Unexpected Application of Lean Six Sigma Tools in Daily Operational Activities

Sometimes operational people from my organisation approach me with “the” 1000-Euro-question: “Six Sigma guy, can you please help me analyze this set of data.” Asking me to help, they usually have an issue they really can’t resolve themselves. Perceiving Lean Six Sigma too far away from daily operations they turn in despair to the Six Sigma guy, hoping he can help and his expensive training program can be justified after all…

My standard answer is (of course, because I want to prove Lean Six Sigma is beneficial for daily operations) “Sure, let’s sit down for 10 minutes and discuss your question”. Sometimes, I find there is no clearly defined purpose for the data analysis, or there is no clear answer to the simple, but fundamental question “What do you want the data to tell you?” In such case, Six Sigma guy can’t help either…

Luckily, many times, the purpose is clear, or the situation described above can be cleared out asking the 5 why questions. Then it sometimes astonishes what results you can get using simple tools as Pareto charts, histograms, run charts, box plots and/or scatter plots.

Last week I even astonished myself when a manager approached me to help him organize price quote data from multiple subcontractors in such a way that decisions could be made. Each subcontractor quoted unit costs for more then 100 different specific tasks. You can image how that data matrix looked like in Excel…

Together with the manager, I applied 2 different tools to organize the data and draw conclusions:

1. Box plots. The box plot outliers showed us clearly which quotes were unrealistically high and low. Using the brush function in Minitab, we could deduct that 80% of the high value outliers were caused by one subcontractor. On the low values, we saw a similar phenomenon.

2. Using the characteristics of the normal distribution. Per task we have 8 data values available. Not all these data sets were normally distributed, but we did the following: calculation of the average x-bar and the standard deviation s. A characteristic of the normal distribution is that 68% of the data is covered by x-bar +/- 1 s. Or, 32% of the data can be found outside these values, meaning 16% at the lower end and 16% at the higher end. Guess what? 99% of the data points in this lower and upper end of the distribution also showed up as outliers in our box plot analysis.

In the past the decision would have been to grant the contract to the cheapest subcontractor without asking further questions. This has lead to quality losses and rework in the past. Today, we’ll invite the cheapest and ask clarification to ensure us of the understanding of our requirements, thus ensuring us of 1st time right quality at the right price.

And I can assure you, by this unexpected application of simple tools, this manager is now 200% convinced of the value add Lean and Six Sigma tools can bring to the daily business of the organisation!