The Privacy-First Monitoring Playbook: how to measure work without surveillance

Surveillance backfires because it costs the trust that good work depends on and measures the wrong things. Research published since 2022 links monitoring with more stress, more plans to quit and more rule-breaking, and has found no evidence that it improves performance. Measuring how work time is spent, openly and with the minimum of data, is a different practice, and it is the subject of this playbook.

Monitoring is common, and people notice it

Most US employers already use software to monitor some part of work. In an OECD survey of 6,047 firms in six countries, 90% of US firms said they use software to monitor workers, most often their working time and whether tasks get done (OECD, February 2025). Workers see it too: 51% of US workers say their employer uses technology such as computers, software, cameras or bar-code scanners to monitor them while they work (American Psychological Association, 2023). So the real question for most organizations is not whether to measure anything, but what to measure, how openly, and what to do with the result.

Monitored workers report more stress and more plans to leave

The American Psychological Association's 2023 Work in America survey of 2,515 employed US adults compared workers who know they are monitored with workers who say they aren't. The survey shows an association, not a cause, but the gap is wide.

Online survey of 2,515 employed US adults by The Harris Poll for the American Psychological Association, April 17 to 27, 2023.

The public expects the same reaction. In a Pew Research Center survey of 11,004 US adults, 81% said workers would feel inappropriately watched if employers used AI to collect and analyze information about how they do their jobs, and 70% opposed using facial recognition to analyze employees' facial expressions, while 9% favored it (Pew Research Center, April 2023).

There is no evidence that it raises performance

A meta-analysis of 94 independent samples covering 23,461 people found no evidence that electronic performance monitoring improves worker performance. Monitoring was associated with higher stress whatever form it took, and organizations that monitored more transparently and less invasively could expect more positive attitudes from their workers (Ravid and colleagues, Personnel Psychology, 2023).

It can make rule-breaking more likely

In two studies, researchers found that monitored employees were more likely to break rules, including deliberately working at a slow pace. Monitoring appeared to weaken people's sense of agency, and with it their sense of responsibility for their own conduct; when employees felt they were treated fairly, the effect was smaller (Thiel and colleagues, Journal of Management, 2023, summarized by the researchers in Harvard Business Review, June 2022).

What this means for your team

None of this says that measuring work harms people. It says that monitoring carries a cost, which rises when monitoring is invasive, secret or used to punish: screenshots, keystroke logs and cameras that capture far more than any work question needs, introduced without explanation and read without context. The research also points to what helps: monitoring that is transparent, collects less and is applied fairly. Chapter 2 turns that finding into five principles.

Ethical employee monitoring rests on five principles: start from a purpose, collect the minimum, tell people exactly what is collected and why, give employees access to their own data, and use the data fairly, with a person making every decision about people. Each principle below comes with what it looks like in practice and one question to test your program against.

1. Start from a purpose

Write down the question the data must answer before you choose what to collect, and use the data only for that purpose. "Where does our team's time go, and who is overloaded?" is a purpose. "Just to have it" and "to catch people slacking" are not.

• In practice: list the purposes in your policy, along with the uses you rule out, such as tracking people outside working hours or monitoring union activity.
• Test: could you explain to every employee, in one sentence, why each type of data is collected?

2. Collect the minimum

Use the least invasive data that answers the question. For workload and capacity questions, it is enough to know which apps and websites are used, for how long, and whether the computer was active. Screenshots, keystroke content, webcam images and audio capture private content that no workload question needs. California's privacy law already requires covered businesses to keep collection "reasonably necessary and proportionate" to its purposes (Civil Code § 1798.100).

• In practice: start from the table in chapter 3 and justify every data type you add to it.
• Test: for each data type, what decision would you make differently without it?

3. Tell people exactly what and why

Before monitoring starts, employees should know what is collected, what is never collected, why, who can see it and how long it is kept. Secret monitoring destroys trust when it is discovered, and several US states require notice anyway (chapter 4).

• In practice: a written notice and acknowledgment, a plain list of collected and never-collected data, a Q&A before launch, and visible tracking wherever the device allows it.
• Test: would an employee reading your notice be surprised by anything in the data?

4. Give employees their own data

People should be able to see what is recorded about them, ask for a copy and correct mistakes. Data that people can see is data they can explain: a long idle stretch turns out to be a workshop, not a mystery.

• In practice: give employees access to their own activity and tool data, name a contact for copies and corrections, and answer within a set number of days.
• Test: can every employee see the numbers their manager uses?

5. Use the data fairly, with a person deciding

Read trends over weeks, compare people only with others in the same role, and treat the data as one input among several. No decision about discipline, pay or termination should rest on monitoring data alone, and the same rules should apply to leaders as to everyone else. From July 1, 2027, California's SB 947 writes a version of this into law for decisions that rely on automated systems (chapter 4).

• In practice: team-level views first, individual follow-up as a conversation, and a documented human review before any decision.
• Test: if a number looked bad, would your first step be a conversation with the person?