Why the Obsession With How Employees Feel?

This is a Reveal post. Every third post in the series takes one decision made in building HumanSafe Career and explains why it was made that way. This one is about the step at the centre of it.

In the Toyota Production System, overburden has a name. It is muri, one of the three forms of waste the system was built to remove, alongside muda and mura.

Muri describes people or equipment pushed beyond a reasonable capacity. Toyota treats it as a property of the flow rather than a characteristic of the person standing in it, and it holds that unevenness upstream, mura, is what produces the overburden downstream.

That work was done for physical production, and it was done decades ago. The same condition in cognitive work is routed somewhere else entirely.

Why it went somewhere else

There is a practical reason, and it is not that anybody stopped caring.

Overburden on a production line is externally visible. You can stand at the point where the work piles up, count the queue, time the cycle and see the unevenness that caused it. The condition presents itself as a property of the process, because that is where it can be observed.

Cognitive overburden has no external instrument. Nothing about a saturated working memory shows up in a queue, and the attempts to observe it from outside, through message volume or calendar density or response time, are inference dressed as measurement.

So it surfaces in the only place it can, which is the person. It arrives as a feeling, and it arrives with no structural information attached to it.

Everything downstream then treats the feeling as the subject. Resilience training, mindfulness, wellbeing programmes and self-care content all take the sensation as the thing to be worked on, because the sensation is the only part of the event anybody has access to.

Ratings are feelings, with arithmetic on top

The organisations that do collect this are not measuring something other than the feeling. They are measuring the feeling and putting a number on it.

An engagement score, a pulse rating, a one to ten on how manageable the week was. Each takes the same sensation and converts it into arithmetic. The number is more tractable than the sentence it replaced, and it is not more objective, because nothing was added to it.

What the conversion does add is a subject. A feeling described in words is ambiguous about what it refers to. A feeling expressed as a score attaches firmly to the person who gave it, and can then be averaged, tracked, compared across teams and read as a property of the people who reported it.

The structural information is not lost at the point of scoring. It was never captured. A rating preserves the intensity of the reading and discards the only part that would have said what produced it.

That is why a decade of measurement has coexisted with no improvement in the conditions being measured. The instrument was pointed at the right event and recorded the wrong half of it.

The decision

We treat the feeling as an instrument reading rather than a subject.

The person declares that a piece of work felt a particular way. That declaration is converted into a structural condition, a signal such as Overloaded or Pressured, which carries with it the circumstances known to produce that condition and the approaches associated with relieving it.

What comes back is not a description of the person. It is a description of an environment that produces this reading in people.

The transformation runs inside a constraint layer we call the constitutional safety membrane, and it runs one way. Once a declaration has become a structural signal, the original wording is not retained, so there is no emotional record, no mood history, and nothing to reconstruct later.

Where the transformation happens

It also happens somewhere specific, and the location is doing as much work as the mechanism.

The transformation does not run inside the employer's systems. The declaration is made in a protected space the organisation is not party to, and the conversion into a structural condition takes place there, before anybody sees anything.

That matters because of what a declaration is. An accurate reading of your own conditions is only available where the person giving it is not simultaneously managing how it will be received.

Inside an employer's instrument they always are. Every engagement survey in existence is answered by someone doing two things at once, reporting the week and calculating the consequences of reporting it, and the second calculation contaminates the first.

A protected space is not a privacy feature added to the transformation. It is the precondition for the input being worth transforming.

Nothing an organisation sees is the declaration itself. What it can see, and on what terms, is a separate decision and belongs to a later post in this series.

What is refused, and why

The feeling does not survive the step. That is the refusal, and it is easy to mistake for a data-handling nicety when it is the whole design.

A stored feeling is a fact about a person. A structural condition is a fact about a place. Keep the first and you have built an emotional record of somebody, which is precisely the instrument this reader already has too many of pointed at them.

There is also a use question underneath it. If a feeling is retained, it can be compared, tracked and eventually read as a trend in the person, and at that point the tool has quietly changed its subject from the environment to the individual.

Someone who has spent six months asking whether the problem is them or the place has one question they need answered. What use is an instrument that answers it by keeping a record of them?

The position

Our position is that a feeling reported at work is data about an environment, and that treating it as data about a person is a category error, of a kind production engineering identified in its own domain decades ago.

The error is not primarily an ethical one, though it has ethical consequences. It is an analytical one. A signal generated by the conditions is being attributed to the sensor, which in any other engineering context would be recognised immediately as a fault in the diagnosis rather than a fault in the instrument.

Toyota did not respond to overburden by training the operator to tolerate more of it. It went upstream and looked at what the flow was doing.

The provocation we would leave open is this. Every organisation already collects the raw material, through surveys, check-ins and pulse tools of one kind or another, and most of them have been collecting it for years.

What they hold is a long series of numbers describing how people felt, attached to the people who felt it. It is worth asking what that same archive would have been worth if each reading had been recorded against the environment that produced it.

More on how HumanSafe Career works: humansafe.tech

Declaration of Generative AI and AI-assisted technologies in the writing process:
The author made use of Generative AI or AI-assisted technologies in the preparation of this post.

Sources
Toyota UK, "Muda, Muri, Mura, Toyota Production System guide"
Art of Lean, "Muda, Mura, and Muri: The Three Lenses for Finding Waste in the Toyota Production System"

The contents of this article are for informational purposes only and do not constitute professional, legal, or financial advice.