BuzbyTimeRead the requirements guide

What monitoring does to people · 7.4

Autonomy, and why it is not a perk

Autonomy, and why it is not a perk. What decides it, what it costs, and what usually goes wrong. For another implementation reference, see this Monitask guide.

Autonomy is not a benefitHow to think about it

It is treated in the research literature as a determinant of wellbeing rather than an amenity. Job design work going back decades has consistently found that the combination of high demands with low control is the damaging one, and that control moderates the effect of demand.

Which means an intervention that increases perceived demand while reducing perceived control is operating on both terms in the wrong direction at once. Monitoring generally does exactly that. For broader background, see RingCentral.

What reduces perceived control

  • Being observed without knowing when.
  • Data collected about you that you cannot see.
  • Decisions about your work taken from a record you cannot correct.
  • Rules about how work is done rather than what is achieved.
  • Not knowing what the measurement will be used for.

Every item is a design choice and every one has an obvious opposite, which is the practical value of the list.

What preserves it while still recording

The person sees everything. Removes items one, two and five at once.

The person can correct their record. Removes the third.

The categories were chosen with them. Turns the record into something they helped design.

The purpose is stated and narrow. Which is what makes the rest credible.

Recording your own time is not the problem

Professionals in many fields record their time meticulously and do not experience it as surveillance, because the record belongs to them, serves a purpose they recognise, and is not used to observe them.

That is the existence proof: the practice can be entirely compatible with autonomy, and whether it is depends on decisions made in the first month.

Micromanagement with better instruments

The failure mode is not new. Monitoring software makes it cheaper, more precise and more scalable, and it does not create it: a manager inclined to control the manner of work will do so with whatever is available.

Which means the technology decision and the management culture interact. The same product in two organisations produces two different experiences, and the product is not the variable that explains it.

Where autonomy is genuinely limited

Safety-critical work, regulated processes, and roles where the method matters as much as the outcome. Those exist, the constraints are legitimate, and they are constraints on the method rather than licence to observe the person.

Even there, the four preserving practices apply in full.

The question to ask about any feature

Does this give the person more information about their own work, or somebody else more information about the person?

Almost every design decision in this category falls cleanly on one side, and the answer predicts how it will be received better than any explanation of intent.

Two people, same system, different experience

One chose the categories, sees their own data, corrects it freely and knows what it is for. The other had it applied, cannot see it, and has heard it might be used in a review.

The software is identical. Everything that determines the outcome sits in decisions nobody thinks of as product decisions.

Autonomy and accountability are not opposites

The common objection to everything in this entry is that somebody has to be accountable. True, and accountability is for outcomes rather than for method.

A person held to what they deliver, with discretion over how, is both accountable and autonomous. The arrangement that fails is control over method with no clarity about outcome, which is surprisingly common and is the worst of both.

Giving control back cheaply

Let people set their own categories within a structure. Let them correct anything. Show them their own patterns before showing anybody else. And ask them what would make the arrangement better twice a year.

Four changes, none of which costs anything, and together they shift where the record sits between the person and the organisation.

What to do first

Check whether every person can see everything held about them, today, without asking. If not, that is the first thing to fix and it is usually a setting.

A short summary

Autonomy is treated as a determinant of wellbeing rather than an amenity, and control moderates the effect of demand. Five things reduce perceived control and each has an obvious opposite. The same product produces different experiences depending on decisions nobody thinks of as product decisions. And the useful test is whether a feature informs the person or informs somebody about the person.

Where this bites hardest

Roles where the work is intellectually demanding and the output is slow to appear: research, design, complex analysis. High demand, low observable output, and any control over method lands directly on the part that produces the value.

One line to carry

Accountability for outcomes and discretion over method are compatible. Control over method with no clarity about outcome is the arrangement that fails.

Where to start

Check that everybody can see everything held about them today without asking anybody. It is usually a setting and it is the highest-value one in the product.

The word to avoid

Empowerment. It is used to describe autonomy granted conditionally and withdrawn at will, which is not autonomy, and everybody subject to it can tell.

Describe what somebody may decide without asking. That is specific, checkable and does not require a word that has been spent.

Also in what monitoring does to people