When Your Boss Can't Tell the Difference Between Work and Theater

I'll never forget the week my team lead installed keystroke monitoring software. Not directly—it was buried in an IT security email about "productivity metrics." I spent three days watching my colleagues game the system: opening and closing Slack, holding down keys with a coffee cup, refreshing the same PR they'd already read. Within a month, the tool was quietly removed. But not before it had poisoned something that took years to build back: trust.

That experience crystallized something I've been thinking about for a while now. When a company reaches for keystroke counters and activity monitors, they're not implementing a policy. They're making a public confession about their own failures in management.

The Contradiction Nobody Talks About

Here's what kills me about the surveillance trend: the same companies pushing keystroke counters are the ones evangelizing AI tools like Copilot. These tools literally exist to reduce the typing you do. Write fewer keystrokes. Better code. Faster iteration.

So we're in this absurd position where leadership simultaneously says: "Use AI to write smarter code faster" and "We'll measure you by how many keys you press per hour." It's not just contradictory—it reveals the actual problem. They don't know how to measure what matters, so they measure what's easy to count.

I've written my cleanest code in moments of complete stillness. Staring at a problem. Sketching on a whiteboard. Sometimes deleting more than I wrote because I realized a better approach. None of that registers on a keystroke counter. Meanwhile, someone frantically fixing their own mistakes registers as "hardworking." The metric punishes thoughtfulness and rewards chaos.

The Real Cost Nobody Measures

The University of Pittsburgh study from 2024 caught my attention precisely because it shows what I've observed firsthand: strict RTO mandates with activity monitoring don't improve financial performance or productivity. What they do improve? Turnover rates among senior engineers jump by 14%. Time-to-fill open roles stretches to 23% longer.

Your best people leave first. They can leave—they have options. Then you spend months trying to replace them with someone who doesn't yet know your codebase. Meanwhile, you're spending budget on monitoring software that isn't solving the actual problem.

I've talked to three senior developers in Islamabad who left their last role specifically because of surveillance culture. Not because they were slacking—because they couldn't work in an environment where their judgment wasn't trusted.

What Actually Matters

If I were building a culture of real accountability, I'd measure outcomes. Deployment frequency. Bug reoccurrence. Time spent in code review providing feedback. How many junior developers your senior people mentored. Whether the system stayed up. Whether we shipped what mattered this quarter.

Those are hard to measure. They require understanding context. They demand taste and judgment. They're not reducible to a spreadsheet.

But here's the thing: if you can't distinguish good work from bad work without a keystroke counter, the problem wasn't the employees. It was the hiring, the onboarding, or the management. You picked the wrong people, didn't set clear expectations, or never learned what they actually do.

My Question for You

I want to ask you honestly: has any monitoring metric your company adopted actually improved work quality? Or did it just get better at making people look busy?

In my experience, the metrics shape behavior more than they measure it. Activity monitoring selects for people who are good at appearing productive, not people who are actually good at their craft. That's a race to the bottom.

If you're managing engineers and considering surveillance tools, ask yourself first: do I trust my team? If yes, you don't need the tool. If no, the tool won't fix the real problem—you should probably fix the hiring and management instead.

The keystroke counter isn't a safety net. It's an admission that something broke, and instead of fixing it, you're going to watch everyone more closely.

Source: This post was inspired by "The keystroke-counting RTO metric is a confession, not a policy" by Dev.to. Read the original article