A ternary plot helps scientists visualize how changes in the proportions of ingredients in a formulation affect a response. Scientists color-code responses, letting them quickly learn more about the properties of their formulation.
Analytics upskilling transforms an organization on the move
When early benchmarking at Parker LORD showed that JMP® could have potentially dramatic time saving benefits, employees led the way in changing the culture
|Rapid growth in the electric vehicles market puts pressure on manufacturers to double the speed of innovation while also raising product performance and quality. Employees at Parker LORD saw early on how statistical methods like mixture designs and other forms of design of experiments (DOE) could help the company remain nimble in a competitive marketplace. Implementation, however, hit roadblocks because domain experts were reluctant to adopt statistical methods they did not fully understand.
|An initial trial of JMP enabled formulation chemists, working together with Six Sigma and quality engineering leaders, to dramatically reduce experimentation times and deepen process knowledge. A customized training course developed in partnership with SAS Education helped the Parker LORD R&D teams to gain the skills they needed to maximize the software’s value. Finally the scientists had an industrially relevant statistics course, mentoring and the tools to help them.
|In just 18 months, access to JMP – and JMP training from SAS Education – has accelerated a cultural shift toward fast-paced, data-driven, precise innovation. Early benchmarking showed cost savings and faster experimentation times, which have a direct impact on costs of goods sold – and lead the industry with cutting-edge performance. These successes have persuaded Parker LORD leadership to expand JMP use at sites around the world.
For more than 90 years, Parker LORD has worked in close partnership with customers to develop highly reliable adhesives, coatings, motion management devices and sensing technologies that significantly reduce risk and improve product performance. The company’s innovative oil and gas, aerospace, defense, automotive and industrial solutions today help bridges withstand earthquakes, strengthen critical medical equipment, safely contain corrosive chemicals and make electric vehicles more affordable for more drivers.
A long history of technological leadership, together with penetration in key high-growth markets, is now bringing value to Parker Hannifin Corporation, a Fortune 250 global leader in motion and control technologies. Parker’s strategic acquisition of LORD Corporation in late 2019 set in motion the integration of two organizations’ strong materials science, electrification and aerospace capabilities.
Now known as Parker LORD, the organization has rapidly transitioned from its focus as a custom adhesive manufacturer to serve the broader automotive market with larger, more demanding customers. In particular, fast growth in the electric vehicles market has increased the pressure to reduce development cycle time to develop products with unprecedented quality and performance requirements.
Global supply chains call for a strategic approach to scaling and quality control
“Our customers value speed because they are also developing very quickly to be competitive. They're transferring those same pressures onto us while maintaining the same level of quality expectations for both legacy and automotive products,” explains Tim Holtz, a Process Engineering Manager at Parker LORD. Holtz leads a global process development team working between R&D, manufacturing, sourcing and quality on implementation and technology transfer across three continents. The team, he explains, is responsible for scaling product development and creating manufacturing processes that can be deployed to facilities around the world.
The globalization of supply chains has accelerated the demand for international operations, explains Quality Leader and Six Sigma Black Belt Bernard Goguelet. Quality expectations, however, remain unchanged even as teams navigate varying site-specific conditions like raw material suppliers, equipment, workforce norms and even climate conditions. That’s why it is imperative that quality be designed into the product during development rather than in production, Goguelet says: “That’s where we can maximize the business value. It’s why I come to work every day.”
Whereas quality-related analytics at Parker LORD were designed for high-mix, low-volume products, Parker LORD has transitioned since its acquisition from producing a small volume of high-value products to producing a high volume at high value. As a result, the team is challenged to constantly improve manufacturing throughput, an area where analytics, Goguelet says, “can really help. We’re seeing the use of analytics increasing in our operations too – not just design.”
Changing the culture, from gut feel to data-driven decisions in 18 months
With a deep bench of domain experts, Parker LORD has long had the chemical expertise to guide decisions at all levels. With mounting pressures in the marketplace to operate more nimbly, Holtz and Goguelet are moving the organization toward integrating statistical approaches that reduce risk and decision-making timelines. This shift puts statistical approaches like design of experiments (DOE) in the hands of experts.
“[Scientists] should look at statistics as a supplement to make what they do easier, to make the next decision easier,” Holtz says. Methods like full-factorial DOE, for example, can confirm a hunch and point scientists to the next step more quickly. “On a speed and quality level, we have to make every experiment count,” Holtz says. “We want to minimize the amount of experiments and be able to learn as much as we can with the materials, both with product development and process development.”
Time and cost savings also emanate from further down the pipeline, where Holtz explains that reproducible approaches build the process knowledge needed to scale operations. “If we can build a knowledge base as we go, we can respond faster to customer needs,” he says. Analytical tools help teams adopt both reactive and proactive quality measures; not only can they use analytics to solve any quality issues that appear during the scaling process, they can refine the ways in which materials are transferred from the product development stage.
“The culture is starting to shift,” Holtz says, explaining that what began initially with just a handful of scientists showing interest has in less than 18 months blossomed into an organizationwide analytics initiative that is opening the door to wide-ranging improvement opportunities. The initiative gained traction in large part because of Holtz and Goguelet’s efforts to test and document early proofs-of-concept and bring quantifiable ROI to the attention of management. But before that, the pair had to gain scientists’ buy-in.
“Breaking through that initial skepticism was the hardest part – but also the inflection point,” Goguelet says. Once scientists saw how custom DOE could help them fit their problems into a framework, they began to understand the benefit. And, says Goguelet, the inflection point came when they saw JMP.
“The way JMP has proliferated at Parker LORD is through word of mouth. It’s through seeing friends excel at something. We want to develop a cohort that’s really motivated to perform well, that way they can help their peers. That brings the mean competency up in the organization.”
– Bernard Goguelet, Quality Leader
When early adopters evangelize the power of analytics – and JMP® – to colleagues, a culture transformation is born
Getting an accessible software like JMP into the hands of scientists – most of whom had little previous training in statistics – meant decreasing the learning burden on experts already busy with scientific work. Most had used Excel for years, and the thought of using the tool for more sophisticated statistics seemed daunting, Goguelet says. “With Excel, you’d have to do huge studies to look at everything… and we’d be lucky if we could even get an R-squared value!” he laughs.
A menu of ready-made analysis tools like the Custom DOE platform in JMP quickly showed benefits over existing best practices. “That’s when we really see the difference,” Holtz says; “They're like, ‘Wow, this makes my life easier somehow.’ Or ‘it helps me learn something that with a ladder study used to take me 1,000 experiments and I can now learn in 20.’”
“Our scientists learned how the way you choose to set up the experiment has a huge impact on the quality of the data you get,” Goguelet remarks. “JMP’s Custom DOE module helps us make sure each experiment yields the most information.”
A few early adopters embraced the tool immediately, working together with Holtz and Goguelet to explore how JMP might help solve challenges. “They built cases to help show the power of [analytics], and it started to look more approachable,” Holtz explains. “These [early adopters] were willing to help others learn – they became the experts. So we didn’t have to break through with everyone, just with a few people and use them as lead adopters.”
Goguelet says early enthusiasm has helped accelerate the culture transformation. “We’re already seeing signs of people integrating it into their daily thinking: I’m going to develop a new product and use my intuition to get close – that’s the subject matter expertise. Then I’ll use JMP to define the region and to optimize it. They’re learning things about their models that they wouldn’t otherwise have known.”
The Prediction Profiler lets process engineers visualize how sensitive the formulation is to changes in the proportions of ingredients.
Mixture design saves time (and reduces scrap)
Several of the team’s early wins came on the chemical formulation side of the business, where chemists had some legacy knowledge about formulations but were tasked to adapt mixtures to new conditions. The team had some intuition on how to reach the desired mixture by adding more of one of the key ingredients but limited statistical evidence to back it up. When one of the formulators approached Goguelet to ask about statistical approaches, they worked together to set up a mixture design in JMP. The results showed, in direct contrast to the team’s assumption, evidence of strong nonlinear effect between two key ingredients previously thought to be of little consequence. Adding more of the ingredient wouldn’t have any impact.
In Parker LORD process development, making one batch can take a day of work. Ladder studies, the previous tool of choice, could easily require 30 to 60 batches – that means a month or two. “We used JMP’s mixture design module and unlocked better insight with dramatically fewer batches than our previous method,” Goguelet said.
“Had we not had JMP, we would have sent that [instruction] to the plant and kept adding more [of the ingredient] as we have for the previous 10 generations. There’s a direct impact on our costs of goods sold. It would have meant extra cycle time eaten up trying to get that thing at the tolerance. Worst case, it would have been scrapped or delivered late to the customer…. That was the first time I thought ‘wow.’”
The formulation chemist was also impressed with the result. “The funny thing is they were a complete skeptic at first!” Holtz adds. “We had to say, just try it. [But after that result], they said ‘I’m sold. I’m going to use this all the time. How do I get a JMP license? How do I get training?’”
Training through SAS Education helps Parker LORD maximize the value of their investment in JMP
With more and more scientists showing an interest in JMP, it became clear to Holtz and Goguelet that training was indeed needed if users were to take advantage of the full spectrum of the software’s features. Mixture design, a methodology of which Parker LORD formulation chemists were not previously aware, was just one example where the team saw transformative potential but lacked the expertise to roll out new methods in a meaningful way.
“We – along with our management – looked at it like this: If we’re going to invest in JMP, then we have to invest [in training],” Holtz says. “We realized we didn’t have the competency in-house and weren’t going to develop it out of thin air.”
“We started learning how much we didn’t know,” Goguelet adds. “JMP is only as powerful as the person who's driving it... I realized in order to get the actual value from [JMP], some additional investment was needed. That’s when we decided to engage SAS Education.”
Unlike other programs, SAS Education offered the opportunity to work with experts at JMP to design a customized training program. Benchmarking off other companies helped Holtz and Goguelet to determine a range of industry standard tools. Experts from SAS Education with expertise in the chemical space then worked from this list to identify the skills Parker LORD’s formulation chemists would need most. “SAS helped us fill in the blanks: we told them what we were trying to achieve, and they told us the tools we needed to get there,” Goguelet explains.
Parker LORD scientists dispense a laboratory-scale batch onto a substrate. The scientists will evaluate the mechanical properties of the cured adhesive, then use JMP to help unlock insights about how factors from their designed experiment affect those properties.
A custom training program helps scientists learn the skills they actually need
Whereas previous trainings at Parker LORD had struggled to retain participants’ attention, SAS Education’s bespoke approach focused only on the skills formulators would need on the job. “In order to get to implementation, you have to focus on the problems people care about,” Holtz says. “If you fix something for them and make their job easier, then they’re going to use it all day long. You’ve solved a problem that was taking up a lot of their time and driving them nuts.
“The Education team really tailored the training for what we needed. I’ve heard from many of the team members who attended that this was one of those trainings they actually found useful. We were able to really leverage SAS’ knowledge and experience interacting with different customers to help identify what we needed and how best to approach it…. It’s been a good partnership.”
From 10 JMP® users at one site to many at many sites in 18 months
The success of the first seven-day training at the Parker LORD North Carolina division headquarters soon led to two new contracts for sites in Hilden, Germany and Erie, PA. Mentorship will be an important part of continuing the partnership moving forward, and the team is also discussing the possibility of an annual mixture design “booster” course with SAS Education to address new challenges along the way.
“I’m really thrilled with the leadership support that we have now,” Goguelet says, noting that upper-level management beginning with a Parker LORD Director of Technology, have not only signed on to analytics transformation, but are now encouraging that statistical approaches become the company standard. “We’ve actually embraced that more since doing the training with [SAS Education]. We now set expectations around JMP and the tools with it. They’re part of our process development now – it’s an expectation.”
One and a half years after starting with just 10 JMP licenses, Parker LORD has since rolled out JMP on a much larger scale. “The way JMP has proliferated at Parker LORD is through word of mouth,” Goguelet says. “It’s through seeing friends excel at something. We want to develop a cohort that’s really motivated to perform well, that way they can help their peers. That brings the mean competency up in the organization.”
As the company grows, Holtz and Goguelet aim to bring people into the culture and set expectations of competency at certain milestones during onboarding.