Within his current purview of process control, Steve says it’s critical to think not just of one aspect of a process, but the entire value stream. Early changes in any process can have knock-on effects and by taking a broad view of the entire analytics workflow, he has been able to introduce targeted improvements that reduce rework, defects and costs downstream.
In a recent conversation, @shampton82 shared his thoughts – and some advice – on how he’s gotten more value out of JMP, both for his organization and his own career.
Meg: Introducing any new method or tool can be disruptive at first – particularly to an organization that sees itself as already working with the most cutting-edge tools on the market. But that doesn’t mean there aren’t overwhelming benefits to some of those transitional growing pains! What kinds of ROI or improvements have helped to build momentum for the data initiatives you’ve introduced at PCC?
Steve: This question reminds me of the activation energy curves from my school days. Initially, there is a huge energy barrier for a reaction to take place, but by adding in an appropriate catalyst, that energy is greatly reduced. I see JMP – and the support structure that comes with the software – as that catalyst for our company.
At PCC, we’ve had a long history with NEI and the DMAIC mindset, with Minitab being our main stats package. But we were also pretty stuck in an ANOVA and bivariate regression analysis rut. Even worse, with the limitations of Minitab and Excel, engineers would have to hop between the two multiple times during an analysis, and many just ended up using Excel because that was what they knew and could be efficient with. The result was some OK visualizations of our data and processes but a backslide on statistical rigorousness around decisions.
At the time when I was first introduced to JMP, I was a Minitab superuser and, to be honest, didn’t immediately see the value. Nevertheless, once I made the switch with help from @Jordan_Hiller’s onsite training and a few excellent on-demand webinars, I never looked back. Everyone who has made the transition feels the same way.
The turning point for my plant was a pretty large process excursion where we weren’t getting anywhere using our typical tools – but it was costing us hundreds of thousands of dollars as the problem persisted. The project team went back to basics with process mapping and brainstorming and started using JMP’s powerful visualization tools to quickly test theories in a group setting. Then we could run a statistical test to see if they were worth pursuing further. This led to a rapid understanding of what the main experiment priorities should be and a fundamental shift in what we prioritized in controlling our process.
Out of that, not only did we solve our process excursion (that was initially focused around a few parts), but we also saw improvement across all parts. The speed at which JMP allowed us to get to actionable items and the flexibility in distilling down complex issues with Graph Builder and the Profiler allowed PCC’s upper management to see the value of JMP for our company. Just this year, we are now training all our technical rotation personnel with the STIPS course to ingrain this thinking and the ability to use JMP from the beginning of their careers.
Meg: You’re describing two different kinds of newcomers to JMP. One is a person, maybe fresh out of university, who hasn’t yet used tools like JMP in an industrial workplace. The other is more advanced in their career and already has well-established preferences when it comes to statistical tools. And as a former “Minitab superuser,” you sound like you were previously in that second category!
One of the most common objections we hear – most often from that more advanced-career group – is that some fear statistical approaches in JMP will supplant their domain expertise. How has having JMP changed the way you and your team apply domain knowledge and skills?
Steve: Graph Builder and the Profiler are the Rosetta Stone for domain experts and JMP nerds. The capability for dynamic visuals basically allows domain experts to see what the data is saying across different settings and in different graphical environments. That’s really the “ah-ha” moment because they see what they have in their head – and JMP allows them to explain the links between process elements very clearly to other team members.
So, in project team meetings, I quickly see our domain experts driving the conversation – “can we look at this factor versus that one” or “let’s overlay this and that” – and then start talking to the team about what they are seeing and what can be done.
Meg: Tell us about your relationship with the JMP organization. There are some people who just use the software – and that’s fine – but there are others who really engage with JMP as an organization. What kind of value, in your experience, have you gained from taking advantage of some of the resources JMP offers?
Steve: I’ve been using JMP now for about six years (and JMP Pro for two), but it wasn’t until a couple of years ago that I really got involved in the JMP Community. I have been pretty active on the JMP Community discussion boards (mostly asking for scripting help), attended the 2019 Discovery Summit (which was awesome!) and was lucky enough to get selected to present at least year’s Discovery Summit.
I also regularly watch the on-demand webinars, have gone through the STIPS training, and have taken the online DOE training courses, so at this point you could say I’m all in!
Everything I’ve listed is the immense value-add that most people don’t know about. Or maybe they don’t understand how much value there is outside of the JMP application itself. Having a local technical resource like Jordan who I can bounce ideas off of and who can do group training sessions for us is really a game-changer when it comes to a JMP user group growing and thriving.
Meg: How would you say your advocacy for analytics transformation has impacted your career trajectory?
Steve: I’m not sure I’d be in my current role as Process Control Manager right now if it wasn’t for my love of all things data and stats. I’d held multiple operations and engineering roles before I joined the Process Control team at PCC, but I believe what helped me land the role was my understanding of how data and data-based decisions can lead to step-function changes in the control and improvement of a process. That and my interest in converting others into stat nerds!
It has also help me with exposure to other plants, networking across divisions, and exposure at technical conferences as I help others with their data analytics problems and present on what we’re doing new at my plant.
Meg: What advice would you give to someone looking to cultivate a more mature analytics culture in their organization?
Steve: I think the No. 1 thing is to force people to take the time to learn. Starting out, trying to understand data analytics, and using a new program like JMP requires a manager to say it’s OK to go slower in the short term with analytic problems so that people feel empowered to actually learn rather than resort to their old ways. When the house is on fire, people won’t naturally grab the tool that will slow them down at first, so that’s our job as change agents to let them know it’s OK to go slower at the outset to end up going faster and farther overall.
I also would recommend finding someone with a passion for data analytics and empowering them to have time to share their passion and be a contagious force. They can encourage others to take every available opportunity to try new ways of exploring data and having it lead to actionable items.
Finally, you need to share wins and make sure people are not reinventing the wheel on best analytical practices – so user group meetings, tech portals, and best practice report-outs are critical.
Meg: What advice would you give to someone who is just starting out with JMP?
Steve: Here are my top 5 tips:
- Take advantage of the free resources JMP has to offer.
- Dedicate yourself to using JMP enough to develop a “button” muscle memory.
- Make the platform more efficient by using custom toolbar buttons and scripts for frequently used actions and analysis.
- ALWAYS visualize your data before you try to do any analysis.
- Get involved in other people’s problems. You’ll be amazed at how much you learn exploring data you’re not used to and questions you’ve never thought of how to answer before.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.