Latest / The Productivity Podcast with Fexingo: Output, Efficiency, and Long-Term Economic Growth / The Productivity Cost of Workplace Surveillance
Transcript
- Lucas: There's a number that's been stuck in my head since I read a study from the University of Oxford's Internet Institute last year. Workers who are actively monitored by surveillance software — keystroke logging, screen captures, mouse movement tracking — are about 12 percent less productive on complex problem-solving tasks than workers who aren't. Luna: Twelve percent less productive even though the whole point of the software is to make sure people are working. That is a brutal irony. Lucas: Exactly. And it's not just productivity — the same study found that employee turnover in firms using this kind of real-time surveillance is about 8 percent higher than in comparable firms that don't. So you're getting less output and you're losing people faster. Luna: Is there a specific company case in that study or is it aggregate data? Lucas: They had some really telling firm-level comparisons. One was a midsize software company — about 400 engineers — that rolled out keystroke monitoring in 2023. Over the next eighteen months, voluntary bug reporting dropped by 14 percent and code-review quality scores actually declined. The engineers were spending so much cognitive energy trying to look productive that they had less bandwidth for actual problem-solving. Luna: Right — they're gaming the metrics instead of doing the work. That's a textbook example of Goodhart's Law, isn't it? When a measure becomes a target, it ceases to be a good measure. Lucas: That's exactly the framework. And the researchers explicitly cite Goodhart. They argue that keystroke counts and mouse movement are proxy metrics that become useless once you start rewarding or punishing based on them. Workers adapt their behaviour to the surveillance, not to the actual output. Luna: But there's also a simpler psychological channel at play. Even if you're not trying to game anything, being watched changes how you think. I remember a study — I think from Harvard Business Review — where monitored workers reported higher anxiety and lower job satisfaction, which then showed up in their performance metrics. Lucas: The Oxford study found the same thing. They measured self-reported stress levels alongside actual output, and there was a clear negative correlation. The surveillance itself is a cognitive load. Your brain is processing the fact that someone might be watching at any moment, and that uses attention you could be spending on the task. Luna: So the cost isn't just gaming the system — it's the distraction of being monitored, full stop. Lucas: Right. And this is where I think the privacy debate often misses the point. People say, 'If you have nothing to hide, you have nothing to fear.' But the research shows that even workers who report having nothing to hide still perform worse under surveillance. It's not about hiding something — it's about your brain's attentional system being disrupted by the awareness of observation. Luna: That's a really important distinction. So the productivity cost is universal, not just for people who are slacking off. Lucas: Exactly. And the magnitude is not trivial. Twelve percent on complex tasks — that's roughly the productivity boost you'd get from adding a full hour to a knowledge worker's effective day. You're essentially giving back that gain through surveillance. Luna: Do we know which types of work are most affected? I imagine data entry might be less harmed than, say, creative strategy. Lucas: The study broke it down by task complexity. For routine, repetitive tasks — like data entry or basic customer service — the productivity loss was only about 3 percent. But for tasks requiring problem-solving, creativity, or collaboration, it jumped to that 12 percent figure. And for tasks involving high-level strategic thinking, it was even higher — around 16 percent. Luna: So the more valuable the work, the more surveillance damages it. That is a terrible incentive structure for any company that claims to value innovation. Lucas: Right. And there's a counter-case in the same research. They looked at a company that actually moved in the opposite direction — a financial services firm that had used monitoring for years and then removed all real-time surveillance in 2024, switching to purely outcome-based management. Within six months, voluntary error reporting went up 30 percent, and the quality of client-facing work improved measurably. Luna: Because people felt safe to surface problems instead of hiding them. Lucas: Exactly. When you're being watched, you hide your mistakes. When you're evaluated on outcomes, you want to find and fix mistakes before they become client issues. It flips the incentive completely. Luna: I think a lot of managers reach for surveillance tools because they're concrete and measurable. Trust is abstract. A dashboard with green dots feels like control. But the research is showing that control has a real price tag. Lucas: And it's not just the productivity loss. There's also the selection effect — the kind of workers who are most comfortable under surveillance tend to be those who are less innovative, less autonomous. Over time, you're shaping your workforce toward compliance rather than initiative. Luna: That's a longer-term cultural cost that probably doesn't show up in a quarterly earnings report but absolutely affects competitiveness over five or ten years. Lucas: The Oxford researchers actually modelled that. They simulated a five-year horizon for a typical mid-size firm and found that the cumulative effect of elevated turnover and reduced innovation could wipe out roughly 4 percent of annual revenue growth. That's a substantial number. Luna: So what's the alternative? You still need some way to evaluate performance, especially in remote or hybrid settings. Lucas: The alternative is outcome-based management — which is harder to implement because you have to define what a good outcome looks like for each role. But the firms that do it well tend to use very few metrics, reviewed quarterly, not in real time. They focus on deliverables and impact, not on activity. Luna: And presumably those metrics are harder to game because they're tied to actual business results. Lucas: Exactly. And they require managers to actually understand what their teams do, which is a different skill from watching a dashboard. But the firms that invest in that managerial capability see better retention and higher productivity — the Oxford study found that the outcome-based firms in their sample had 5 percent higher total factor productivity than the surveillance-heavy ones. Luna: I think there's a parallel here to what we talked about in the episode on managerial overhead. Surveillance is a form of overhead — it consumes resources and attention without producing value. Lucas: That's exactly right. And the irony is that the overhead of surveillance often replaces the overhead of good management. If you're watching keystrokes, you might feel like you don't need to have the harder conversation about what good work looks like. Luna: If today's episode has given you a new lens on your own workplace — or maybe validated something you've felt — that's exactly what we're here for. These conversations stay ad-free because of listener support, and if you want to help keep them that way, you can do that at buy me a coffee dot com slash fexingo. Lucas: Yeah, it genuinely makes a difference. A lot of people contribute a small amount each month, and that's what lets us spend the time digging into studies like this Oxford one instead of chasing sponsors. So thank you to everyone who does. Luna: And on that note — one thing I want to push back on a little. You've framed this as a pure productivity cost argument, and I think that's powerful. But there's also a dignity argument. Even if surveillance didn't hurt productivity at all, there's something fundamentally dehumanizing about being watched every moment of your workday. Lucas: I don't disagree. I focus on the productivity angle because it's the language that gets through to managers who are making the purchasing decisions on these tools. But you're right — the human cost is real even if the numbers were neutral. Luna: And in a tight labour market, that dignity factor becomes a retention factor. Workers have choices, and a lot of them are choosing not to work under surveillance. The data on quit rates in surveilled firms bears that out. Lucas: The study I mentioned earlier — the one with the 8 percent higher turnover — that's in a labour market where workers had options. So that's a real cost to the firm in recruitment and training. And it's entirely avoidable. Luna: So the productivity case and the dignity case converge on the same conclusion: less surveillance, more trust, better outcomes. Lucas: That's the bottom line. And I'd love to see more research on the long-term effects — most of these studies cover one to two years. What happens to a company's innovation pipeline after a decade of surveillance culture? My guess is it's not pretty. Luna: Maybe that's the episode we do in a year or two, when the data catches up. Lucas: I'll be watching for it.