On Nov. 16, Chris Gotwalt will present a new webcast on Best Practices in Model Comparison, which will include a sneak peek of JMP 10 (due to be released this upcoming March). I recently caught up with Chris to find out how he was enjoying his new role at JMP, what new features we can look forward to in JMP 10 and what inspires him and his teams to create these new features we’ll soon be enjoying.
In April, Chris was promoted to Director of Statistical Research and Development overseeing both statistical development and testing teams and has been with JMP for 10 years. His undergraduate degree is in mathematics, and he holds a PhD in statistics. He discovered his interest in the field during the last semester of college, after he realized that statistics combined many of his diverse interests, from math to computer science and philosophy of science.
Chris’ main focus in his new role also happens to be what he’s enjoying most -- collecting customer feedback and working hard to ensure that a product that’s being released is completely customer-motivated.
“When we begin to create and look at revisions and enhancements, we actually visualize that there is someone on the other side anticipating and waiting to use what we are developing. I’m motivated by making things easier for our customers and by improving elements that will help successfully accomplish their tasks. Everything we do, we do because we have someone very particular in mind,” he says.
Other primary goals of the statistical development team and testing teams are to shake out usability snags swiftly and to provide more and more reliable analytical capabilities with each release. What kinds of questions does R&D ask? Here’s a good starting template from Chris to provide the foundation for a new emergence of ideas:
Ask yourself every week: What can we do better?
What would make JMP easier to teach?
Where do people get frustrated?
What’s something our customers have to do over and over again that could be automated?
What can we do to help our customers who are working to fulfill organizational or governmental mandates? What tools would make things easier for them?
Members of Chris’ team seek out multiple outlets for gaining this feedback, internally from sales, SAS education and marketing peers, but also externally through conferences, WebEx preview sessions, alpha and beta presentations, meetings with leading university professors, statistical counsels, related professional organizations and users groups.
One of Chris’ feedback channels provided unexpected results, leading the direction for huge changes in JMP 10. Chris began reviewing tracks from technical support and found that control charts were the fourth most-used platform in JMP, as measured by number of technical support tracks generated. “I was very surprised to make that discovery. We hadn’t done anything substantial with that platform in six or seven years, but once we saw the tech support logs, I immediately met with Leo Wright in product marketing to examine what would be the next generation for control chart tools.”
Also, some of the greatest perspectives come from the instructors in SAS education, which has for many years been a source of ideas for the development of JMP. Every class is 30 new people and 30 new perspectives on usability, analyses and questions related to industry-specific challenges. “Sometimes instructors may get a question for which they don’t have an answer. So, they write a script. The script gets popular, and then people add to it to use and share. It’s when this happens that we realize it should be native to JMP,” Chris explains.
What is the result of all of this feedback? JMP 10 launches in March of 2012 and will include these new capabilities:
Control Chart Builder
Nonlinear Regression made simple
Design evaluation platform
New measurement systems analysis platform
Partial Least Squares
Model Comparison platform
One-Click Bootstrapping of any quantity in JMP
Reliability Growth Analysis
My final question to Chris was "What recommendations would you make to folks just getting started with analytics?" His answer: "Find out where your company’s data is, load it into JMP and start asking questions. Look for surprises. Where are the problem areas? What are people asking questions about? I frequently shift development priorities, based on seeing what our customers are trying to use our product for each day. Share your information with others and ask what does this mean? Most importantly, allow yourself to be surprised."
The learning loop is continuous, and the JMP statistical development team is wrapping up work on JMP 10, and will soon be adding new nonlinear models and data mining techniques that will be part of JMP 11. Chris concludes, “We’re excited because there’s a lot we can do to make our product more modern and usable.”