Visual Six Sigma: making data analysis Lean

Introduction

This paper introduces the idea of “Visual Six Sigma”, a practical and pragmatic approach to data analysis and process improvement. This approach has been developed in response to a growing business need to broaden the use of six sigma type thinking beyond the realms of highly trained and “statistically savvy” Black Belts and Green Belts.

In the typical business environment of process improvement, what people are looking for today are simple to use tools that can be widely used by everyone at all levels to rapidly explore and interpret data, and then use that understanding to drive improvement. By making these tools highly visual and engaging we can accelerate the process of analysis and eliminate the need for advanced statistical analysis in all but the most complex of situations.

We can also broaden and deepen the application of Six Sigma thinking in the organisation by making the tools intuitive, easy to use, and the results easy to interpret.

This article describes and illustrates the Visual Six Sigma approach based around a case study, but first lets set the scene typical of many business environments and ask a critical question:

So Be Honest... How Much Heavy Duty Statistics Do We Really Need To Drive Process Improvement?

Many of you will be familiar with the Dabawalla story. A story of Bombay’s extraordinarily efficient lunch system which has operated for more than a century. Last spring The Times (UK) reported :

"Just after 11am every workday, Bombay’s famous dabbawallas stream off the city’s railway network into the downtown business district to deliver hot, home-made meals to an army of hungry office workers. Carrying tiffin boxes lovingly packed by wives and mothers in nearly 200,000 surburban kitchens, these 5,000 lunch delivery workers are part of one of the world’s most admired distribution systems. Employing a complex colour-coded logistics process, the dabbawallas (can-carriers) complete a door-to-door service across 15 miles (25km) of public transport and 6 miles (10km) of road with multiple transfer points in a three-hour period."

In a system finely tuned over 120 years they maintain an error rate of only one in eight million ( >7 sigma performance)... and they do this without statistical analysis at all!

In a recent analysis of lean six sigma deployment in a large multinational we also made some very interesting observations (see figure 1). Not only are about 80% of the typical business population either terrified or very uncomfortable with statistics, also > 80% of the project value comes from projects where only very basic tools and/or modest statistical analysis was required to identify and deliver the improvement.

Figure 1: Comfort levels with statistical methods

So, if most people are terrified of statistical methods and try to avoid the methods taught at black belt level and above, how can we make data analysis simple, quick, intuitive, practical and engaging for the typical business so that they achieve data driven solutions rapidly and with minimum overhead.

In this article we describe the approach we call “Visual Six Sigma” based around exploiting the capabilities of JMP software. This approach focuses heavily on using a range of very powerful and easy to understand visual tools to rapidly identify “Hot Xs”. The statistical rigour can be used (but only as much as required) to underpin this and then to easily build models to simulate and do “what ifs” to assess improvement opportunities.

Not only is this software analysis package easy to use – Its visual capabilities and accessible output makes the findings very easy to communicate and to engage with leaders in getting support for the improvement activities….. another challenge in many continuous improvement deployments!

The Visual Six Sigma approach is described in section 3 and then demonstrated in some detail in section 4 based around a fictional (but fairly typical case study).