Lyndsay Noble has both a BS and an MS in mathematics with a focus in statistics from McNeese State University and an ABD in applied statistics and research methods from the University of Northern Colorado. She has been working as an analytics professional since 2001.

In 1995, I was a second-year, undergraduate, transfer student in the math department and needed to fulfill a state requirement for an introductory math class. Since I had already taken a calculus class, I could not take college algebra, and the only class that met the state requirement was elementary statistics (strange rules I still don’t understand, but what a fortuitous happenstance). I was drawn into the subject matter and haven’t looked back.

When I reached the end of my bachelor’s program, the only job opportunities I was qualified for were programming or teaching. While I could not have told you why, neither of those roles sounded like the right fit for me. Thus, I decided to pursue a master’s to qualify for different opportunities.

At the end of that program, I felt educated, but still didn’t really understand what one does as a statistician. On the recommendation of my adviser, I entered a PhD program. My adviser thought my academic skills would serve me well in academia, and I realized more education would enable me to be selective when I finally left school to pursue a career.

I worked several hours per week in a statistical consulting lab during that program, helping students, and even did some freelance work helping local businesses. At that point, I started to understand the value I could bring to the business world and the fun of being a statistician.

When I became ready to search for full-time, gainful employment (2004), I had no idea what industry or role I wanted to try first. I must have sent out hundreds of job applications. I used every website I could find and included both the government and public sectors in my search. The only conscious decision I made was to not pursue any applications that asked for code samples because, while I had learned a common statistical scripting software for school, I was not comfortable as a coder. I’m not sure that would even be a choice in today’s environment, with most entry-level statistician roles requiring SQL, R, python, or experience with some other scripting language.

I accepted a position as a statistician in the consumer insights and product development area at the largest winery in the world after about six months of searching. Oh, my goodness. What a job! I am often asked what a statistician does at a winery. The short answer is they apply everything they have ever learned in school and then learn some more to apply. During the six years I worked at the winery, I primarily supported marketing, production, scientific research, and consumer research, but I think I might have worked with each department in the company at some point.

While working with the marketing department, the statistical applications were straightforward. We designed and analyzed conjoint studies to select new labels or brand names. We also defined control and test markets for point-of-sale marketing and price testing and advertising. We combined our learnings from the control store work with other data purchased from grocery chains and built price elasticity models using ARIMA time series with covariates. None of these methods was explicitly taught in my coursework, but creating a control group follows logically from ANOVA.

My work was predominantly sampling plans for defect testing and the application of statistical product control methodologies in the production environment. I was also involved in some LEAN/Six Sigma projects. I think production work might have been the most ‘typical statistician work’ I did. In fact, there was so much opportunity for the use of sampling plans and statistical process control methodologies that I developed a training program to educate the process engineers on the methodology so they could run it without my input.

The scientific research was fascinating stuff. The company not only sold wine, but they grew, harvested, and crushed the grapes; made the wine; and aged it. Each step involves nearly infinite variables that can be tuned. Their manipulation affects cost, quality, and the nuanced flavor and aroma of the wine. My role was to work with the scientists to design and analyze their experiments. Some experiments were about multivariate optimization, some about mixtures, and some about simple incomplete blocks. The three experimental design classes I took in graduate school only hinted at the diversity of applications.

The consumer research part of my work at the winery was the original job description; everything else was fun stuff I found to work on. Consumer research was about answering the following questions:

• What kind of wine do people like? Performing cluster analysis on the wines and people
• Why do they like it? Using sensometric and chemometric data with high multicollinearity to model consumer preference measured on a Likert scale
• Why don’t they buy more of it? Designing and analyzing consumer surveys to discover root cause and potential market impact and recommend product development to meet the consumer need

After six years at the winery, I chose to explore a different industry. Consumer food and beverage is fascinating, but big data was starting to be a catch phrase so I transitioned to a company focused on internet advertising. The data in that world is definitely big! I joined the team as a statistician, and my role was twofold: create and test targeting algorithms to get the right ad in front of the right people at the right time and analyze recent historical data to identify the root cause of unexpected changes in algorithm performance. In no time, I was leading that team and, shortly thereafter, to managing a larger team. As manager (and eventually director), I was no longer analyzing data, but prioritizing projects and working with business leaders to identify new areas in which the team could add value. The team I managed was made up of statisticians, machine learning scientists, and data engineers. I walked in knowing nothing about computer programming (some C++ in college really doesn’t count) and learned enough to help the team succeed.

When I left the winery and went into digital advertising, I started to think about where I wanted to go with my career. I have always said I love the application of statistics because it teaches me about the world. I know so much about how sun exposure in grapes changes the flavor of the wine and what levels of isobutyl methoxypyrazine in combination with low levels of fruity esters will make the wine taste vegetal. I understand the data economy on the internet and what privacy really means. I want to keep learning and be able to apply the numerous methodologies I have learned to more industries. Thus, I decided to try consulting.

I joined a start-up within a giant company in 2014. The company provides products and services to asset managers based on the transfer and processing of data. Given this abundance of data, the company chose to incubate a group that would focus on developing data-driven products and providing data management and data usage consulting to its clients. I came in to build the consulting part of that business. Four years in, I am still building, and the focus has morphed to fit the needs of the market. I spend most of my consulting time helping clients decide where to focus their energy around analytics and advising them about measuring the value of their analytics efforts. In addition to the consulting, because of the exposure I have to the clients, I create solutions for and advise our analytics product development team.

I’m not sure what will come next in my career. Perhaps I will stay on the consulting path. Alternatively, I could go deeper into leading analytics product development. There is even the lure of going back to consumer goods as a leader. I plan to work as long as my brain holds out, and there is no reason I can’t explore all these options and probably many more I haven’t even dreamed of.