How Would Taiichi Ohno Leverage Bezosism?

An article in the Wall Street Journal titled The Way Amazon Uses Tech to Squeeze Performance Out of Workers Deserves Its Own Name: Bezosism argues that the way the company manages its warehouse workers is not people-friendly. Safety incidents are above the industry average. The system of making rate discourages ad hoc bathroom breaks. There is a lot of walking. The author points out that while Amazon leaders talk about how their use of automation makes the job of the associates easier, these leaders have failed until recently to face the facts. These tools intended to be labor-saving appear to be grinding down the associates both physically and psychologically. For this, he coins the term Bezosism.

How Would Taiichi Ohno Leverage Bezosism?

Amazon chief Jeff Bezos is famously a student of the Toyota Production System. His company is no stranger to modern management practices such as Six Sigma, Agile and Lean. So it was peculiar to read the hypothetical the author posed, about a third of the way into the article.

Imagine how much additional just-in-time efficiency in inventory levels, capital allocation, and automated reordering Taiichi Ohno and Eiji Toyoda, creators of the Toyota Production System in postwar Japan, would be able to extract from a system that knew the precise moment a worker plucked an item from a shelf and sent it on its way.

To my understanding, Toyota and many other automotive manufacturers have been using this capability to optimize their use of capital for decades. They call it the kanban system. It began as cards attached to containers, but barcode scanners and digital systems have been in use since at least the mid-1990s. The kanban system is a way to trigger reordering, movement of goods and production based on a withdrawal signal of the customer or consuming process. Perhaps the author’s point, in the context of name-checking Taylorism, the Gilbreths, and Toyota, was to say, “Imagine the harm to worker well-being these people would have done,” with Amazon-like access to data.

Taiichi Ohno Stood Against Labor Intensification

But this is an odd question, since “inventory levels, capital allocation and automated reordering” has little to do with the author’s central argument around labor conditions at Amazon warehouses. Furthermore, direct labor is a small part of the cost of manufacturing an automobile. Most of it is in the design, materials, and overhead costs. Setting this aside, what would Taiichi Ohno have done if he had Amazon-like amounts of data on how long it took every worker to perform every task, at his fingertips?

Ohno and his disciplines from the Shingijutsu consulting firm, whom I spent some years with, often scolded their clients against what they called rodo-kyouka (労働強化) or labor intensification. This means the effort to raise productivity without removing waste to make the work easier. If the work takes a fixed amount of time, and we do nothing to reduce this, the only way to get more done is to work faster. There are rare cases in badly managed companies where people were standing around not working, and when put to work, doubled their output without any kaizen. Some may argue that’s labor intensification. But standing around and not working isn’t labor, so no.

People can raise productivity by speeding up their pace of work, up to a point. Industrial and process engineers understand human factors and set standard times for doing tasks, walking, etc. with allowances baked in. What Amazon is doing by using aggregate data to set a performance rate for everyone. Humans aren’t aggregate lifeforms. They are individuals, so only those fit enough can keep up with the rate. This is working as intended at Amazon, up to a point.

Keep on Fighting the Robot Fad

At Amazon there seems to be a combination of labor intensification and the use of automation to make the process easier. Ohno argued in a chapter in his book Workplace Management, not to rush into automation or the use of robots before first simplifying the process. He pointed out, “If the robot fad continues and we replace people with robots this is a problem. Robots don’t complain or ask for raises. Using robots to reduce cost may be good for the corporation, but is it good for the whole?” When a process is automated, and the human intelligence to say, “This task is hard,” is removed, progress can stop.

The author takes issue with Amazon’s use of labor-saving automation. He calls it a “classic mistake” to think that a new technology that makes people more productive will ultimately mean they have to do less work. That seems reasonable until he makes a sweeping claim without evidence in the next sentence

History shows that every time we automate a task, we tend to use more of the product or service requiring that task, in combination with others, to accomplish some other, more complicated, or difficult end.

When we no longer have to take our bars of lye soap down to the river because we have a washing machine in our home, how does this make the process more complicated or difficult? Do we do more laundry simply because we can? Or do most of our washing machines sit idle, while we spend the time we save not scrubbing wet fabric on rocks doing something more interesting and productive?

Why Doesn’t Amazon Pay a Day’s Wage for One-third Day’s Work?

There is a basic difference between working for oneself, taking on the risks and rewards of investment, and working for someone else who has done so. When I buy labor-saving automation for my home, like a washing machine, it gives me free time. When the industrial laundromat where I work invests in labor-saving machinery, it does not give me free time. One of the article’s main points is that Amazon is using automation to make it more productive, but not to make sure the work is truly easier for the employee. The result can be labor intensification.

However, in a surprising passage for the notably not anti-capitalist Wall Street Journal, the author seems to indicate that a capitalist should pass on savings from their labor-saving investments directly to the employee.

But it wasn’t as if the Kiva-using companies then reduced all their warehouse employees’ hours to a third of what they once were while paying them the same wage.

If there was a world in which the cost of installing and maintaining labor-saving robots was zero, it might make sense to pay the workers for a full day’s wage for a third of a day’s work. If a farmer invests in a tractor or combines to reduce backbreaking labor, now they have a loan to pay off. It’s not as if they can harvest their crop in a third of the time and stop. They have to get more out of their land, their time, and other resources, in order to be more productive using the new technology.

Just-in-Time and that Other Pillar

The other pillar of the Toyota Production System that we, Jeff Bezos & co., and authors of critical articles on Amazon mustn’t forget, is jidoka. This is not simply automation but autonomation which is sometimes translated as automation with human intelligence, or automation with a human touch. It began with the automatic loom over a century ago, which detected broken threads and stopped before ruining the fabric. It has evolved to mean giving people and processes the autonomy to detect, stop, call for help, and fix the problem. Jidoka not only allows every process to check for quality or process abnormality, it also allows workers to point out unreasonable conditions or overburden. The Amazon warehouses appear to be all about speeding up and simplifying the flow of goods, without the process-level stop-and-fix.

When we attempt to support a structure intended to be held up by two pillars with only one of them, bad things can happen. When we accelerate a system without building in the ability to detect errors and stop, it has the potential to go speeding off in the wrong direction. Perhaps this is happening at Amazon warehouses. I’ve also heard that newspapers are making use of machine learning and artificial intelligence to lower the cost of writing articles. I don’t know if the WSJ uses AI to assist in writing articles. Would that be an example of the author’s “classic mistake” that fails to make journalists more productive? But perhaps one day, AI could inject some jidoka to detect abnormalities, inconsistencies, and build quality into the process of writing articles.

3 Comments

  1. Keely Maitland

    October 11, 2021 - 10:44 pm
    Reply

    Thanks for a great article looking at a very nuanced aspect of Lean. I’ve been studying this stuff for 15 years and there is always more to learn.

  2. Lloyd Francis

    October 15, 2021 - 10:19 am
    Reply

    Great article. critical, educational, and not attacking.

  3. rodney brack

    October 16, 2021 - 2:38 am
    Reply

    I wonder if the people designing the building where tasked with developing a warehouse with maximum storage area at minimum cost (ie centralised toilet facilities). I think in 2 seconds Taichhi would have been asking the manager the root cause for why they think employees take so long to go to the bathroom.

    Hopefully he would have told them to draw a circle on the ground at a point far from the toilet and stand there until they worked it out. No toilet breaks allowed.

Leave a Reply

Your email address will not be published. Required fields are marked *