Jan 24 2022
Does Amazon Use Lean, Six Sigma, or Lean Six Sigma?
In 2019, Christoph Roser posted six articles on his blog about the inner workings of Amazon Fulfillment Centers, based on visits to locations in the US and Germany. His blog is called AllAboutLean but the word “Lean” appears nowhere in his articles about Amazon. “Six Sigma” does not appear either, and Christoph does not mention meeting any black belt.
In addition, in Working Backwards: Insights, Stories, and Secrets from Inside Amazon (2021), Amazon alumni Colin Bryar and Bill Carr make no reference to Lean, and all they report about Six Sigma is using DMAIC to define metrics.
Yet you find some published descriptions of Amazon as a showcase for Lean, Six Sigma, or Lean Six Sigma but, if you consider them without confirmation bias, the evidence is underwhelming. The keywords appear, along with a few more, like “Operational Excellence” or “Scrum.”
Based on the small amount of published data, the leaders of Amazon, starting with Jeff Bezos, “learned a bunch of techniques, like Six Sigma and lean manufacturing and other incredibly useful approaches.”
In other words, they learned everything they could get their hands on while staking out uncharted territory. Then they developed their own system. Now they are sharing with outsiders a few homilies but no details, as is their privilege. Their system is to retail as Toyota’s is to manufacturing. It’s not reducible to Lean, Six Sigma, or Lean Six Sigma.
“Six Sigma as a problem-solving methodology causes many hang-ups for Japanese managers. Many Americans seeking training in Six Sigma in Japanese organizations face resistance with little explanation as to why. This often leads to frustration and contempt towards management. They write off the Japanese resistance to the training as resistance to change, preventing growth and feeling unrepresented.
Apr 3 2026
Deviating Standard Deviations
This basic concept deserves revisiting. The following is from a blog post from 2022 hosted by a supplier of statistical software intended to explain the meaning of some notations in plain, simple terms:
The author calls two different things by the same name. If the standard deviation of each variable is 1, how could its expected value be anything else? The confusion within this nonsensical statement is the same we make when we equate the temperature of a soup with a thermometer reading. In our mental model of a bowl of soup, it has a temperature that exists regardless of our ability to measure it, and the thermometer reading is only an estimate of it.
For the purposes of eating soup, confusing the two is harmless, unless the thermometer, poorly calibrated, always gives you an answer that is 15°F off. This is the situation we have with the most commonly used estimator of the standard deviation of a random variable from a small sample. It is biased, and c_4(n) is a correction factor applicable when the random variable is Gaussian.
To describe c_4(N) accurately, we need to dig into probability theory. It is, in fact, the expected value of the estimator S=\sqrt{\frac{1}{N-1}\sum_{i=1}^{N}\left ( X_i -\bar{X} \right )} of the standard deviation from a sample of N independent Gaussian variables \left ( X_1, \dots, X_N \right ) with unit standard deviation, \sigma = 1. This is an accurate statement, but every term in it needs an explanation.
Share this:
Like this:
By Michel Baudin • Technology 0 • Tags: Control Charts, Probability, Quality, Six Sigma, SPC, Standard Deviation, statistics