Measurement · 5.3

Monitoring, and what it establishes

Monitoring, and what it establishes. What is actually the case, and how it compares with what is repeated.

For a practical comparison of the systems discussed here, see employee monitoring software, which covers how monitoring can be configured and reviewed.

Electronic monitoring of employees expanded sharply after 2020 and the question it raises is narrow: what does it establish, and what is it being used to conclude.

What is measured

Time at a device. Keystrokes and mouse movement. Applications in focus. Screenshots at intervals. Location, for drivers and field staff. Message content, in some systems. Badge and sensor data for movement within a building.

Every one of these measures activity, presence or position. None measures output, and none measures quality.

The gap that follows

A metric of activity is satisfied by activity. Devices sold specifically to simulate mouse movement exist and sell well, which is the clearest possible demonstration that the measure and the thing are separable.

This is the mechanism from the previous entry, applied to the most easily gamed measure available.

What the evidence says about effects

Mostly that monitoring reduces trust and raises turnover intention, and that perceived fairness of the monitoring moderates the effect substantially: monitoring explained in advance, limited to what the job requires, and applied to everybody produces much less damage than monitoring discovered.

There is a serious counter-example and it should be stated. A study of theft monitoring in restaurant chains found not only reduced theft but increased revenue per shift, apparently because staff redirected effort into selling. Where the behaviour being monitored is discrete, attributable and genuinely undesirable, monitoring can work as intended.

That is a narrow condition and most workplace monitoring does not meet it.

The legal position, in outline

In the European Union, processing employee data requires a lawful basis, must be necessary and proportionate to a specified purpose, and generally requires an impact assessment for systematic monitoring. Consent is a weak basis in an employment relationship because it cannot be freely given. In some member states works councils must agree.

In the United States the position is considerably weaker, with federal law permitting most employer monitoring of employer systems and several states requiring notice.

Nothing here is legal advice and the rules differ substantially by jurisdiction.

What this site will not help with

Covert monitoring. It is unlawful in many jurisdictions, it destroys the trust that makes an organisation work when discovered, and it is discovered. No entry here will describe how to do it, and the omission is deliberate rather than an oversight.

The question to ask before installing anything

What decision will be made differently as a result. If the answer is vague, the system will collect data that is used retrospectively to justify decisions taken on other grounds, which is the commonest actual use and the least defensible one.

If the answer is specific, the follow-up is whether a narrower instrument would answer it. Frequently one would.

Proportionality in practice

Location tracking on a delivery vehicle during working hours is proportionate to an obvious purpose. The same tracking outside working hours is not, and systems that cannot distinguish are common.

The general test applied in European law is whether the same purpose could be achieved by a less intrusive means, and asking it honestly eliminates a large share of what is deployed.

What monitoring cannot fix

A manager who does not know what their team is producing has a management problem, and activity data will not supply the answer. It will supply numbers, which is different and considerably more comfortable.

The organisations that report the least value from monitoring are generally the ones that installed it in place of deciding what good work looked like.

For the person being monitored

Ask what is collected, for what purpose, how long it is retained and who sees it. In several jurisdictions you are entitled to be told, and in most workplaces nobody has asked, so the answer takes a while to arrive.

The productivity claim

Monitoring vendors sell on the promise of productivity gains, and the supporting evidence is generally vendor case studies rather than controlled comparison.

Where independent evaluation exists, effects on output are small and effects on measured activity are large, which is exactly what one would predict from the entry on Goodhart's law.

Telematics in vehicles

The best-established case for monitoring: harsh braking, speeding and cornering data has a defensible safety purpose and a measurable safety effect in several fleet studies.

It is also the case where scope creep is most visible, since the same device records location continuously and the temptation to use it for timekeeping is immediate.

Who is monitored most

Lower-paid workers, consistently, across every survey that asks. Monitoring intensity correlates inversely with pay and with autonomy, which means the least measured people are the ones whose decisions carry most consequence.

For broader context, consult ICO guidance on monitoring workers.