Activity Is Not Work
The central measurement problem, stated plainly: the signal and the thing it stands for are only loosely related.
Monitoring captures activity at a keyboard. Work is what gets achieved. The relationship between them is weak, inconsistent, and varies by exactly the thing that matters most.
The distinction in “Activity Is Not Work” is important when operational records are interpreted. Organisations researching dual n back training can use cognitive training and dual n-back to add time and project context, while outcomes, direct feedback and human review remain necessary to explain what the numbers do not show.
Where they come apart
Thinking produces no input and is most of the valuable work in many roles.
For an independent reference relevant to “Activity Is Not Work”, consult the European Commission data-protection resources; it provides a useful external check on scope, terminology, governance and the claims made during procurement or review.
Reading produces almost none.
A conversation that resolves a problem registers as a gap.
Meanwhile copying files, reorganising folders and reformatting a document produce a great deal of measured activity and no output.
The inversion
The more skilled the work, the weaker the relationship.
Data entry has a genuine correlation between keystrokes and output.
Design, analysis, writing, diagnosis and advice have close to none, and in places an inverse one — the person who solved it in ten minutes of thought looks worse than the one who spent two hours flailing.
Why this is not a tuning problem
It is not that the thresholds are wrong.
The signal does not contain the information.
No amount of configuration extracts output from input counts, and vendors who claim otherwise are claiming to have solved a problem that is not solvable with this data.
What the signal is genuinely good for
Detecting that somebody is not at the machine at all, over long periods.
Identifying specific prohibited activity where that is defined and narrow.
Establishing time at a terminal where a contract requires it.
These are real, narrow and are not productivity measurement, which is the whole distinction.
The averaging trap
Across a large team over a long period, activity and output correlate somewhat, because people who do nothing do appear.
Which produces a tempting conclusion: the measure works in aggregate.
It does, weakly, and that aggregate correlation is then applied to individuals, where it does not hold at all. That step is where the harm happens.
What to measure instead
Whatever your business already counts: things delivered, resolved, produced, sold.
Imperfect, gameable, and about outcomes rather than about keystrokes.
Its own note covers measuring output, and the short version is that most organisations already have these figures and do not look at them.
The sentence to hold
You are measuring the shape of somebody's hands, not the value of their work.
Everything that follows in this section is a consequence of that.
What to check
For your roles, how much valuable work involves no keyboard?
Would your best performer score well?
Has anybody compared activity scores against actual output?
And what does your business already count that you are not looking at?