A dashboard can count keystrokes, logins, online minutes and closed tickets with impressive precision. That does not mean it has measured judgement, care, complexity or useful work. When leaders turn convenient activity traces into performance verdicts, the danger is not only privacy intrusion. It is false confidence: a clean number can hide a poor question.
Employee monitoring may be justified for security, safety, legal compliance or a clearly defined operational problem. It becomes weak leadership when data is collected “just in case”, interpreted without context or used to rank people before they can see and challenge it. The practical task is therefore not to decide whether all monitoring is good or bad. It is to decide what a signal can legitimately support—and where human judgement must take over.
Begin with the decision, not the dashboard
Before choosing a tool or metric, write down the decision you expect the data to inform. “We want more visibility” is not a decision. “We need to know whether overnight access attempts indicate a security incident” is. So is “We need to understand why customers wait more than two working days for a first response.”
The UK Information Commissioner’s Office says organisations must define their purpose for monitoring, avoid monitoring workers “just in case” and collect no more information than is needed. Its guidance also warns against function creep—using information for wider purposes than those originally stated. A security log gathered to detect unauthorised access should not quietly become a productivity score without a fresh legal, ethical and operational review.
- Name the problem: What specific risk, bottleneck or obligation are you addressing?
- Name the decision: What will a manager do differently after seeing the data?
- Name the boundary: What decision must the data never make on its own?
- Name the review date: When will you decide whether the monitoring still serves its purpose?
Separate a system signal from a performance judgement
Most monitoring data is a proxy: it records something easy to observe in the hope that it represents something more important. Good leaders make that gap explicit.
| Signal | What it can show | What it cannot establish alone |
|---|---|---|
| Login time | When an account accessed a system | Concentration, contribution or whether the person was solving a problem elsewhere |
| Keystrokes or mouse activity | Device interaction during a period | Useful output, thinking quality or customer value |
| Response time | Elapsed time before a recorded reply | Care, accuracy, case complexity or work completed through another channel |
| Tickets closed | Recorded throughput | Whether the issue stayed resolved or was closed prematurely |
| Time marked online | Platform presence | Availability, collaboration or sustainable performance |
This distinction matters because people adapt to what is measured. If a dashboard rewards visible busyness, a conscientious employee may spend more time producing signals and less time doing the quiet work that prevents errors. The metric has not merely observed behaviour; it has begun to shape it.
Test necessity, proportionality and alternatives
Acas advises employers to explain monitoring, consult employees and their representatives, and carry out an impact assessment. That assessment should identify the reason and likely benefit, possible negative effects, legal considerations and less intrusive alternatives. The ICO similarly treats accountability and proportionality as central, with data protection impact assessments required before processing likely to create a high risk to people’s interests.
Ask what would happen if you did nothing. Sometimes the cost of no monitoring is substantial: a regulated business may need audit logs, or a support team may need queue data to prevent requests being lost. In other cases, the existing problem may be better addressed through clearer goals, sampled quality reviews, better workload planning or a conversation with the team.
The least intrusive method that answers the legitimate question is usually the strongest starting point.
Also account for the full operating cost. That includes software, configuration, manager review time, training, data storage, security, responding to access or correction requests, and investigating false alerts. A cheap subscription can create expensive management work if it produces more noise than insight.
Treat trust and psychological reactance as operating conditions
A 2024 peer-reviewed cross-sectional survey of 633 participants examined attitudes to keyboard and camera monitoring. Privacy concerns were associated with stronger psychological reactance, and reactance was negatively correlated with attitudes towards monitoring and job satisfaction. Trust in the employer was positively correlated with attitudes towards monitoring. Because this was an observational survey, it cannot prove that monitoring caused the outcomes. It does show why a technically accurate system may still fail socially when people experience it as intrusive or illegitimate.
Trust is not a soft extra added after implementation. It affects whether people report errors, explain unusual patterns and believe that a review will be fair. Leaders can protect it by stating the purpose in plain language, limiting access, avoiding covert monitoring except in exceptional justified circumstances, and documenting how long data will be retained.
Check the work design around the data
The Health and Safety Executive’s Management Standards identify six areas of work design associated with stress when poorly managed: demands, control, support, relationships, role and change. Monitoring can interact with every one of them. A response-time target may increase demands; continuous screen capture may reduce control; unexplained scoring may damage relationships; and a rapidly introduced system may turn change itself into a stressor.
- Demands: Does the metric encourage an unsafe pace or discourage breaks?
- Control: Can people choose reasonable methods and sequence their work?
- Support: Does the data trigger coaching and resource decisions, or only criticism?
- Relationships: Will the system increase suspicion between colleagues and managers?
- Role: Are expectations clear, including which activities the system cannot see?
- Change: Have people been consulted, trained and given time to adapt?
Build a correction path before a score affects anyone
A monitored worker should be able to understand what is collected, see relevant records, explain missing context and challenge an error before the data supports a significant employment decision. The ICO gives the example of software that measures time in a case-management system but misses work done outside it because of reasonable adjustments. Treating that partial record as complete would be unfair and inadequate.
A practical correction path should answer four questions: Who can see the data? How can a person add context or correct an error? Who reviews disputes? What evidence is required before the data influences appraisal, discipline or reward? If those answers are vague, the system is not ready for high-impact use.
Illustrative example: turn a ranking into a process review
This is a fictional example designed to show the method, not a reported customer case.
A small digital-services company notices that one support adviser closes fewer tickets than colleagues. A league table would place that person last. Before judging performance, the manager samples the work and learns that the adviser handles complex account migrations, records detailed notes and prevents repeat contact. The dashboard counted volume but ignored case mix and durable resolution.
The team keeps the queue data because it is useful for staffing, but stops using raw closure counts as an individual ranking. It adds case complexity and reopen rates, reviews a sample for quality, and lets advisers flag work completed outside the ticketing system. The same data becomes more useful because its claim is narrower.
A seven-question leadership review
- What exact problem and decision justify this monitoring?
- Is the information necessary, proportionate and limited to that purpose?
- What important work or context will the system fail to observe?
- Could a less intrusive method answer the question?
- How might the measure affect demands, control, support, relationships, role and change?
- Can workers see, explain and correct the data before it affects them?
- When will we review usefulness, accuracy, retention and unintended effects—and switch it off if the case no longer holds?
Use monitoring as evidence, not a verdict
Monitoring is most defensible when a real need is clear, the method is restrained and the interpretation remains open to context. It is least defensible when leaders outsource judgement to a score simply because the score is available.
The leadership standard is straightforward: measure the work you genuinely need to understand, explain the purpose, invite challenge and keep consequences proportionate to the evidence. A dashboard can support a decision. It should not become the decision-maker.
Sources and further reading
- Information Commissioner’s Office: Data protection and monitoring workers (guidance under review following the Data (Use and Access) Act; accessed 17 October 2026)
- Acas: Monitoring performance (accessed 17 October 2026)
- Health and Safety Executive: What are the Management Standards? (accessed 17 October 2026)
- Brinson, Amini and Lemon: Electronic performance monitoring—reactance, trust, privacy concerns and job satisfaction, Cyberpsychology (2024), DOI: 10.5817/CP2024-5-10
Discover more from Marychuks.com AI, Psychology, Business & CreativeVerse
Subscribe to get the latest posts sent to your email.
