Why aren’t more Lean Six Sigma practitioners studying Data Science?
When I began debating going into data science earlier this year, I assumed that the internet would be awash with professionals who had paired the in-demand expertise with an equally-hot Lean Six Sigma (LSS) certification. After all, the role of Process Improvement Professionals (PIPs), those who practice LSS, is to find ways to minimize waste and maximize efficiency. Clearly, machine learning and data science is the most efficient way to process the large amounts of data PIPs require to perform their roles. What I found, however, is that the worlds have remained mostly separate, and that research has only recently started into looking at how the two specializations might be able to aid one another.
First, however a brief explanation of data science and LSS for those who may be unfamiliar with one or the other:
- “Data science is an interdisciplinary field that uses scientific methods, processes, algorithms, and systems to extract knowledge and insights from structural and unstructured data… It is a ‘concept to unify statistics, data analysis, machine learning, domain knowledge, and their related methods’ in order to ‘understand and analyze actual phenomena’ with data.”
- “Lean Six Sigma is a method that relies on a collaborative team effort to improve performance by systematically removing waste and reducing variation. It combines lean manufacturing… and Six Sigma to eliminate the eight kinds of waste: Defects, Over-Production, Waiting, Non-Utilized Talent, Transportation, Inventory, Motion, and Extra-Processing”
Why should someone study both data science and Lean Six Sigma?
While their stated end-goals may differ, they leverage similar methods: both fields rely heavily on educated hypothesis creation, copious data gathering, and advanced knowledge of statistical modeling. Both can benefit from a practitioner with strong business acumen, and seek to leverage large amounts of data to reduce “noise” and provide clear direction.
Here’s how the two could interact:
In LSS, a practitioner works with both observed (“go-see”) and gathered data (such as business KPIs, error reporting, and the like) to establish a baseline. They then report on that baseline’s efficiency using various statistical reporting methods and models. In this case, employing data science could increase both the depth and complexity of the models used to establish the baseline, as well as creating a more complex vision for the current state of the business.
After a baseline has been established, further statistical analysis is conducted to determine the relationships between causes (known as “X’s”) and the outputs of the processes (the “Y’s”) in order to determine the most important changes to increase efficiency and reduce waste. In this case, data science could allow the consideration of multiple concurrent variables, as well as increasing the amount of data that could be considered. This could allow for a greater number of concurrent changes because you could test the impact of changing multiple X’s at once, which would speed up the number of iterations required to maximize a process. What’s more, data scientists are adept at creating dashboards to more effectively communicate data to key stakeholders, making the oftentimes cumbersome burden of proof that a PIP faces far easier.
So… why don’t more LSS practitioners use both?
It would seem that the simplest answer is coding knowledge. While this is reasonable as a barrier to entry, this only reinforces the need for LSS-trained data scientists. For a PIP, the ultimate goal is to be more effective in day-to-day work; if data science makes someone more effective in that role, it stands to reason that it is the next logical step in the evolution of the practice. If LSS was the addition of statistical modeling to business practice, doesn’t the addition of machine learning and complex computation models make sense as the next iteration?
Just as Lean and Six Sigma became a methodology that was stronger together, perhaps Machine Learning Lean Six Sigma makes sense as the next iteration?
Ultimately, the reasons above are why I decided to go into data science: from my perspective, the writing is on the wall that the future of process improvement lies in the work of data scientists, and those who want to get ahead should make sure they have both skillsets now.
