The Observer Effect in Engineering Management
- Marko Belusic
- Jul 31
- 3 min read
Introduction – From Quantum Labs to Engineering Teams
Most managers trust their metrics. That’s their first mistake.
In physics, the observer effect tells us the act of measuring changes the thing being measured. At the quantum scale, you can’t look at a particle without altering its behavior. In teams, you can’t track a metric without changing how people work, and often, what you end up measuring isn’t reality, it’s the performance your measurement system incentivized.
The Observer Effect in Physics
At the subatomic scale, particles don’t behave like billiard balls, they behave more like probabilities. To learn anything about them, scientists have to interact with them.
For example, to detect an electron’s position, you bounce a photon off it. That collision reveals where the electron is, but it also nudges the electron, changing what it was doing.
In quantum experiments, measurement is inseparable from outcome. There’s no “passive” way to observe without influencing the system.
The lesson: observation always alters reality. In physics, scientists must account for this. In engineering management, so should we.
The Observer Effect in Engineering Management Teams
Engineering teams aren’t made of electrons, but they are made of humans, humans who adapt behavior when they know they’re being measured, monitored, or watched.
Sometimes the adaptation is conscious, like shaping a sprint update to sound better. Sometimes it’s unconscious, like changing how you speak in a meeting because a senior leader is present.
This is why the observer effect in engineering management shows up so often in metrics, what you measure can distort how teams behave.
Examples of Distortion in Engineering Management
Velocity Tracking Measuring story points delivered per sprint often pushes teams to inflate estimates or split work into meaningless tickets. Numbers go up, but real value doesn’t.
Bug Count Metrics Using raw bug counts as a proxy for quality can incentivize hiding or reclassifying issues. On paper, quality “improves,” but real problems remain.
Standups When senior leaders join standups, engineers shift into reporting mode, downplaying blockers.
Other Variations Commit frequency, Jira throughput, and even A/B tests can all be gamed in similar ways, optimized for optics rather than outcomes.
So, What’s the Solution?
Just as physicists can’t avoid measurement, managers can’t avoid metrics. The answer isn’t to stop measuring, it’s to design observation methods that minimize distortion and account for their effects.
Measure What Matters
Focus on metrics that reflect value, not just activity.
For example: cycle time can be more meaningful than raw commit counts, and customer adoption or retention can be more valuable than number of features shipped.
Balance multiple measures so no single one drives all behavior.
Use Metrics as Learning Tools, Not Surveillance
Co-create metrics with the team so they’re seen as shared feedback, not policing.
Treat data as a conversation starter, always paired with context and qualitative insight.
Be Thoughtful in How You Observe
Don’t micromanage in real time, favor retrospectives, sampling, and periodic reviews.
Be transparent about what’s measured and why, making clear it’s about improving the system, not individuals.
Invest in Team Culture
Build psychological safety so people can be honest about blockers and setbacks.
Frame metrics as a team mirror, not a personal microscope.
A strong culture acts like “measurement shielding” , it won’t erase distortion, but it reduces it.
Closing – Lead Like a Scientist
In quantum mechanics, measurement changes reality. In engineering management, observation changes behavior.
Our job is to design measurements that bring us closer to the truth, while building a culture that resists distortion.
“When a measure becomes a target, it ceases to be a good measure.” - Goodhart's law
