of what we bought. And, to my point earlier, not everything, but you’re going to find certain datasets are going to have sufficient quality, they’re going to be reusable for today’s problems and today’s opportunities. Those are the ones to focus on, to go back and use and apply to your opportunities of the current business.

Bryce Meredig: You’ve been a technology leader at a number of large materials companies, and you’re clearly a proponent of these data-driven methods. If you were to put your CTO hat back on and you were going to make a case to your colleagues, to other executives, to invest in materials informatics, how would you make that case?

Cynthia Arnold: In the past, how I approach making any capital investment, an investment that’s substantive for a company, is to go back and do a process map of how we currently work, and lay out a new way of working that’s supported by materials informatics, how that would reshape how we work, and importantly, quantify how this could improve outcomes for our business.

Cynthia Arnold: Innovation and speed to market leads to market sharing margin, so you can build in some assumptions about how being faster to market and how having better products is going to lead to those financial opportunities. More efficient product development with a higher probability of success makes the whole organization more efficient. You should be able to get more work out of the existing organizational staff, they should be able to work more productively once the system is effectively set up.

Cynthia Arnold: Materials informatics is also necessary, for instance, to effectively utilize lab automation. A lot of materials, chemical companies are looking at low-skill labor that’s repetitive, doesn’t require a lot of institutional knowledge or scientific knowledge, and trying to use lab automation as a way of speeding up product development. But lab automation necessarily requires materials informatics in order to function well repeatedly, which is the goal of the system. So that would be an opportunity for a double advantage, to be more productive and faster to market as well.

Cynthia Arnold: I think looking at current ways of working, and you can take specific scenarios: Whether it’s doing a new product that’s relevant to the business today, or whether it’s adopting lab automation and changing the productivity about how the organization works, or whether it’s to support translating knowledge and getting global colleagues up to speed, all of these are specific to your business environment, but doing a process map and laying out the financial benefit is the way I’d go forward to try to justify the investment.

Bryce Meredig: So you focus your argument here on financial and productivity reasons. Do you think this parallels at all the arguments that companies used, or the case that companies used, to invest in Six Sigma?

Cynthia Arnold: I think it does. Six Sigma was used twofold, but predominantly it was used to deliver productivity. That proved to be of big value. Design for Six Sigma, a little bit different flavor of the Six Sigma protocols, was used for product development and to be faster and more rigorous in developing the right attributes to meet specific customer needs. So for me, Six Sigma really brought in the voice of the customer, and doing the experiments to validate quantitatively that you can hit the targets of what that voice of the customer was asking for.

Cynthia Arnold: I believe materials informatics goes well beyond that, although it’s kind of Six Sigma on steroids in many cases because, by referencing historical data and having larger datasets, necessarily the outcomes that you would use in a design of experiments in Six Sigma protocols should become much better at meeting customer needs, because you’re accessing a larger dataset.

Cynthia Arnold: But I think materials informatics goes beyond that because it allows you, not only to reach inside your current design space and be a lot faster, but it builds on some knowledge of physical relationships that are happening through that experimental database, and allows you to step outside of your current design space much more effectively and to really innovate.

Cynthia Arnold: Six Sigma was designed to hit targets, and again, the opportunity to innovate I think is propelled much further forward with materials informatics as we’re thinking of it with the tools available now.

Bryce Meredig: Well, we’ve had other guests here on Data Lab that have mentioned an industry demographics issue, specifically that there is a generation of professionals who are getting ready to retire, they’re exiting the workforce, and they’re taking decades of irreplaceable institutional knowledge with them. How, in your view, can we facilitate the transfer of this knowledge to the next generation of employees coming into these companies?

Cynthia Arnold: Bryce, that’s a really important point that all chemical and materials companies face today. Workforce talent is indeed a big issue in these industries. We’re not generating enough technologists and scientists to fill the positions that are being vacated from retirement.

Cynthia Arnold: The most common approach that you see is really old-fashioned, it’s to hire new people and have them overlap with their more experienced colleagues for a while before they retire. And given how budgets are managed, you’re usually allowed a period of one to maybe two years. That being said, in some of the organizations that I’ve managed I’ve had many directors come to me and say, “I need to hire new people, but it’s going to take five to seven years to get a new person up to speed, to understand what I do, before I retire.”

Cynthia Arnold: So here you have this mismatch about the time of overlap that’s allowed financially versus what is really required from just working together experientially over an extended period of time in order to translate a lot of that institutional knowledge. That’s clearly not appropriate, this is a type of transition that’s clearly fraught with risk to lose value.

Cynthia Arnold: A more concerted effort to document and structure the deep experience and the insights of the more experienced workforce is clearly required. And again, this is where culture change is required. It takes time to sit down with someone, although it’s incredibly valuable to leverage 10, 20, 30 years of experience in these industries where these people have had their focus. Even without the imminent retirement cliff on top of it, translating knowledge across a global organization can really get enhanced through better knowledge management practice.

Cynthia Arnold: So I strongly encourage a concerted and specific effort to document and structure the experience and insights of experienced people, and to think about that in respect to the data strategy that you have is how do you upload some of that relevant knowledge through spending some time in documenting it? You want a taxonomy of that knowledge, you want some structure to make that knowledge accessible for the future, but to think about that in terms of a strategic data management roadmap.

Bryce Meredig: Whenever you’re talking about investing additional time and effort to structure organized data, you’ll often encounter some pushback as to, “Why should we do this?”, especially from people who are very experienced. But of course, if materials informatics is an enabler here and knowledge transfer is a goal that can be enabled by materials informatics, it would seem that that would help us make the case to invest the additional effort, the additional time and energy, into organizing our data better from the outset.

Cynthia Arnold: Yes, absolutely, I couldn’t agree with you more on that. And again, you mentioned about culture, culture’s really important, but to get people to understand the value of just taking a timeout to get some content experts in particular to spend time to organize their data and document what their key learnings are is very critical.

Bryce Meredig: Given your wide array of experiences as an executive and a board member, including here at Citrine, in the materials and chemicals industries, what are you most excited about in terms of the future of these industries?

Cynthia Arnold: Well you have to be clearly excited about the value proposition we’ve been talking about, Bryce. Innovation, productivity, efficiency, capturing knowledge that could otherwise be lost, translating it more effectively. I think many CTOs, Chief Technology Officers, those in leadership positions in product development organizations, I think this absolutely resonates with them.

Cynthia Arnold: But boards are now starting to talk for instance about digital transformation. There’s been a lot of talk about cybersecurity, but the broader opportunities of using digital tools are really becoming recognized at the board level and executive levels of companies. And materials informatics is a key element of this.

Cynthia Arnold: If I look back historically on my own career, in the 1980s it was considered quite novel to have project management processes and systems in place. And then how many of us have gone through putting ERP systems in place? This has happened throughout our careers.

Cynthia Arnold: In the late 90s it was lean management and Six Sigma, where data was more rigorously generated and used throughout organizations, again, as I’ve mentioned, to mostly capture productivity, but also to be much more targeted at hitting customer needs through voice of customer assessments. That was considered a big change in how businesses were building upon project management processes and ERP systems. Then CRM systems have come along, so that’s supporting, again, more voice of customer and how Six Sigma was leveraging.

Cynthia Arnold: Materials informatics is part of this next wave for business transformation in my mind. The opportunity for innovation, productivity, efficiency is huge, but more importantly it’s built upon business process improvements that have been progressively put in place over the last three or so decades. So we’re ready for it.

Cynthia Arnold: Materials and chemical companies shouldn’t see this as something different, this is part of a progression of business transformation, and it’s building upon these previous generations of improvements, these other software tools and processes and culture changes that have already been put into place.

Cynthia Arnold: In the next five years, I really believe we’re going to see leading companies that know how to leverage their data assets, and they will stand apart from the crowd in the market in terms of their financial performance. This, to me, is the most exciting aspect of where we’re heading. It’s going to be an exciting new way of working built upon how the industry has been transforming itself over the last few decades.

Bryce Meredig: Why, in your view, has digital transformation become a board level conversation for materials and chemicals companies?

Cynthia Arnold: The magnitude of the opportunity, the availability of the tools, and the promotion of the tools and processes through other players within the industries broadly. This is being raised in awareness, you see talk of artificial intelligence, although I don’t think this is anything having to do with artificial. This is machine learning like we’ve used in the past for something as simple as computer-aided design and computer-aided engineering. It’s a tool to help organizations.

Cynthia Arnold: And I think now, as part of broader enterprise opportunities, say around managing your supply chain or managing your customer relationships through billing cycles, managing inventory. Materials informatics and what opportunities you have for product development and product line management now are being part of this digital transformation thought process. Boards are now aware of it because obviously computers have raised the issue of cybersecurity, but that’s transforming itself from managing downside risk of cybersecurity into managing upside opportunities.

Bryce Meredig: Well, Cynthia, that’s all the time we have for today. Thank you so much for joining us on Data Lab.

Cynthia Arnold: Good, thank you Bryce, I appreciate it. It’s a great time to be working in materials informatics.

Bryce Meredig: Thanks for listening to Data Lab. If you have questions or an idea for an episode, contact our team at podcast@datalabmi.com.