Living In Halifax

Posted by Stuart Watt on January 14, 2022 · 2 mins read

Scientific customer management and how it breaks products


Scientific management, by the standards of the day, was wildly successful and influential, shaping everything from McDonalds to the Soviet Union – where Lenin, initially skeptical of American ideas, ended up endorsing scientific management and its benefits for “lazy” Russian workers.

Taylor identified five main principles:

  1. Science, not rule of thumb
  2. Harmony, not discord
  3. Mental revolution
  4. Cooperation, not individualism
  5. Development of every person to their greatest efficiency and prosperity

Much as Taylor gets the credit, modern scientific management is at least as influenced by William Sealy Gosset, also known as “Student”, head brewer at Guinness and statistician par excellence.

Science, not rule of thumb

First and foremost, scientific management was about replacing human judgements by “science”.

So, what’s the link between Taylor’s work in pre-Depression steel factories and the modern world? I think there are two strands.

Scientific management and the modern panopticon

First, modern internet technology, and especially artificial intelligence, has enabled scientific management to work at a far greater scale. What we’re building now is essentially a “modern panopticon”. It’s no longer a physical built structure, because it doesn’t need to be. Thanks to the deep integration of analytics into everything, decision-makers can follow activity no matter where we are, even at home, even in the middle of a pandemic.

Perhaps the greatest exponents of modern scientific management are McDonalds and Google and similar big technology companies.

In Google, for example, cooperation is reinforced by explicitly developing a social identity as a “Googler”.

It should, therefore, come as no surprise that Google is as actively anti-union as Taylor was himself. Sometimes, nothing much changes.

Github, for example, is a perfect example of a scientific management tool.

Scientific customer management

“sales engineering” https://en.wikipedia.org/wiki/W._Edwards_Deming

The second strand is related, and also depends on technology to scale. However, instead of focusing on employees, it focuses on customers. Perhaps the most well-documented model of this is Netflix and its use of “consumer science”, which is very substantially exactly the same as scientific management but for customers instead of employees.

Note that Apple, at least under Steve Jobs, didn’t use anything like the same method. Reed Hastings, Netflix’s founder, put it this way:

“Leaders like Steve Jobs have a sense of style and what customers seek, but I don’t. We need consumer science to get there”

So CRMs are tools for McDonaldization, especially in sales and marketing. By using templates and standard emails, they make communication with customers efficient, predictable, and controlled. They collect data on how often we open and read emails, and use that data to test emails — exactly like Gosset tested different types of barley in fields. The difference is, in the case of companies like McDonalds, the science is being applied to people, not plants, and to customers, not employees.


So, I’m not entirely convinced that surveillance capitalism

Is it possible, therefore, that the erosion in products is linked to the use of analytics? Where once was judgement and craft, now there is science?


Yet another lens to look at this is: risk. As Ulrich Beck has argued, the “second modernity” is one of a risk society. And the use of “science” in scientific management is really all about minimizing and eliminating risk. Particularly in a business like that of Netflix, they want to commission series that are likely to be successful. How to do that? That was exactly what Gosset’s work was about.

As T. S. Eliot put it “Where is the wisdom we have lost in knowledge? Where is the knowledge we have lost in information?” Data does not have the answers, the methods do. And it is to people like Gosset we should look for the methods.